UNIVERSAL SERIAL BUS

DEVICE CLASS DEFINITION

FOR
AUDIO DEVICES

Release 4.0
April 2023
Universal Serial Bus Device Class Definition for Audio Devices

SCOPE OF THIS RELEASE

This document is the Release 4.0 of this Device Class Definition.

CONTRIBUTORS

Chris March Apple Inc.

Rhoads Hollowell Apple Inc.
Girault Jones Apple Inc.
Geert Knapen (Co-Chair & Editor) Apple Inc.
E-mail: [email protected]
Eric Schulz Apple Inc.
Abdul Rahman Ismail (Co-Chair) Intel Corporation
Devon Worrell Intel Corporation
Niel Warren Google
Terry Moore MCCI Corporation
Egidio Sburlino Microsoft Corporation
Franz Detro Native Instruments
Mike Kent Roland
Brad Saunders SpecWerkz
Bob Dunstan SpecWerkz
Morten Christiansen Synopsys
Paul E. Berg (Program Manager) USB-IF

REVISION HISTORY

Rev. Date Filename Description

1.0 Mar. 1998 Audio10.pdf Release 1.0
2.0 May 2006 Audio20 final.pdf Release 2.0
3.0 Sep. 2016 Audio30.pdf Release 3.0
4.0 Apr. 2023 Audio40.pdf Release 4.0

Release 4.0 April 2023 2

Universal Serial Bus Device Class Definition for Audio Devices

Copyright © 1997-2023 USB Implementers Forum, Inc.

All rights reserved.

INTELLECTUAL PROPERTY DISCLAIMER

A LICENSE IS HEREBY GRANTED TO REPRODUCE THIS SPECIFICATION FOR INTERNAL USE ONLY. NO OTHER LICENSE,
EXPRESS OR IMPLIED, BY ESTOPPEL OR OTHERWISE, IS GRANTED OR INTENDED HEREBY.

USB-IF AND THE AUTHORS OF THIS SPECIFICATION EXPRESSLY DISCLAIM ALL LIABILITY FOR INFRINGEMENT OF
INTELLECTUAL PROPERTY RIGHTS RELATING TO IMPLEMENTATION OF INFORMATION IN THIS SPECIFICATION. USB-
IF AND THE AUTHORS OF THIS SPECIFICATION ALSO DO NOT WARRANT OR REPRESENT THAT SUCH
IMPLEMENTATION(S) WILL NOT INFRINGE THE INTELLECTUAL PROPERTY RIGHTS OF OTHERS.

THIS SPECIFICATION IS PROVIDED “AS IS” AND WITH NO WARRANTIES, EXPRESS OR IMPLIED, STATUTORY OR
OTHERWISE. ALL WARRANTIES ARE EXPRESSLY DISCLAIMED. USB-IF, ITS MEMBERS AND THE AUTHORS OF THIS
SPECIFICATION PROVIDE NO WARRANTY OF MERCHANTABILITY, NO WARRANTY OF NON-INFRINGEMENT, NO
WARRANTY OF FITNESS FOR ANY PARTICULAR PURPOSE, AND NO WARRANTY ARISING OUT OF ANY PROPOSAL,
SPECIFICATION, OR SAMPLE.

IN NO EVENT WILL USB-IF, MEMBERS OR THE AUTHORS BE LIABLE TO ANOTHER FOR THE COST OF PROCURING
SUBSTITUTE GOODS OR SERVICES, LOST PROFITS, LOSS OF USE, LOSS OF DATA OR ANY INCIDENTAL,
CONSEQUENTIAL, INDIRECT, OR SPECIAL DAMAGES, WHETHER UNDER CONTRACT, TORT, WARRANTY, OR
OTHERWISE, ARISING IN ANY WAY OUT OF THE USE OF THIS SPECIFICATION, WHETHER OR NOT SUCH PARTY HAD
ADVANCE NOTICE OF THE POSSIBILITY OF SUCH DAMAGES.

NOTE: VARIOUS USB-IF MEMBERS PARTICIPATED IN THE DRAFTING OF THIS SPECIFICATION. CERTAIN OF THESE
MEMBERS MAY HAVE DECLINED TO ENTER INTO A SPECIFIC AGREEMENT LICENSING INTELLECTUAL PROPERTY
RIGHTS THAT MAY BE INFRINGED IN THE IMPLEMENTATION OF THIS SPECIFICATION. PERSONS IMPLEMENT THIS
SPECIFICATION AT THEIR OWN RISK.

Dolby™, AC-3™, Pro Logic™ and Dolby Surround™ are trademarks of Dolby Laboratories, Inc.
All other product names are trademarks, registered trademarks, or service marks of their respective owners.

Please send comments via electronic mail to [email protected]

Release 4.0 April 2023 3

Universal Serial Bus Device Class Definition for Audio Devices

TABLE OF CONTENTS

Scope of This Release .................................................................................................................................................. 2

Contributors ................................................................................................................................................................ 2
Revision History........................................................................................................................................................... 2
Table of Contents ........................................................................................................................................................ 4
List of Tables ................................................................................................................................................................ 9
List of Figures ............................................................................................................................................................ 14
1 Introduction ...................................................................................................................................................... 16
1.1 Scope ......................................................................................................................................................... 16
1.2 Purpose ..................................................................................................................................................... 16
1.3 Related Documents ................................................................................................................................... 16
1.4 Terms and Abbreviations .......................................................................................................................... 17
2 Management Overview .................................................................................................................................... 22
2.1 Overview of Key Differences between ADC 2.0 and ADC 4.0 ................................................................... 22
3 Functional Characteristics ................................................................................................................................. 24
3.1 Introduction .............................................................................................................................................. 24
3.2 Basic Audio Device Examples .................................................................................................................... 25
3.3 Backwards Compatibility ........................................................................................................................... 26
3.4 Audio Interface Association (AIA) and Interface Association Descriptor ................................................. 26
3.4.1 Audio Function Class ......................................................................................................................... 26
3.4.2 Audio Function Subclass ................................................................................................................... 26
3.4.3 Audio Function Protocol ................................................................................................................... 26
3.5 Audio Interface Class ................................................................................................................................. 27
3.6 Audio Interface Subclass ........................................................................................................................... 27
3.7 Audio Interface Protocol ........................................................................................................................... 27
3.8 Clock Domains ........................................................................................................................................... 27
3.9 Power Domains ......................................................................................................................................... 27
3.10 Groups ....................................................................................................................................................... 28
3.11 Audio Synchronization Types .................................................................................................................... 28
3.11.1 Asynchronous .................................................................................................................................... 28
3.11.2 Synchronous ...................................................................................................................................... 28
3.11.3 Adaptive ............................................................................................................................................ 28
3.11.4 Implications of the Different Synchronization Types ........................................................................ 28
3.12 Inter Channel Synchronization .................................................................................................................. 31
3.13 Audio Function Topology .......................................................................................................................... 31
3.13.1 AudioControls.................................................................................................................................... 37
3.13.2 Cluster ............................................................................................................................................... 38

Release 4.0 April 2023 4

Universal Serial Bus Device Class Definition for Audio Devices

3.13.3 Input Terminal ................................................................................................................................... 39

3.13.4 Output Terminal ................................................................................................................................ 41
3.13.5 Mixer Unit ......................................................................................................................................... 44
3.13.6 Selector Unit ...................................................................................................................................... 45
3.13.7 Feature Unit ...................................................................................................................................... 46
3.13.8 Sampling Rate Converter Unit........................................................................................................... 47
3.13.9 Effect Unit.......................................................................................................................................... 48
3.13.10 Processing Unit.................................................................................................................................. 53
3.13.11 Extension Unit ................................................................................................................................... 57
3.13.12 Clock Entities ..................................................................................................................................... 58
3.13.13 Connector Entities ............................................................................................................................. 59
3.13.14 Power Domains and Power Domain Entities .................................................................................... 59
3.13.15 ChannelGroups, EntityGroups, and CommitGroups ......................................................................... 60
3.14 Operational Model .................................................................................................................................... 60
3.14.1 AudioControl Interface...................................................................................................................... 61
3.14.2 AudioStreaming Interface ................................................................................................................. 61
3.14.3 AudioControls.................................................................................................................................... 63
3.14.4 Clock Model ....................................................................................................................................... 64
3.14.5 Connector Model .............................................................................................................................. 65
3.14.6 Power Domain Model ....................................................................................................................... 65
3.14.7 Additional Power Considerations and Requirements ....................................................................... 67
3.14.8 Group Model ..................................................................................................................................... 69
3.14.9 Binding between Buttons and AudioControls .................................................................................. 69
4 Descriptors ........................................................................................................................................................ 71
4.1 Standard Descriptors ................................................................................................................................. 71
4.2 Class-specific Descriptors .......................................................................................................................... 71
4.2.1 Traditional Class-specific Descriptor Characteristics ........................................................................ 71
4.2.2 Extended Class-specific Descriptor Characteristics .......................................................................... 72
4.2.3 Common Fields in some Class-specific Descriptors .......................................................................... 73
4.3 Audio Function Descriptor Set .................................................................................................................. 73
4.3.1 Audio Function Descriptor Set Layouts ............................................................................................. 74
4.4 Cluster Descriptor ..................................................................................................................................... 76
4.4.1 Cluster Descriptor Header ................................................................................................................. 77
4.4.2 Cluster Descriptor Block .................................................................................................................... 78
4.4.3 Example Cluster Descriptor ............................................................................................................... 87
4.4.4 CEA-861.2 Channel Mapping ............................................................................................................ 88
4.4.5 Physical versus Logical Cluster .......................................................................................................... 89
4.5 AudioControl Interface Descriptors .......................................................................................................... 89

Release 4.0 April 2023 5

Universal Serial Bus Device Class Definition for Audio Devices

4.5.1 Standard AC Interface Descriptor ..................................................................................................... 89

4.5.2 Class-specific AC Interface Descriptor............................................................................................... 90
4.5.3 Class-specific AC Interface Building Block Descriptors ..................................................................... 91
4.6 AudioControl Endpoint Descriptors ........................................................................................................ 125
4.6.1 AudioControl Control Endpoint Descriptors ................................................................................... 125
4.6.2 AudioControl Interrupt Endpoint Descriptors ................................................................................ 125
4.7 AudioStreaming Interface Descriptors .................................................................................................... 126
4.7.1 Standard AudioStreaming Interface Descriptor.............................................................................. 126
4.7.2 Class-specific AudioStreaming Interface Descriptor ....................................................................... 127
4.7.3 Class-specific AudioStreaming Self Descriptor................................................................................ 127
4.7.4 Class-specific AudioStreaming Valid Frequency Range Descriptor................................................. 128
4.8 AudioStreaming Endpoint Descriptors.................................................................................................... 129
4.8.1 AudioStreaming Isochronous Audio Data Endpoint Descriptors .................................................... 129
4.8.2 AudioStreaming Isochronous Feedback Endpoint Descriptor ........................................................ 130
4.9 Class-specific String Descriptors.............................................................................................................. 131
5 Commands & Requests ................................................................................................................................... 133
5.1 Standard Requests .................................................................................................................................. 133
5.2 Class-specific Audio Function Management Commands ........................................................................ 133
5.2.1 Commit ............................................................................................................................................ 133
5.2.2 Switch Function ............................................................................................................................... 134
5.3 Class-specific AudioControl Commands.................................................................................................. 134
5.3.1 Class-specific AudioControl Command Structure and Layout ........................................................ 135
5.3.2 AudioControl Commands Characteristics ....................................................................................... 138
5.3.3 AudioControl Commands ................................................................................................................ 150
5.3.4 AudioStreaming Commands ........................................................................................................... 187
5.3.5 Additional Commands ..................................................................................................................... 189
6 Interrupts ........................................................................................................................................................ 193
6.1 Interrupt Message ................................................................................................................................... 193
6.1.1 Interrupt Message Header .............................................................................................................. 193
6.1.2 AudioControl Interrupt Message Body ........................................................................................... 194
6.1.3 Extended Descriptor Interrupt Message Body................................................................................ 195
6.1.4 Class-specific String Descriptor Interrupt Message Body ............................................................... 195
6.2 Interrupt Behavior ................................................................................................................................... 196
7 Audio Data Formats ........................................................................................................................................ 197
7.1 Service Interval and Service Interval Packet Definitions......................................................................... 198
7.2 Simple Audio Data Formats..................................................................................................................... 198
7.2.1 Type I Formats ................................................................................................................................. 198
7.2.2 Type III Formats ............................................................................................................................... 202

Release 4.0 April 2023 6

Universal Serial Bus Device Class Definition for Audio Devices

7.3 Extended Audio Data Formats ................................................................................................................ 203

7.3.1 Extended Type I Formats................................................................................................................. 204
7.3.2 Extended Type III Formats .............................................................................................................. 205
7.4 Class-specific AudioStreaming Self Descriptor........................................................................................ 206
7.5 Auxiliary Protocols................................................................................................................................... 207
7.5.1 HDCP Protocol ................................................................................................................................. 207
7.6 Adding New Audio Data Formats ............................................................................................................ 208
7.7 Adding New Auxiliary Protocols .............................................................................................................. 208
8 Backwards Compatibility Considerations ....................................................................................................... 209
8.1 Simple Multi-Mode Device Example ....................................................................................................... 210
8.2 Discovery of Higher Version Level Support for an Audio Function......................................................... 212
8.3 Retrieval of an HRL Audio Function Descriptor Set ................................................................................ 213
8.4 Switching the Audio Function from BRL Operation to HRL Operation ................................................... 214
Appendix A. Audio Device Class Codes ............................................................................................................ 215
A.1 BOS Capability Codes .............................................................................................................................. 215
A.2 Audio Function Class Code ...................................................................................................................... 215
A.3 Audio Function Subclass Codes ............................................................................................................... 215
A.4 Audio Function Protocol Codes ............................................................................................................... 215
A.5 Audio Interface Class Code ..................................................................................................................... 215
A.6 Audio Interface Subclass Codes .............................................................................................................. 215
A.7 Audio Interface Protocol Codes .............................................................................................................. 216
A.8 Audio Class-specific Traditional Descriptor Types .................................................................................. 216
A.9 Audio Class-specific Traditional Descriptor Subtypes ............................................................................. 216
A.10 Audio Class-specific Extended Descriptor Types .................................................................................... 216
A.11 Audio Class-specific Extended Descriptor Subtypes ............................................................................... 216
A.12 Descriptor Variant Types ......................................................................................................................... 217
A.13 Cluster Descriptor Segment Types .......................................................................................................... 217
A.14 Channel Purpose Definitions ................................................................................................................... 218
A.15 Channel Relationship Definitions ............................................................................................................ 218
A.16 Ambisonic Component Ordering Convention Types .............................................................................. 221
A.17 Ambisonic Normalization Types .............................................................................................................. 221
A.18 Terminal Companion Segment Types ..................................................................................................... 221
A.19 Effect Unit Effect Types ........................................................................................................................... 222
A.20 Processing Unit Process Types ................................................................................................................ 222
A.21 Class-specific Request Codes .................................................................................................................. 222
A.22 Class-specific Attribute Codes ................................................................................................................. 223
A.23 Control Selector Codes ............................................................................................................................ 223
A.23.1 Terminal Control Selectors .............................................................................................................. 223

Release 4.0 April 2023 7

Universal Serial Bus Device Class Definition for Audio Devices

A.23.2 Mixer Unit Control Selectors ........................................................................................................... 223

A.23.3 Selector Unit Control Selectors ....................................................................................................... 224
A.23.4 Feature Unit Control Selectors........................................................................................................ 224
A.23.5 SRC Unit Control Selectors .............................................................................................................. 224
A.23.6 Effect Unit Control Selectors ........................................................................................................... 225
A.23.7 Processing Unit Control Selectors ................................................................................................... 226
A.23.8 Extension Unit Control Selectors .................................................................................................... 227
A.23.9 Clock Source Control Selectors ....................................................................................................... 228
A.23.10 Clock Selector Control Selectors ..................................................................................................... 228
A.23.11 Connector Control Selectors ........................................................................................................... 228
A.23.12 Power Domain Control Selectors .................................................................................................... 228
A.23.13 AudioStreaming Interface Control Selectors .................................................................................. 228
A.24 Connector Types ..................................................................................................................................... 229
A.25 AudioControl Capabilities Overview ....................................................................................................... 230
A.26 Interrupt Source Types ............................................................................................................................ 233
Appendix B. Audio Data Format Codes ........................................................................................................... 234
B.1 Audio Data Formats ................................................................................................................................ 234
B.2 SubHeader Codes .................................................................................................................................... 235
B.3 Audio Format General Constants ............................................................................................................ 235

Release 4.0 April 2023 8

Universal Serial Bus Device Class Definition for Audio Devices

LIST OF TABLES

Table 3-1: Output Cluster Configuration and Cluster Content Behavior .................................................................. 67
Table 4-1: Traditional Class-specific Descriptor Layout ............................................................................................ 71
Table 4-2: Class-specific Descriptor Layout............................................................................................................... 72
Table 4-3: Cluster Descriptor Header ........................................................................................................................ 77
Table 4-4: Cluster Descriptor Segment ..................................................................................................................... 78
Table 4-5: End Block Segment ................................................................................................................................... 79
Table 4-6: Channel Relationships .............................................................................................................................. 80
Table 4-7: Information Segment ............................................................................................................................... 85
Table 4-8: Ambisonic Segment.................................................................................................................................. 86
Table 4-9: Channel Description Segment .................................................................................................................. 86
Table 4-10: Cluster Descriptor Example .................................................................................................................... 87
Table 4-11: Standard AC Interface Descriptor .......................................................................................................... 89
Table 4-12: Class-specific AC Interface Descriptor ................................................................................................... 90
Table 4-13: Self Descriptor ........................................................................................................................................ 91
Table 4-14: Input Terminal Descriptor ...................................................................................................................... 92
Table 4-15: Output Terminal Descriptor ................................................................................................................... 94
Table 4-16: Terminal Companion Descriptor Header ............................................................................................... 96
Table 4-17: Terminal Companion Descriptor Segment ............................................................................................. 98
Table 4-18: End Block Segment ................................................................................................................................. 98
Table 4-19: EN 50332-2 Acoustic Level Segment ...................................................................................................... 98
Table 4-20: EN 50332-2 Voltage Level Segment ....................................................................................................... 99
Table 4-21: Bandwidth Segment ............................................................................................................................... 99
Table 4-22: Magnitude Segment ............................................................................................................................. 100
Table 4-23: Magnitude/Phase Segment ................................................................................................................. 100
Table 4-24: Position_XYX Segment ......................................................................................................................... 101
Table 4-25: Position_RΘΦ Segment ....................................................................................................................... 101
Table 4-26: Mixer Unit Descriptor........................................................................................................................... 103
Table 4-27: Selector Unit Descriptor ....................................................................................................................... 105
Table 4-28: Feature Unit Descriptor ....................................................................................................................... 105
Table 4-29: Sampling Rate Converter Unit Descriptor ............................................................................................ 107
Table 4-30: Effect Unit Descriptor........................................................................................................................... 108
Table 4-31: Parametric Equalizer Section Effect Unit Descriptor ........................................................................... 108
Table 4-32: Reverberation Effect Unit Descriptor .................................................................................................. 109
Table 4-33: Modulation Delay Effect Unit Descriptor ............................................................................................. 110
Table 4-34: Dynamic Range Compressor/Expander Effect Unit Descriptor ........................................................... 111
Table 4-35: Common Part of the Processing Unit Descriptor ................................................................................. 113

Release 4.0 April 2023 9

Universal Serial Bus Device Class Definition for Audio Devices

Table 4-36: Up/Down-mix Processing Unit Descriptor ........................................................................................... 114

Table 4-37: Channel Remap Processing Unit Descriptor ........................................................................................ 115
Table 4-38: Stereo Extender Processing Unit Descriptor........................................................................................ 116
Table 4-39: Multi-Function Processing Unit Descriptor .......................................................................................... 117
Table 4-40: Extension Unit Descriptor .................................................................................................................... 119
Table 4-41: Clock Source Descriptor ....................................................................................................................... 120
Table 4-42: Clock Selector Descriptor ..................................................................................................................... 121
Table 4-43: Connector Entity Descriptor ................................................................................................................ 122
Table 4-44: Power Domain Entity Descriptor ......................................................................................................... 123
Table 4-45: EntityGroup Descriptor ........................................................................................................................ 124
Table 4-46: CommitGroup Descriptor ..................................................................................................................... 124
Table 4-47: Standard AudioControl Interrupt Endpoint Descriptor ....................................................................... 125
Table 4-48: Standard AudioStreaming Interface Descriptor................................................................................... 126
Table 4-49: Class-specific AudioStreaming Interface Descriptor ............................................................................ 127
Table 4-50: Class-specific AudioStreaming Self Descriptor..................................................................................... 127
Table 4-51: Class-specific AudioStreaming Valid Frequency Range Descriptor...................................................... 128
Table 4-52: Standard AudioStreaming Isochronous Audio Data Endpoint Descriptor ........................................... 129
Table 4-53: Standard AudioStreaming Isochronous Feedback Endpoint Descriptor ............................................. 130
Table 4-54: Class-specific String Descriptor ............................................................................................................ 132
Table 5-1: Set/Get Function Request Layout .......................................................................................................... 133
Table 5-2: Push and Pull Command Request Layout .............................................................................................. 135
Table 5-3: Push Command Parameter Block Layout ............................................................................................... 136
Table 5-4: Pull Command Set Request Parameter Block Layout ............................................................................ 137
Table 5-5: Pull Command Get Request Parameter Block Layout............................................................................ 138
Table 5-6: Capabilities Attribute DataPart .............................................................................................................. 140
Table 5-7: AudioControl Command AddressPart Layout ........................................................................................ 143
Table 5-8: 1-byte AudioControl CUR or NEXT DataPart .......................................................................................... 145
Table 5-9: 1-byte AudioControl RANGE DataPart (Array of Triplets) ..................................................................... 145
Table 5-10: 1-byte AudioControl RANGE DataPart (ValueList) ............................................................................... 146
Table 5-11: 2-byte AudioControl CUR or NEXT DataPart ........................................................................................ 146
Table 5-12: 2-byte AudioControl RANGE DataPart (Array of Triplets) ................................................................... 146
Table 5-13: 2-byte AudioControl RANGE DataPart (ValueList) ............................................................................... 147
Table 5-14: 4-byte AudioControl CUR or NEXT DataPart ........................................................................................ 147
Table 5-15: 4-byte AudioControl RANGE DataPart (Array of Triplets) ................................................................... 147
Table 5-16: 4-byte AudioControl RANGE DataPart (ValueList) ............................................................................... 148
Table 5-17: AudioControl Table with Default DataPart Layout .............................................................................. 150
Table 5-18: AudioControl Table with Custom DataPart.......................................................................................... 150
Table 5-19: Bypass Control Characteristics ............................................................................................................. 151

Release 4.0 April 2023 10

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-20: Cluster Control Characteristics ............................................................................................................ 152

Table 5-21: Cluster Active Control Characteristics ................................................................................................. 152
Table 5-22: Latency Control Characteristics ........................................................................................................... 153
Table 5-23: Voltage Control Characteristics ........................................................................................................... 154
Table 5-24: Momentary Exposure Level Control Characteristics ........................................................................... 155
Table 5-25: Overload Control Characteristics ......................................................................................................... 155
Table 5-26: Clipping Control Characteristics ........................................................................................................... 156
Table 5-27: Mixer Control Characteristics .............................................................................................................. 157
Table 5-28: Selector Control Characteristics ........................................................................................................... 158
Table 5-29: Mute Control Characteristics ............................................................................................................... 159
Table 5-30: Gain Control Characteristics ................................................................................................................ 159
Table 5-31: Bass Control Characteristics ................................................................................................................. 160
Table 5-32: Mid Control Characteristics ................................................................................................................. 161
Table 5-33: Treble Control Characteristics.............................................................................................................. 161
Table 5-34: Band Numbers and Nominal Midband Frequencies (ANSI S1.11-2004 Standard) .............................. 161
Table 5-35: Graphic Equalizer CUR or NEXT Control Characteristics ...................................................................... 163
Table 5-36: Graphic Equalizer RANGE Control Characteristics ............................................................................... 164
Table 5-37: Automatic Gain Control Characteristics ............................................................................................... 165
Table 5-38: Delay Control Characteristics ............................................................................................................... 165
Table 5-39: Bass Boost Control Characteristics ...................................................................................................... 166
Table 5-40: Loudness Control Characteristics ......................................................................................................... 166
Table 5-41: Input Gain Pad Control Characteristics ................................................................................................ 167
Table 5-42: Phase Inverter Control Characteristics ................................................................................................ 167
Table 5-43: Center Frequency Control Characteristics ........................................................................................... 168
Table 5-44: Qfactor Control Characteristics............................................................................................................ 169
Table 5-45: Gain Control Characteristics ................................................................................................................ 169
Table 5-46: Type Control Characteristics ................................................................................................................ 170
Table 5-47: Level Control Characteristics................................................................................................................ 170
Table 5-48: Time Control Characteristics ................................................................................................................ 171
Table 5-49: Feedback Control Characteristics ........................................................................................................ 171
Table 5-50: Pre-Delay Control Characteristics ........................................................................................................ 172
Table 5-51: Density Control Characteristics ............................................................................................................ 172
Table 5-52: Hi-Freq Roll-Off Control Characteristics ............................................................................................... 173
Table 5-53: Balance Control Characteristics ........................................................................................................... 174
Table 5-54: Rate Control Characteristics ................................................................................................................ 174
Table 5-55: Depth Control Characteristics .............................................................................................................. 175
Table 5-56: Time Control Characteristics ................................................................................................................ 175
Table 5-57: Feedback Control Characteristics ........................................................................................................ 176

Release 4.0 April 2023 11

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-58: Compression Ratio Control Characteristics ......................................................................................... 177

Table 5-59: MaxAmpl Control Characteristics ........................................................................................................ 177
Table 5-60: Threshold Control Characteristics ........................................................................................................ 178
Table 5-61: Attack Time Control Characteristics..................................................................................................... 178
Table 5-62: Release Time Control Characteristics .................................................................................................. 179
Table 5-63: Width Control Characteristics .............................................................................................................. 180
Table 5-64: Algo Present Control Characteristics ................................................................................................... 182
Table 5-65: Enable Algo Control Characteristics ..................................................................................................... 182
Table 5-66: Sampling Frequency Control Characteristics ....................................................................................... 184
Table 5-67: Status Control Characteristics .............................................................................................................. 184
Table 5-68: Clock Selector Control Characteristics ................................................................................................. 185
Table 5-69: Insertion Detect Control Characteristics .............................................................................................. 186
Table 5-70: Power State Control Characteristics .................................................................................................... 187
Table 5-71: Active Alternate Setting Control Characteristics ................................................................................. 188
Table 5-72: Valid Alternate Settings Control Characteristics .................................................................................. 189
Table 5-73: CS String Descriptor Command AddressPart Layout ........................................................................... 189
Table 5-74: Extended Descriptor Command AddressPart Layout .......................................................................... 191
Table 5-75: Paged Extended Descriptor Command AddressPart Layout ............................................................... 192
Table 6-1: Interrupt Message Header Format ........................................................................................................ 194
Table 6-2: AudioControl Interrupt Message Format .............................................................................................. 194
Table 6-3: Extended Descriptor Interrupt Message Format ................................................................................... 195
Table 6-4: Class-specific String Descriptor Interrupt Message Format .................................................................. 195
Table 7-1: Packetization .......................................................................................................................................... 200
Table 7-2: SIPDescriptor Layout .............................................................................................................................. 204
Table 7-3: Class-specific AudioStreaming Self Descriptor....................................................................................... 206
Table 7-4: HDCP SubHeader Layout ........................................................................................................................ 207
Table 8-1: BOS HRL Function Capability Descriptor ................................................................................................ 213
Table 8-2: Function Container Descriptor ............................................................................................................... 213
Table A-1: BOS Capability Codes ............................................................................................................................. 215
Table A-2: Audio Function Class Code..................................................................................................................... 215
Table A-3: Audio Function Subclass Codes ............................................................................................................. 215
Table A-4: Audio Function Protocol Codes ............................................................................................................. 215
Table A-5: Audio Interface Class Code .................................................................................................................... 215
Table A-6: Audio Interface Subclass Codes ............................................................................................................. 215
Table A-7: Audio Interface Protocol Codes ............................................................................................................. 216
Table A-8: Audio Class-specific Traditional Descriptor Types ................................................................................. 216
Table A-9: Audio Class-specific Traditional Descriptor Subtypes ........................................................................... 216
Table A-10: Audio Class-specific Extended Descriptor Types ................................................................................. 216

Release 4.0 April 2023 12

Universal Serial Bus Device Class Definition for Audio Devices

Table A-11: Audio Class-specific Extended Descriptor Subtypes ............................................................................ 216

Table A-12: Descriptor Variant Types ..................................................................................................................... 217
Table A-13: Cluster Descriptor Segment Types....................................................................................................... 217
Table A-14: Channel Purpose Definitions ............................................................................................................... 218
Table A-15: Channel Relationship Definitions......................................................................................................... 218
Table A-16: Ambisonic Component Ordering Convention Types ........................................................................... 221
Table A-17: Ambisonic Normalization Types .......................................................................................................... 221
Table A-18: Terminal Companion Segment Types .................................................................................................. 221
Table A-19: Effect Unit Effect Types........................................................................................................................ 222
Table A-20: Processing Unit Process Types............................................................................................................. 222
Table A-21: Class-specific Request Codes ............................................................................................................... 222
Table A-22: Class-specific Attribute Codes ............................................................................................................. 223
Table A-23: Terminal Control Selectors .................................................................................................................. 223
Table A-24: Mixer Unit Control Selectors ............................................................................................................... 223
Table A-25: Selector Unit Control Selectors............................................................................................................ 224
Table A-26: Feature Unit Control Selectors ............................................................................................................ 224
Table A-27: SRC Unit Control Selectors ................................................................................................................... 224
Table A-28: Parametric Equalizer Section Effect Unit Control Selectors ................................................................ 225
Table A-29: Reverberation Effect Unit Control Selectors ....................................................................................... 225
Table A-30: Modulation Delay Effect Unit Control Selectors ................................................................................. 225
Table A-31: Dynamic Range Compressor/Expander Effect Unit Control Selectors ................................................ 226
Table A-32: Up/Down-mix Processing Unit Control Selectors ................................................................................ 226
Table A-33: Channel Remap Processing Unit Control Selectors ............................................................................. 226
Table A-34: Stereo Extender Processing Unit Control Selectors ............................................................................ 227
Table A-35: Multi-Function Processing Unit Control Selectors............................................................................... 227
Table A-36: Extension Unit Control Selectors ......................................................................................................... 227
Table A-37: Clock Source Control Selectors ............................................................................................................ 228
Table A-38: Clock Selector Control Selectors .......................................................................................................... 228
Table A-39: Connector Control Selectors ................................................................................................................ 228
Table A-40: Power Domain Control Selectors ......................................................................................................... 228
Table A-41: AudioStreaming Interface Control Selectors ....................................................................................... 228
Table A-42: Connector Types .................................................................................................................................. 229
Table A-43: AudioControl Capabilities Overview .................................................................................................... 230
Table A-44: Interrupt Source Types ........................................................................................................................ 233
Table B-1: Audio Data Formats in the wFormat Field and Usage ........................................................................... 234
Table B-2: SubHeader Codes ................................................................................................................................... 235
Table B-3: General Constants .................................................................................................................................. 235

Release 4.0 April 2023 13

Universal Serial Bus Device Class Definition for Audio Devices

LIST OF FIGURES

Figure 3-1: Audio Interface Associations and Interfaces .......................................................................................... 24

Figure 3-2: Example Audio Function Global View ..................................................................................................... 25
Figure 3-3: General PCC Inheritance Rules ............................................................................................................... 33
Figure 3-4: Single Input Pin PCC Inheritance Rules ................................................................................................... 33
Figure 3-5: Inside the Audio Function ....................................................................................................................... 37
Figure 3-6: Input Terminal Icon ................................................................................................................................. 41
Figure 3-7: Output Terminal Icon .............................................................................................................................. 43
Figure 3-8: MEL Procedure ........................................................................................................................................ 44
Figure 3-9: Mixer Unit Icon ....................................................................................................................................... 45
Figure 3-10: Selector Unit Icon.................................................................................................................................. 46
Figure 3-11: Feature Unit Icon .................................................................................................................................. 47
Figure 3-12: Sampling Rate Converter Unit Icon ...................................................................................................... 48
Figure 3-13: PEQS Effect Unit Icon ............................................................................................................................ 50
Figure 3-14: Reverberation Effect Unit Icon ............................................................................................................. 51
Figure 3-15: Modulation Delay Effect Unit Icon ....................................................................................................... 51
Figure 3-16: Dynamic Range Compressor/Expander Transfer Characteristic .......................................................... 52
Figure 3-17: Dynamic Range Compressor/Expander Effect Unit Icon ...................................................................... 53
Figure 3-18: Up/Down-mix Processing Unit Icon ...................................................................................................... 55
Figure 3-19: Channel Remap Processing Unit Icon ................................................................................................... 55
Figure 3-20: Stereo Extender Processing Unit Icon .................................................................................................. 56
Figure 3-21 Multi-Function Processing Unit Icon...................................................................................................... 57
Figure 3-22: Extension Unit Icon ............................................................................................................................... 58
Figure 3-23: Clock Source Icon .................................................................................................................................. 59
Figure 3-24: Clock Selector Icon ................................................................................................................................ 59
Figure 3-25: N-channel AudioControl with Cluster-wide AudioControl ................................................................... 64
Figure 3-26: Typical behavior of Audio Device that implements LPM/L1 ................................................................ 68
Figure 4-1: Audio 2.0 Function Descriptor Set layout ............................................................................................... 74
Figure 4-2: Audio 3.0 Function Descriptor Set layout ............................................................................................... 75
Figure 4-3: Audio 4.0 Function Descriptor Set layout ............................................................................................... 76
Figure 4-4: Cluster Descriptor ................................................................................................................................... 77
Figure 4-5: Cluster Descriptor Block.......................................................................................................................... 78
Figure 4-6: 3D Representation of the Channel Relationships ................................................................................... 84
Figure 4-7: Terminal Companion Descriptor ............................................................................................................. 96
Figure 4-8: Terminal Companion Channel Block ....................................................................................................... 97
Figure 4-9: Mixer internals ...................................................................................................................................... 103
Figure 5-1: OCN:ICN:IPN Example ........................................................................................................................... 144

Release 4.0 April 2023 14

Universal Serial Bus Device Class Definition for Audio Devices

Figure 7-1: Type I Audio Stream .............................................................................................................................. 197

Figure 7-2: Extended Type I Format ........................................................................................................................ 205
Figure 7-3: Extended Type III Format ...................................................................................................................... 205
Figure 8-1: Standard Interface Descriptor Set Layout ............................................................................................ 210
Figure 8-2: ADC 2.0 Audio Function Descriptor Set ................................................................................................ 211
Figure 8-3: ADC 4.0 Audio Function Descriptor Set ................................................................................................ 212

Release 4.0 April 2023 15

Universal Serial Bus Device Class Definition for Audio Devices

1 INTRODUCTION

The following sections provide a brief overview of the scope and purpose of this document. It also includes a list of
related documents.

1.1 SCOPE
The Audio Device Class Definition applies to all Devices or Functions embedded in composite Devices that are used
to manipulate audio, voice, and sound-related functionality. This includes both audio data (analog and digital) and
the functionality that is used to directly manipulate audio signals, such as Gain and Tone Control. The Audio Device
Class does not include functionality to operate transport mechanisms that are related to the reproduction of audio
data, such as tape transport mechanisms or CD-ROM drive control, nor does it include how HID features (volume
up/down, play/pause, etc.) are conveyed. See Universal Serial Bus Device Class Definition for Human Interface
Devices (HID) for more related information.

1.2 PURPOSE
The purpose of this document is to describe the minimum capabilities and characteristics an Audio Function shall
support to comply with the USB Audio Device Class. This document also provides recommendations for optional
features.

1.3 RELATED DOCUMENTS

• USB Core Specifications:

• Universal Serial Bus 2.0 Specification, Revision 2.0 (referred to in this document as the USB 2
Specification). See Chapter 5, “USB Data Flow Model” and Chapter 9, “USB Device Framework.”
• Universal Serial Bus 3.2 Specification, Revision 1.0 (referred to in this document as the USB 3
Specification). This document covers details specific to SuperSpeed and SuperSpeed+ Devices.
• Universal Serial Bus 4.0 Specification, Version 2.0. This document covers details specific to Gen T Devices.
• Universal Serial Bus Device Class Definition for Human Interface Devices (HID), Version 1.11.
• ANSI S1.11-1986 standard.
• MPEG-1 standard ISO/IEC 111172-3 1993. (available from http://www.iso.ch )
• MPEG-2 standard ISO/IEC 13818-3 Feb. 20, 1997. (available from http://www.iso.ch)
• Windows Media Audio (WMA) specification. (available from http://www.microsoft.com)
• Digital Audio Compression Standard (AC-3), ATSC A/52A Aug. 20, 2001. (available from http://www.atsc.org/)
• ANSI/IEEE-754 standard. IEEE Standard for Floating-point Arithmetic.
• ISO/IEC 60958 International Standard: Digital Audio Interface and Annexes.
• ISO/IEC 61937 standard. Interface for non-linear PCM Encoded audio bitstreams applying IEC 60958
• ISO/IEC 80000-13 standard. Quantities and Units.
• ITU G.711 standard. Pulse code modulation (PCM) of voice frequencies.
• ETSI Specification TS 102 114, “DTS Coherent Acoustics; Core and Extensions”. (Available from
• EN 50332-2:2013 specification. Available from several sources. (https://www.en-standard.eu/ )
• EN 50332-2:2017 specification. Available from several sources. (https://www.en-standard.eu/ )
• IETF RFC 4122 GUID definition: Available from https://www.ietf.org/rfc/rfc4122.txt
• HDCP 2.3 HDCP Specifications: Available from https://www.digital-cp.com/hdcp-specifications.
• IEEE Std 269-2019. "IEEE Standard for Measuring Electroacoustic Performance of Communication Devices”.

Release 4.0 April 2023 16

Universal Serial Bus Device Class Definition for Audio Devices

1.4 TERMS AND ABBREVIATIONS

This section defines terms used throughout this document. For additional terms that pertain to the Universal Serial
Bus, see Chapter 2, “Terms and Abbreviations,” in the USB Core Specifications.

AC-3 Audio compression standard from Dolby Labs.

AEC Acronym for Audio Echo Cancellation.

Armed AudioContol An AudioControl that has its NEXT Attribute preloaded with a valid
value.

ASRC Acronym for Asynchronous Sampling Rate Converter.

Audio data stream Transport medium that carries audio information.

Audio Function Independent part of a USB Device that deals with audio-related
functionality.

Audio Function Descriptor Set (AFDS) The collection of the Interface Association Descriptor and the full
set of Standard Descriptors and class-specific Descriptors that
together comprise the entire Audio Function.

Audio Interface Association (AIA) Grouping of a single AudioControl Interface along with zero or
more AudioStreaming Interfaces that together constitute a
complete interface to a particular version (compliance level) of the
Audio Function.

Audio Slot A collection of Audio Subslots, each containing a PCM audio sample
of a different physical audio channel, taken at the same moment in
time.

Audio Stream A continuous sequence of advancing, time-ordered audio slots.

Typically, a set of one or more channels that have a common rate,
direction, and phase.

Audio Subslot Holds a single PCM audio sample.

AudioControl Logical object that is used to manipulate a specific audio property.

Examples are Gain Control, Mute Control, etc.
AudioControl Attribute Parameter of an AudioControl. Examples are Current, Next,
Minimum, Maximum and Resolution Attributes of a Gain Control.

AudioControl Interface USB interface used to access the AudioControls inside an Audio
Function.

AudioStreaming Interface USB interface used to control the USB transport of audio streams
into or out of the Audio Function.

Bus Interval Time interval between two consecutive SOF tokens while the Bus is
in L0. This is not the same as the Service Interval.

byte An 8-bit quantity. Byte fields start with the letter b (or ba for
arrays), followed by the field name.

Release 4.0 April 2023 17

Universal Serial Bus Device Class Definition for Audio Devices

ChannelGroup A grouping of audio channels inside of a Cluster that carry tightly

related audio.

Clock Domain A zone within the Audio Function that is served by sampling clocks
that are all derived from the same Reference Clock.

Clock Selector (CX) Selects from several input clock signals.

Clock Source (CS) Generates an audio-related clock signal (Main clock or sample
clock).

Cluster A grouping of audio channels that carry tightly related audio,

streaming over a connection inside the Audio Function. An
instantiation of a Cluster Configuration.

Active Cluster A grouping of audio channels that currently carries valid audio.

Inactive Cluster A grouping of audio channels that currently does not carry valid
audio.

Incoming Cluster The Cluster entering an Entity on a single Input Pin. Note that some
Entities have multiple Input Pins and therefore have multiple
incoming Clusters.

Outgoing Cluster The Cluster leaving an Entity on its single Output Pin.

Cluster Configuration The definition of the Cluster as detailed in the Cluster Descriptor.

Cluster Descriptor An Extended Descriptor describing the characteristics of a Cluster,

such as the number of channels, Purpose, and Relationship of each
channel.

Connector Entity Provides access to the AudioControl(s) that impact the behavior of
the Connector.

Device Term used to describe a device that has USB functionality.

DTS Acronym for Digital Theater Systems.

DVD Acronym for Digital Versatile Disc.

dword A 4-byte quantity. Dword fields start with the letter d (or da for
arrays), followed by the field name.

Effect Unit (EU) Provides advanced audio manipulation on the incoming logical
audio channels.

Encoded Audio Bit Stream A continuous sequence of advancing, time-ordered encoded audio
frames.

Encoded Audio Frame A sequence of bits that contains an encoded representation of

audio samples from one or more physical audio channels taken
over a fixed period.

Entity Addressable logical object inside an Audio Function.

Release 4.0 April 2023 18

Universal Serial Bus Device Class Definition for Audio Devices

EntityGroup A logical grouping of Entities that are associated and have a

relationship with one another.

Extended Descriptor A new representation for class-specific Descriptors that allows for
potentially large (>255 bytes) and dynamic Descriptors.

Extension Unit (XU) Applies an undefined process to a set of logical input channels.

Feature Unit (FU) Provides basic audio manipulation on the incoming logical audio
channels.

IAD Acronym for Interface Association Descriptor.

ICN Acronym for Input Channel Number. The ordinal number of an

audio channel in a Cluster on an Input Pin of an Entity.

Input Pin Logical input connection to an Entity. Carries a single Cluster.

Input Terminal (IT) Receptacle for audio information flowing into the Audio Function.

Interface Association A group of two or more Interfaces associated with the same Audio
Function.

IPN Acronym for Input Pin Number. The ordinal number of an Input Pin
of an Entity.

kbits kilobits. Equals 1000 bits per ISO/IEC 80000-13 standard.

KiB kibibyte. Equals 1024 bytes per ISO/IEC 80000-13 standard.

Logical Audio Channel Logical transport medium for a single audio channel. Makes
abstraction of the physical properties and formats of the
connection. Is usually identified by spatial location. Examples are
Front Left channel, Surround Array Right channel, etc.

LSb Least Significant bit.

Mixer Unit (MU) Mixes one or more logical input channels into one or more logical
output channels.

MPEG Acronym for Moving Pictures Expert Group.

MSb Most Significant bit.

Node A location within an Entity’s internal topology, identified by an

OCN:ICN:IPN Triplet.

OCN Acronym for Output Channel Number. The ordinal number of an

audio channel in a Cluster on the Output Pin of an Entity.

OCN:ICN:IPN Triplet The Output Channel Number, Input Channel Number, and Input Pin
Number combination that identifies a specific AudioControl within
an Entity’s internal topology.

Output Pin Logical output connection from an Entity. Carries a single Cluster.

Output Terminal (OT) An outlet for audio information flowing out of the Audio Function.

Release 4.0 April 2023 19

Universal Serial Bus Device Class Definition for Audio Devices

PCM Acronym for Pulse Coded Modulation.

Pin Channel A physical channel a Unit or Terminal supports/implements on an

Input or Output Pin.

Pin Channel Count (PCC) The maximum number of physical channels a Unit or Terminal
supports/implements on an Input or Output Pin.

Phase Locked Loop A hardware or software control system that generates an output
signal whose phase is related to the phase of an input signal.
Typically used in clock recovery applications.

Power Domain A grouping of one or more Entities whose power states are
simultaneously controlled by a single Power Domain Entity.

Power Domain Entity Provides access to the AudioControl(s) that impact the behavior of
the Power Domain.

Processing Unit (PU) Applies a predefined process to one or more logical input channels.

Reference Clock The single clock generator that is used to derive all clocks in a Clock
Domain.

Reserved “Reserved” is a keyword indicating reserved bits, bytes, words,

fields, and code values that are set-aside for future
standardization. Their use and interpretation may be specified by
future extensions to this specification and, unless otherwise stated,
shall not be utilized, or adapted by vendor implementation. A
reserved bit, byte, word, or field shall be set to zero by the sender
and shall be ignored by the receiver.

Request Error The Device returning a STALL PID when an error is detected in the
content of the Request. See Section 9.2.7, “Request Error”, of the
Universal Serial Bus Specification Revision 2.0 for more details.

Sampling Rate Converter Unit (RU) Converts the incoming audio data stream, running at a sampling
rate that is synchronous to a first Main clock, into an outgoing
audio data stream that is running at a sampling rate that is
synchronous to a second Main clock, which is either free running
with respect to the first Main clock (asynchronous sampling rate
conversion) or synchronous to the first Main clock (synchronous
sampling rate conversion).

Selector Unit (SU) Selects from several input Clusters.

Service Interval The period between consecutive transfers on a periodic Endpoint,

specified by the value in the bInterval field of the Endpoint
Descriptor. This is not the same as the Bus Interval.

Service Interval Packet A packet that contains all the audio slots that are transferred over
the bus during a Service Interval.

Terminal Addressable logical object inside an Audio Function that represents

a connection to the Audio Function’s outside world.

Release 4.0 April 2023 20

Universal Serial Bus Device Class Definition for Audio Devices

Unit Addressable logical object inside an Audio Function that represents

a certain audio sub-functionality.

USB-C® A 24-pin USB connector system.

WMA Acronym for Windows Media Audio.

word A 2-byte quantity. Word fields start with the letter w (or wa for
arrays), followed by the field name.

Release 4.0 April 2023 21

Universal Serial Bus Device Class Definition for Audio Devices

2 MANAGEMENT OVERVIEW

The USB is very well suited for transport of audio, ranging from low fidelity voice connections to high quality, multi-
channel audio streams. The USB has become a ubiquitous connector on modern computers and is well-understood
by most consumers today. As such, it has become the connector of choice for many peripherals and is indeed the
simplest and most pervasive digital audio connector available today. Consumers can count on this medium to meet
all of their current and future audio needs. Many applications from communications, to entertainment, to music
recording and playback, can take advantage of the audio features of the USB.

In principle, a versatile bus specification like the USB provides many ways to propagate and/or control audio
signals. For the industry, however, it is very important that audio transport mechanisms be well-defined and
standardized on the USB. Only in this way can interoperability be guaranteed among the many possible Audio
Devices on the USB. Standardized audio transport mechanisms also help keep software drivers as generic as
possible. The Audio Device Class described in this document satisfies those requirements. It is written and revised
by experts in the audio field. Other Device Classes that address audio in some way should refer to this document
for their audio interface specification.

An essential issue in audio is synchronization of the data streams. Indeed, the smallest artifacts are easily detected
by the human ear. Therefore, a robust synchronization scheme on isochronous transfers has been developed and
incorporated in the USB Core Specifications. The Audio Device Class definition adheres to these synchronization
schemes to transport audio data reliably over the bus.

This document contains all necessary information for a designer to build a USB-compliant Device that incorporates
Audio functionality. It specifies the standard and class-specific Descriptors that shall be present in each USB Audio
Function. It further explains the use of class-specific Commands that allow for full Audio Function control. Several
predefined data formats are listed and fully documented. Each format defines a standard way of transporting
audio over the USB.

Many of the changes introduced in this version of the USB Specification for Audio Devices are inspired by the
desire to use USB Audio in modern portable devices. Special attention has been paid to make the Audio Device
Class more power-friendly by providing new tools to selectively enable and disable parts of the Audio Function and
by supporting burst mode data transfers for longer sleep times in between data transfers. In addition, the
specification supports new CODEC types and data formats for consumer audio applications, provides numerous
clarifications of the original specification and extensions to support various changes in the core specification.

Inherent restrictions that were present in previous versions have been alleviated. For example, it was impossible to
have a Cluster with more than 61 channels on the Input Pin of a Feature Unit due to the limited descriptor space (<
256 bytes) available to describe per-channel AudioControls.

This ADC 4.0 specification supersedes the ADC 3.0 version. It is highly recommended to design devices using this
latest 4.0 version as the previous version imposed some restrictions (especially in the realm of backwards
compatibility) that made it cumbersome to implement. The use of the ADC 3.0 version is therefore highly
discouraged.

2.1 OVERVIEW OF KEY DIFFERENCES BETWEEN ADC 2.0 AND ADC 4.0
The following list is not an exhaustive list of all changes that have been introduced. For complete information,
refer to the full specification.

• Dolby Processing Unit removed

• Encoder and Decoder support removed

Release 4.0 April 2023 22

Universal Serial Bus Device Class Definition for Audio Devices

• Copy protection (S/PDIF-style) removed

• bmAttributes (MaxPacketsOnly bit) field in class-specific isochronous Endpoint Descriptor removed
• Type II and Extended Type II and Type IV Audio Data Formats removed
• Additional Type III formats introduced
• New Power Domains
• New Multi-Function Processing Unit
• New Cluster Descriptor
• New Terminal Companion Descriptor (previously named Extended Terminal Descriptor in ADC 3.0)
• New Extended Descriptors (previously named High Capability Descriptors in ADC 3.0)
• Introduced a new backwards compatibility mechanism
• New CAP and NEXT Attribute for AudioControls
• New Commit Control
• New ChannelGroups and EntityGroups
• Modified the meaning of the bm(a)Controls field and renamed to bm(a)OptControls
• Selector Units now support an off position
• Several Descriptors have a slightly different layout and some added fields
• Active and Inactive Clusters
• Clarified difference between Cluster channels and Pin channels
• AudioStreaming Interfaces now exclusively used for USB Streaming Interfaces
• Removed bAssocTerminal from Terminal Descriptor. Replaced with EntityGroup concept
• Removed Terminal Types
• Converted most class-specific Descriptors from Traditional Descriptors to Extended Descriptors
• Added Channel ID to enable channel traceability throughout the Audio Function
• EN 50332-2 and EN 50332-3 support
• Added Channel Remap PU
• Optional support for LPM/L1 and clarified the concept
• Added StartDelay concept to AudioStreaming Interfaces
• Removed LockDelay concept from endpoint descriptors. Replaced by the StartDelay concept
• Eliminated class-specific Endpoint Descriptors
• Removed Overrun, Underrun, Pitch Controls
• Additional sources for interrupts
• New class-specific String Descriptors
• Added Connector Entities
• Removed the RAW_DATA concept
• New Control Command structure to overcome the limitations in allowed parameter bit widths in the
current Control request structure
• Removed the MEM Attribute and its associated infrastructure
• Redefined the Clock Source Entity
• Removed the Input Gain Control
• Renamed the Volume Control to Gain Control

Release 4.0 April 2023 23

Universal Serial Bus Device Class Definition for Audio Devices

3 FUNCTIONAL CHARACTERISTICS
The following sections describe the functional characteristics of a USB Audio Function.

3.1 INTRODUCTION
In many cases, Audio functionality is incorporated with other USB Class functionality within a single (composite)
Device. The Audio Function is thus located at the interface level in the Device Class hierarchy. The following figure
provides details.
Figure 3-1: Audio Interface Associations and Interfaces

Device Configuration

Audio Function
AudioControl + AudioStreaming

Audio Function
AudioControl + MIDIStreaming

Function P
(Ex. CDC)

Function R
(Ex: HID)

Audio Interface Association Audio Interface Association

AudioControl I/F AudioControl I/F

AudioStreaming I/F MIDIStreaming I/F

AudioStreaming I/F MIDIStreaming I/F

An Audio Function is considered to be a ‘closed box’ that has very distinct and well-defined interfaces to the
outside world. Audio Functions are described through an Audio Interface Association (AIA). The AIA groups all USB
interfaces that together provide access to the Audio Function for control and streaming purposes.
An AIA shall have a single AudioControl Interface and may have zero or more AudioStreaming Interfaces. The
AudioControl (AC) Interface is used to access the AudioControls inside the Audio Function whereas the
AudioStreaming (AS) Interfaces are used to encapsulate their associated USB Endpoint(s) and provide access to
AudioControls that impact the transport of USB audio streams into and out of the Audio Function.

The following figure illustrates the concept of an Audio Function and its associated interfaces:

Release 4.0 April 2023 24

Universal Serial Bus Device Class Definition for Audio Devices

Figure 3-2: Example Audio Function Global View

A Device may have multiple independent Audio Functions located in the same composite Device. They are each
accessed through their own Audio Interface Association.

All functionality pertaining to controlling parameters that directly influence audio perception (like volume) are
located inside the central rectangle and are exclusively controlled through the AudioControl Interface. Streaming
aspects of the communication to or from the Audio Function are handled through separate AudioStreaming
Interfaces, as necessary. Each USB audio stream shall be represented by an AudioStreaming Interface. All control
information that is related specifically to the streaming behavior of the USB interface is conveyed through the
AudioStreaming Interface.
Also note that the connection between the AudioStreaming Interfaces and the Audio Function is not ‘solid’. The
reason for this is that when seen from the inside of the Audio Function, each audio stream entering or leaving the
Audio Function is represented by a special object, called a Terminal (see further). The Terminal concept abstracts
the actual AudioStreaming Interface inside the Audio Function and provides a logical view on the connection
rather than a physical view. This abstraction allows audio channels within the Audio Function to be treated as
‘logical’ audio channels that do not have physical characteristics associated with them anymore (analog vs. digital,
format, sampling rate, bit resolution, etc.).

Non-USB audio streams, such as an S/PDIF connection, shall never be represented by an AudioStreaming Interface.

3.2 BASIC AUDIO DEVICE EXAMPLES

The Basic Audio Device Definition documents that were introduced in ADC 3.0 have been deprecated and are
replaced with example documents that outline for each of the previous Basic Audio Device Definition Device Types

Release 4.0 April 2023 25

Universal Serial Bus Device Class Definition for Audio Devices

how they should be implemented using the concepts defined in this ADC 4.0 specification. The appropriate
Descriptor Sets for each are included and should provide implementers with a ready-to-use implementer’s guide.

3.3 BACKWARDS COMPATIBILITY

This 4.0 version of the Audio Device Class Definition (ADC 4.0) as such is not backwards compatible with any of the
previous versions of the same specification. However, a method to provide the broadest interoperability with
existing ecosystems has been incorporated into this specification and is described in detail in Section 8,
“Backwards Compatibility Considerations.”

3.4 AUDIO INTERFACE ASSOCIATION (AIA) AND INTERFACE ASSOCIATION

DESCRIPTOR
An Audio Function is completely defined by its interfaces. The standard USB Interface Association mechanism is
used to describe the Audio Interface Association, i.e. binding those interfaces together. An Audio Function shall
have exactly one Audio Interface Association (AIA). Each AIA shall have one AudioControl Interface and may have
zero or more AudioStreaming Interfaces. All Audio Function interfaces shall be part of an Audio Interface
Association, even if the Audio Interface Association contains only a single AudioControl Interface.

Interface Association is expressed via the standard USB Interface Association Descriptor (IAD). Every Interface
Association Descriptor has a bFunctionClass, bFunctionSubClass and bFunctionProtocol field that together identify
the function that is represented by the Association. The following paragraphs define these fields for the Audio
Device Class.

3.4.1 AUDIO FUNCTION CLASS

The Audio Function Class is contained in the bFunctionClass field of a standard Interface Association Descriptor.
This specification requires that the Function Class code be the same as the Audio Interface Class code.

The Audio Interface Class code and therefore the Audio Function Class code is assigned by the USB-IF. The assigned
codes can be found in Appendix A.2, “Audio Function Class Code.”

3.4.2 AUDIO FUNCTION SUBCLASS

The Audio Function Subclass is contained in the bFunctionSubClass field of a standard Interface Association
Descriptor. This field is not used by this specification and shall be set to FUNCTION_SUBCLASS_UNDEFINED.

The assigned code can be found in Appendix A.3, “Audio Function Subclass Codes.” of this specification. All other
Subclass codes are unused and reserved by this specification for future use.

3.4.3 AUDIO FUNCTION PROTOCOL

The Audio Function Protocol is contained in the bFunctionProtocol field of a standard Interface Association
Descriptor. The Function Protocol code is used to reflect the compliance level of the Audio Function so that
enumeration software can decide which driver version needs to be instantiated. This specification requires that
the Function Protocol code be AF_VERSION_04_00.

The assigned Protocol codes can be found in Appendix A.4, “Audio Function Protocol Codes” of this specification.
All other Protocol codes are unused and reserved by this specification for future use.

Release 4.0 April 2023 26

Universal Serial Bus Device Class Definition for Audio Devices

3.5 AUDIO INTERFACE CLASS

The Audio Interface Class code is contained in the bInterfaceClass field of the standard interface Descriptor of any
Interface in an AIA. All USB Interfaces which provide functionality through this specification use the same Audio
Interface Class code.

The Audio Interface class code is assigned by the USB. The assigned code can be found in Appendix A.5, “Audio
Interface Class Code”.

3.6 AUDIO INTERFACE SUBCLASS

The Audio Interface Subclass code is contained in the bInterfaceSubClass field of the standard interface Descriptor
of any Interface in an AIA. All USB interfaces that provide functionality through this specification shall use one of
the following Audio Interface Subclass codes as provided by this specification:

• AUDIOCONTROL Interface Subclass

• AUDIOSTREAMING Interface Subclass

The assigned codes can be found in Appendix A.6, “Audio Interface Subclass Codes” of this specification. All other
Subclass codes are unused and reserved by this specification for future use.

3.7 AUDIO INTERFACE PROTOCOL

The Audio Interface Protocol code is contained in the bInterfaceProtocol field of the standard interface Descriptor
of any Interface in an AIA. The Interface Protocol code is used to reflect the compliance level of the Audio
Function.

All Audio Functions compliant with this ADC 4.0 specification shall use Interface Protocol Code IP_VERSION_04_00.
The assigned codes can be found in Appendix A.7, “Audio Interface Protocol Codes” of this specification. All other
Protocol codes are unused and reserved by this specification for future use.

3.8 CLOCK DOMAINS

A Clock Domain is defined as a zone within which all sampling clocks are derived from the same Reference clock.
Therefore, within the same Clock Domain, all sampling clocks are frequency-locked, and their timing relationship is
constant. However, the sampling clocks may be at different sampling frequencies, but related exactly by the ratio
of two integers. The Reference Clock may be generated in many ways. An internal crystal could be used as the
Reference Clock; the USB start of frame (SOF) or even an externally supplied clock could serve as the Reference
Clock. Different sampling clocks within the same Clock Domain may then be derived from the Reference Clock
through a combined use of Clock Source and Clock Selector Entities.

Multiple different Clock Domains may exist within the same Audio Function. Clock Domains are described via a
topology map, as part of the various Clock Source Descriptors that are members of the Clock Domain.

3.9 POWER DOMAINS

A Power Domain is defined as a zone within the Audio Function that groups one or more Entities and allows the
Host to control power consumption levels for that Domain. This way, the Host can potentially switch parts of the
Audio Function to lower power states when these parts are currently not in use, leading to an overall decrease in
power consumption. As an example, a USB headset may have two separate Power Domains, one for the output-
related functionality (headphones) and one for the input-related functionality (microphones). When the headset is
used to simply listen to music, the input-related functionality may be temporarily switched to a low power state

Release 4.0 April 2023 27

Universal Serial Bus Device Class Definition for Audio Devices

(or even completely off) to conserve power. This is especially important if the headset is used in conjunction with a
battery-powered device, such as a mobile phone, where extending battery life is important.

Multiple different Power Domains may exist within the same Audio Function. Power Domains are identified by a
Power Domain Entity with a unique Power Domain ID within the Audio Function.

3.10 GROUPS
Three types of Groups are defined by this specification:

• ChannelGroup: Part of a Cluster. A subset of channels in the Cluster that are logically related. A ChannelGroup
shall have only channels as its Members.
• EntityGroup: A collection of Entities that the Audio Function wishes to advertise as functionally or physically
related. An EntityGroup shall have only Entities as its Members.
• CommitGroup: A collection of AudioControls, potentially from multiple Units, that the Audio Function marks
for synchronous updates, using the Commit Capability.

Group nesting (of any type) is prohibited: i.e., a Group shall never be a Member of another Group.

Multiple different Groups may exist within the same Audio Function.

3.11 AUDIO SYNCHRONIZATION TYPES

Each isochronous audio Endpoint used in an AudioStreaming Interface belongs to a synchronization type as
defined in Section 5 of the USB Core Specification. The following sections briefly describe the possible
synchronization types.

3.11.1 ASYNCHRONOUS
Asynchronous isochronous audio Endpoints produce or consume data at a rate that is locked either to a clock
external to the USB or to a free-running internal clock. These Endpoints are not synchronized to a start of frame
(SOF).

3.11.2 SYNCHRONOUS
The clock system of synchronous isochronous audio Endpoints can be controlled externally through SOF or Bus
Interval synchronization. Such an Endpoint shall lock its sample clock to the SOF tick marking the start of a new Bus
Interval.

3.11.3 ADAPTIVE
Adaptive isochronous audio Endpoints can source or sink data at any rate within their operating range. This implies
that these Endpoints shall run an internal process that allows them to match their natural data rate to the data
rate that is imposed at their interface.

3.11.4 IMPLICATIONS OF THE DIFFERENT SYNCHRONIZATION TYPES

The following sections provide some information on the impact the different synchronization types have on both
the Device and Host implementation. Several scenarios are considered and for each synchronization type, the
implications on Host and Device side are listed.

Release 4.0 April 2023 28

Universal Serial Bus Device Class Definition for Audio Devices

3.11.4.1 SINGLE DIRECTION SINK ENDPOINT

A typical example for this scenario is a USB speaker, implementing a sink Endpoint that receives streaming audio
data from the Host.

• Asynchronous
o Host: The Host driver needs to be able to handle an explicit feedback Endpoint. From the
feedback data, the Host then decides how many samples to send over the data streaming
Endpoint in subsequent Service Intervals.

Note: The expected value returned by the feedback Endpoint is the number of full and partial
samples expected on average for every Service Interval expressed by the bInterval field
of the Isochronous data Endpoint. See the description of feedback Endpoints in the
appropriate USB core specification for more information about the format of feedback
Endpoints.

o Device: The Device has its own local, free-running audio sample clock, which determines how
many samples are consumed by the Device each Service Interval. The Device shall implement an
explicit feedback Endpoint as well as the necessary logic to provide the correct feedback values
to send over said Endpoint back to the Host. The advantage of this mode of operation is that it is
rather easy to generate a robust, stable, low-jitter, high-quality audio sample clock (derived from
a crystal-based Main clock, for example).

Note: The expected value returned by the feedback Endpoint is the number of full and partial
samples expected on average for every Service Interval expressed by the bInterval field
of the Isochronous data Endpoint. See the description of feedback Endpoints in the
appropriate USB core specification for more information about the format of feedback
Endpoints.

• Synchronous

o Host: The Host needs to send out a known number of bytes for each packet going to the Device.
The Host may need to generate a (fixed) pattern of audio samples to achieve the desired
sampling rate. As an example, to generate a sampling rate of 44.1 kHz in a Full-Speed
implementation, the Host needs to send a repeating pattern of nine packets containing 44 audio
samples, followed by one packet containing 45 audio samples.

o Device: This synchronization type requires the Device to implement either an audio clock PLL or
an ASRC.

• Adaptive

o Host: The Host may use any method or means to determine how many samples per Service
Interval to transmit. Effectively operating as a “Synchronous-to-SOF” Source is an easy approach,
but not the only one allowed by the USB core specification.

o Device: This synchronization type requires the Device to implement either an audio clock PLL or
an ASRC to adapt to the average number of samples arriving over a certain period.

3.11.4.2 SINGLE DIRECTION SOURCE ENDPOINT

A typical example for this scenario is a USB microphone, implementing a source Endpoint that sends streaming
audio data to the Host.

Release 4.0 April 2023 29

Universal Serial Bus Device Class Definition for Audio Devices

• Asynchronous

o Host: The Host needs to operate as an Adaptive Sink. Depending on the Host’s system design,
this may require an ASRC or an audio clock PLL on the Host side.

o Device: The Device has its own local, free-running audio sample clock, which determines how
many samples are produced by the Device each Service Interval. The advantage of this mode of
operation is that it is rather easy to generate a robust, stable, low-jitter, high-quality audio
sample clock (derived from a crystal-based Main clock, for example).

• Synchronous

o Host: The Host receives a known number of audio samples in each packet coming from the
Device. The Host should expect a fixed pattern of audio samples to achieve the desired sampling
rate.
o Device: This synchronization type requires the Device to implement either an audio clock PLL to
lock on to the SOF or the start of Bus Interval and generate a high-quality audio sample clock
directly or use an ASRC as a bridge between the local and USB Clock Domains. The Device shall
generate a fixed packet size pattern as described in Section 7, “Audio Data Formats.”

• Adaptive (not recommended for implementation)

o Host: This scenario requires the implementation of a feedforward OUT Endpoint in the Device
that allows the Host to inform the Device how many samples per Service Interval to send to the
Host over the streaming data IN Endpoint.

Note: Declaring a source endpoint as Adaptive without providing the feedforward OUT
endpoint is a violation of the USB core specification.

o Device: This synchronization type requires the Device to implement an audio clock PLL or an
ASRC to adapt to the sample rate being communicated by the Host via the feedforward OUT
Endpoint.

3.11.4.3 DUAL DIRECTION SOURCE AND SINK ENDPOINT PAIR

A typical example for this scenario is a USB headset, implementing both a sink Endpoint that receives streaming
audio data from the Host and a source Endpoint that sends streaming audio data to the Host.
In general, the Endpoints may be treated independently, using the same rules and guidelines as stated above. This
includes the ability for the Device to support separate clocks on each Endpoint.

However, if the Device has its own clock and both data streams share this clock, the following two special cases
can be used to allow implicit feedback:

• Both Endpoints are Asynchronous. In this case, the data rate that appears on the Source Endpoint can be
used by the Host as implicit feedback to adjust the data rate transmitted to the Sink Endpoint.

• Both Endpoints are Adaptive. In this case, the data rate that appears on the Sink Endpoint can be used by
the Device to adjust the data rate transmitted by the Source Endpoint.

Note: Using either of these two implicit feedback mechanisms would preclude the ability of an ADC 4.0
compliant Device from being able to use Power Domains to shut down either the source or the sink,
as data shall remain flowing in both directions to provide the feedback/feedforward information.

Release 4.0 April 2023 30

Universal Serial Bus Device Class Definition for Audio Devices

3.12 INTER CHANNEL SYNCHRONIZATION

An important issue when dealing with audio, and 3-D audio in particular, is the phase relationship between
different physical audio channels. Indeed, the virtual spatial position of an audio source is directly related to and
influenced by the phase differences that are applied to the different physical audio channels used to reproduce the
audio source. Therefore, it is imperative that USB Audio Functions respect the phase relationship among all related
audio channels. However, the responsibility for maintaining the phase relation is shared among the USB host
software, hardware, and all the audio peripheral Devices or functions.

3.13 AUDIO FUNCTION TOPOLOGY

To be able to manipulate the physical properties of an Audio Function, its functionality is divided into addressable
Entities where each Entity shall have a unique non-zero Entity ID. The AudioControl Interface encapsulates these
Entities and is itself addressed via its interface number and Entity ID zero. AudioStreaming Interfaces are also
addressed through their interface number and Entity ID zero.

This specification defines several different Entity types:

• Units
• Terminals
• Clock Entities
• Connector Entities
• Power Domain Entities

Units provide the basic building blocks to fully describe the internals of most Audio Functions. Audio Functions are
built by connecting several of these Units. A Unit has one or more Input Pins and a single Output Pin, where each
Pin carries a Cluster of logical audio channels (see Section 3.13.2, “Cluster”).

Units are wired together by connecting their I/O Pins according to the required topology. Note that it is perfectly
legal to connect the Output Pin of an Entity to multiple Input Pins residing on different other Entities, effectively
creating a one-to-many connection.

In addition to Units, Terminals are defined as well. There are two types of Terminals. The Input Terminal (IT) is an
Entity that represents a starting point for audio channels inside the Audio Function. The Output Terminal (OT)
represents an ending point for audio channels inside the Audio Function. From the Audio Function’s perspective, a
USB streaming Endpoint is a typical example of an audio source or sink that is represented by an Input or Output
Terminal. It either provides USB data streams to the Audio Function (IT) or consumes data streams coming from
the Audio Function (OT). Likewise, a Digital to Analog converter, built into the Audio Function is represented as an
Output Terminal in the Audio Function’s model.

The Input Terminal connects into the Audio Function through the Input Terminal’s single Output Pin. The Audio
Function connects to the Output Terminal through the Output Terminal’s single Input Pin.

Input Pins of a Unit are numbered starting from one up to the total number of Input Pins on the Unit. The single
Output Pin number is always one.

Input Pin one is somewhat special in that it is considered the ‘dominant’ Input Pin on most Units. This means that
for both single input and multi-input Units that have a Unit Bypass function, the Cluster that enters the Unit on
Input Pin one shall be passed unaltered to the Output Pin whenever the Unit Bypass function is engaged. It is
always possible to override this standard behavior by not implementing the optional Bypass Control and instead

Release 4.0 April 2023 31

Universal Serial Bus Device Class Definition for Audio Devices

provide a bypassing Selector Unit. When an Entity has both a Cluster Control and a Bypass Control, the Bypass
Control takes precedence when engaged.

Input Terminals have only one Output Pin and its number is always one. Output Terminals have only one Input Pin
and it is always numbered one.

The Clusters traveling over I/O Pins and their interconnects are not necessarily of a digital nature. It is perfectly
possible to use the Unit model to describe fully analog or even hybrid Audio Functions. The mere fact that I/O Pins
are connected is a guarantee (by construction) that the protocol and format, used over these connections (analog
or digital), is compatible on both ends.

Also, depending on certain Audio Function settings, such as some parts of the Audio Function going to a low power
state, or a Selector Unit’s Selector Control set to the no-connect (n.c.) position, or the absence of valid clock signal
on a Terminal, etc., it is possible that an Entity’s Output Pin does not carry a valid audio signal, i.e. the Cluster on
that Output Pin is Inactive (see Section 3.13.2, “Cluster”). In that case, any Entity that receives this Inactive Cluster
on one of its Input Pins shall disregard any signaling on that Input Pin and internally drive Silence on all its physical
channels it is designed to support on that Input Pin (irrespective of the advertised number of logical channels in
the Inactive Cluster).

Clusters may change dynamically over time. For example, a Selector Unit that has two Input Pins where Input Pin
one receives a two-channel Cluster and Input Pin two receives a six-channel Cluster, then the Selector Unit’s
Output Pin will switch between a two-channel an a six-channel Cluster whenever the Unit’s Selector Control is
flipped between position one and two.

An Audio Function is allowed to autonomously switch the Cluster on any of its internal connections from one of
the supported Cluster Configurations to another supported Cluster Configuration on that node. However, if an
Output Terminal that is associated with an AudioStreaming Interface experiences a change in Cluster Configuration
at its Input Pin or that Cluster becomes Inactive, then that AudioStreaming Interface shall switch to Alternate
Setting zero.

The effect on the behavior of a physical output interface when a Cluster Configuration change occurs at the Input
Pin of the Output Terminal that represents that interface is implementation dependent.

Units and Terminals by design have a maximum number of channels they support on any of their I/O Pins. This
number is called the Pin Channel Count or PCC of an I/O Pin. The channels themselves are called Pin channels to
distinguish them from the logical channels in the Cluster that travels over the Pin.
For all connections within the Audio Function, the PCC of the Output Pin of any Unit or Terminal shall match the
PCC of every Input Pin to which that Output Pin is connected. Rather than advertising the (identical) PCC for each
individual Pin that participates in a connection, the PCC is advertised in an Entity’s Descriptor as a parameter
associated with its Output Pin (the single source of the connection).

At any given time, the actual Cluster flowing over a connection between an Output Pin and the connected Input
Pin(s) shall never have more logical channels than indicated by the PCC value.

Note: It is allowed for implementations to advertise a PCC value that is larger than strictly needed to
support each of the available Clusters on a connection, i.e., the advertised PCC value may be larger
than the maximum of the number of channels in each of the available Clusters.

Internally, Units and Terminals inherit the PCC on each of their Input Pins from the Entity’s Output Pin to which the
Input Pin is connected, and the number of Pin channels on each Input Pin is determined by that number. The

Release 4.0 April 2023 32

Universal Serial Bus Device Class Definition for Audio Devices

number of Output Pin channels is determined directly by the PCC value advertised for that Output Pin. The
following figure illustrates this.
Figure 3-3: General PCC Inheritance Rules

For single-input Units that do not alter the number of Pin channels between their Input and Output Pin, the Unit
does not advertise the Output Pin PCC value. Rather, the PCC value is inherited from upstream and determined by
the first Entity upstream that defines a PCC value on its Output Pin as illustrated below.
Figure 3-4: Single Input Pin PCC Inheritance Rules

Note that in some cases multiple single-input Units may need to be traversed before finding the upstream Entity
that defines a PCC value on its Output Pin.

Any signaling on currently unused Pin channels on an Input Pin shall be disregarded and the Unit or Output
Terminal shall internally use Silence on those unused Pin channels for any processing within the Unit or Output
Terminal.

Release 4.0 April 2023 33

Universal Serial Bus Device Class Definition for Audio Devices

Likewise, If the Unit or Terminal receives an Inactive Cluster on an Input Pin, or if the clock on an Output Terminal’s
Clock Input Pin is currently invalid, it shall disregard any signaling on that Input Pin and internally drive Silence on
all of its Input Pin channels.

Every Unit in the Audio Function is fully described by its associated Unit Descriptor. The Unit Descriptor contains all
necessary fields to identify and describe the Unit. Likewise, there is a Terminal Descriptor for every Terminal in the
Audio Function. In addition, these Descriptors provide all necessary information about the topology of the Audio
Function. They fully describe how Terminals and Units are interconnected.

This specification describes the following types of standard Units and Terminals that are considered adequate to
represent most Audio Functions:

• Input Terminal (IT)

• Output Terminal (OT)
• Mixer Unit (MU)
• Selector Unit (SU)
• Feature Unit (FU)
• Sampling Rate Converter Unit (RU)
• Effect Unit (EU)
• Processing Unit (PU)
• Extension Unit (XU)

Besides Units and Terminals, the concept of a Clock Entity is introduced. Three types of Clock Entities are defined
by this specification:

• Clock Source Entity (CS)

• Clock Selector Entity (CX)

A Clock Source Entity provides a certain sampling clock frequency to all or part of the Audio Function. A Clock
Source Entity may represent an internal sampling frequency generator, but it may also represent an external
sampling clock signal input to the Audio Function. An internal sampling frequency generator can be either
asynchronous or it can be derived from the USB SOF signal. In addition, it can use a Clock Domain Reference Clock
that may be shared with other Clock Source Entities within the Audio Function. This would indicate to the Host that
these Clock Source Entities, although they potentially produce sampling clocks at different sampling frequencies,
are frequency-locked.
A Clock Source Entity has a single Clock Output Pin that carries the sampling clock signal, represented by the Clock
Source Entity. The Clock Output Pin number is always one.

A Clock Selector is used to select between multiple sampling clock signals that may be available inside an Audio
Function. It has multiple Clock Input Pins and a single Clock Output pin. Clock Input Pins are numbered starting
from one up to the total number of Clock Input Pins on the Clock Selector. The Clock Output Pin number is always
one.

By using a combination of Clock Source and Clock Selector Entities, complex clock systems can be represented and
exposed to Host software.

Clock Input and Output Pins are fundamentally different from Input and Output Pins defined for Units and
Terminals. Clock Pins carry only clock signals and therefore cannot be connected to Unit or Terminal Input and
Output Pins. They are only used to express clock circuitry topology.

Release 4.0 April 2023 34

Universal Serial Bus Device Class Definition for Audio Devices

Each Input and Output Terminal has a single Clock Input Pin that may be connected to a Clock Output Pin of a Clock
Entity. The clock signal carried by that Clock Output Pin determines at which sampling frequency the hardware
represented by the Terminal is operating. If there is no need to expose to the Host which clock signal is used by a
Terminal or if the Terminal represents an analog port in the system, then the Clock Input Pin of the Terminal may
be left unconnected.

A Sampling Rate Converter Unit has two Clock Input Pins that are typically connected to the Clock Output Pins of
two different Clock Entities. The clock signals carried by those Clock Output Pins determine the sampling
frequencies between which the Sampling Rate Converter Unit is converting and whether the conversion is
synchronous (the two clock signals are derived from the same Main clock) or asynchronous (the two clock signals
are derived from different, independent, Main clocks).

Each Clock Entity is described by a Clock Entity Descriptor. The Clock Entity Descriptors contain all necessary fields
to identify and describe the Clock Entities. In addition, these Descriptors provide all necessary information about
the clock topologies within the Audio Function.

A physical Connector, typically user-accessible, is represented by a Connector Entity and described by a Connector
Entity Descriptor. The Connector Entity is always associated with one or more Input and/or Output Terminals
through which the signals, present on the Connector, enter or leave the Audio Function.

Connector Entities and Power Domain Entities are slightly different in concept from the other Entities in the sense
that they do not convey topological information. They are conceptual constructs that provide addressability for the
AudioControls that reside within them, and they are described by a Connector Entity Descriptor or Power Domain
Entity Descriptor respectively.

The Descriptors are further detailed in Section 4, “Descriptors” of this document.

The ensemble of AudioControl Interface Descriptor, AudioStreaming Interface Descriptors, Endpoint Descriptors,
Unit Descriptors, Terminal Descriptors, Clock Entity Descriptors, Connector Entity Descriptors, and Power Domain
Entity Descriptors provide a full description of the Audio Function to the Host. This information is typically
retrieved from the Device at enumeration time. By parsing the Descriptors, an Audio Class driver should be able to
fully control the Audio Function.

Important Note:

The complete set of Audio Function Descriptors provides only a static initial description of the Audio
Function. During operation, several events may happen that force the Audio Function to change its
state. Host software shall be notified of these changes to remain ‘in sync’ with the Audio Function at
all times. An extensive interrupt mechanism is in place to report all state changes to Host software.

This specification defines a set of symbols to graphically represent the building blocks, discussed above. This allows
the creation of standardized topology diagrams to describe Audio functionality. See Figure 3-5, “Inside the Audio
Function” for an example diagram. Input Terminals are typically located at the far left, while Output Terminals are
at the far right of the topology diagram. The symbols, representing Audio Function Entities have their Input Pins on
the left side of the symbol, and their single Output Pin on the right side. Input Pins are always numbered in
ascending order, starting with Input Pin 1 in the upper left corner of the symbol, so that there is no need to
explicitly provide Input Pin number labels on the symbols.

Figure 3-5, “Inside the Audio Function” illustrates some of the concepts defined above. Using the symbols defined
further, it describes a hypothetical Audio Function that incorporates 15 Entities: three Input Terminals, five Units,
three Output Terminals, three Clock Source Entities, and a Clock Selector Entity. Each Entity has its unique ID (from
1 to 15) and Descriptor that fully describes the functionality of the Entity and how that particular Entity is

Release 4.0 April 2023 35

Universal Serial Bus Device Class Definition for Audio Devices

connected into the topology of the Audio Function. Note that the AudioControl Interface itself is also considered
an implicit addressable Entity with unique ID zero.

Input Terminal 1 (IT 1) is the representation of a USB OUT Endpoint used to stream audio from the Host to the
Audio Device. IT 2 is the representation of an analog Line-In connector on the Audio Device whereas IT 3 is an
analog Microphone-In connector on the Audio Device. Selector Unit 4 (SU 4) selects between the audio coming
from the Host and the audio present at the Line-In connector. Feature Unit 5 (FU 5) is then used to manipulate the
audio (Gain, Bass, Treble …) before it is presented to Output Terminal 9 (OT 9). OT 9 is the representation of a
Headphone Out jack on the Audio Device.

At the same time, all three input sources (USB OUT, Line-In, and Microphone-In) are connected to a Mixer Unit 6
(MU 6) that effectively mixes the three sources together. The output of the Mixer is then fed into a Processing Unit
7 (PU 7) that performs some audio processing algorithm(s) on the mix. The result is in turn sent to FU 8 where
some final adjustments to the audio (Gain …) are made. FU 8 is connected to OT 10 and OT 11. OT 10 represents
speakers incorporated into the Audio Device and OT 11 represents a USB IN Endpoint used to send the processed
audio to the Host for recording purposes.

Clock Source Entity 12 (CS 12) represents an internal sampling frequency generator, running at 96 kHz for instance.
Clock Source Entity 15 (CS 15) is the representation of an external reference sampling clock input that may be used
to synchronize the Device to an external source. Clock Selector Entity 13 (CS 13) enables selection between the
two available Clock Source Entities. The output of CS 13 provides a sampling frequency to IT 1, IT 2, IT3, OT 10, and
OT 11 of 96 kHz. Clock Source Entity CS 14 further provides a sampling frequency of 48 kHz to OT 9 for driving the
headphone. Since all sampling frequencies used inside the Audio Function are at all times derived from a single
Main clock (internal or external) as indicated in the Clock Source Entity Descriptors, all audio streams in the Audio
Function are synchronous.

The Descriptors, associated with each Entity clearly indicate to the Host what the exact nature of each Entity is. For
instance, the IT 2 Descriptor contains a field that indicates to the Host that it represents an external connector on
the Device, used as an analog Line-In. Likewise, the MU 6 Descriptor has a field that indicates that its Input Pin 1 is
connected to the Output Pin of IT 1, Input Pin 2 is connected to the Output Pin of IT 2, and Input Pin 3 is connected
to the Output Pin of IT 3.

For further details on Descriptor contents, refer to Section 4, “Descriptors” of this document.

Release 4.0 April 2023 36

Universal Serial Bus Device Class Definition for Audio Devices

Figure 3-5: Inside the Audio Function

0
AudioControl Interface

Headphone
USB OUT Selector Unit Feature Unit
nc
OUT
1 4 5 9

IT OT

IT Descr SU Descr FU Descr OT Descr

Analog Line
Speakers
IN
2 10

IT OT

Mixer Unit Processing Unit Feature Unit

IT Descr OT Descr
6 7 8

∑ PU

Analog Mic
IN USB IN
MU Descr PU Descr FU Descr
3 11

IT OT

IT Descr OT Descr
12

13 14
nc
CS Descr

15
CX Descr CS Descr

CS Descr

3.13.1 AUDIOCONTROLS
Inside an Entity, functionality is further described through AudioControls. An AudioControl typically provides
access to a specific audio or clock property. Each AudioControl has a set of Attributes that can be manipulated or
that present additional information about the behavior of the AudioControl. An AudioControl has the following
Attributes:

• Capability (CAP) Attribute (Mandatory)

• Current (CUR) Attribute (Mandatory)
• Next (NEXT) Attribute (Optional)

For details about AudioControl Attributes and their Read-Write privileges, see Section 5.3.2.2, “AudioControl
Attributes” and Section 5.3.2.3, “AudioControl Read/Write Privileges.”

As an example, consider a Gain Control inside a Feature Unit. By issuing the appropriate Pull Commands, the Host
software can obtain values for the Gain Control’s Attributes and, for instance, use them to correctly display the
Gain Control to the user. All relevant information that Host software needs to interact with the Gain Control can be
retrieved via a Pull Command to the CAP Attribute. The Gain Control’s CUR Attribute allows the Host software to

Release 4.0 April 2023 37

Universal Serial Bus Device Class Definition for Audio Devices

directly change the level setting of the Control. Setting the Control’s NEXT Attribute allows the Host software to
prepare a change to the setting of the Control. This change will only take effect when Host software issues a
Commit Command. For more details, refer to Section 3.14.3, “AudioControls.”

3.13.2 CLUSTER
A Cluster is a grouping of audio channels that carry tightly related synchronous audio information and that travels
over the connections among Terminals and Units. Inside the Audio Function, complete abstraction is made of the
actual physical representation of the audio in the Cluster. Each audio channel in the Cluster is a logical channel and
all the physical attributes of the channel (sampling frequency, bit width, bit resolution, etc.) are not specified and
considered irrelevant in the context of the Audio Function as seen through the AudioControl Interface. The fact
that an Input Pin and an Output Pin are connected in the Audio Function’s topology is a guarantee (by
construction) that the audio flowing out of the Output Pin on the source Entity is compatible with and consumable
by the Input Pin of the receiving Entity or Entities. This may involve some conversion processes that happen
“behind the scenes”. The details of these conversion processes, however, are beyond the scope of the logical view
exposed by the Audio Function.

Channel numbering in the Cluster starts with channel one up to the number of channels in the Cluster. Channel
Number zero is used to reference the Primary channel. See Section 3.14.3, “AudioControls” for more details.

A typical Cluster is characterized by several Attributes:

• A unique identifier for the Cluster

• The number of audio channels in the Cluster
• For each Channel in the Cluster:
• The purpose of the audio channel, such as Voice, Speech, etc.
• The audio channel relationship (or spatial location)
• A Function-wide unique Channel ID
• An indication whether the audio channel is part of a ChannelGroup

Note: This specification also supports Ambisonic Clusters that have somewhat different Attributes. See
Section 4.4, “Cluster Descriptor” for more details.

Each Cluster has an associated Cluster Descriptor that fully describes the Cluster.

A Cluster that is currently carrying valid audio information is called an Active Cluster. This specification also defines
the Inactive Cluster, in which case the audio channels in the Cluster do not carry any audio and the audio content
of the channels is effectively undefined.

Note: This is different from a Cluster that carries Silence data in its audio channels. Such a Cluster is still
considered Active where audio is streaming, but the audio signals (samples or voltage levels)
correspond to a zero audio level (muted audio).

The transition between Active and Inactive is always the result of a change within the Audio Function. The
following situations may occur:

• The Output Cluster on an Entity’s Output Pin may switch between Active and Inactive due to a change in
Power State of the Power Domain to which the Entity belongs.
• The Output Cluster on a Selector Unit’s Output Pin may switch between Active and Inactive depending on the
position of the Unit’s Selector Control.

Release 4.0 April 2023 38

Universal Serial Bus Device Class Definition for Audio Devices

• The Output Cluster on an Input Terminal’s Output Pin may switch between Active and Inactive depending on
whether audio data is present or not on the interface the Input Terminal represents. Likewise, a switch may
occur depending on the presence of a valid clock on the Clock Input Pin of an Input Terminal.

Note that the Cluster Descriptor itself does not indicate whether the Cluster is Active or Inactive. The state of a
Cluster (Active or Inactive) is always derived from another state in the Audio Function (see above). The Cluster’s
active state is exposed via a Read-Only Cluster Active Control, reflecting the current state of the Cluster (Active or
Inactive). When a Cluster becomes Inactive, its Cluster Descriptor becomes undefined and shall not be used for any
purpose. For example, an Entity that advertises a static output Cluster on its Output Pin may carry an Inactive
Cluster at some point in time due to events that happen upstream in the audio path. In this case, although the
Cluster Descriptor is still available and unchanged, the Cluster is Inactive, and the Cluster Descriptor shall not be
used.

3.13.2.1 CHANNEL ID
Each channel in a Cluster shall be uniquely identifiable by its Channel ID. The Channel ID is a non-zero value in the
wChannelID field in the channel’s Information or Ambisonic Segment of the Cluster Descriptor. The Channel ID
shall be Function-wide unique. The following rules apply:

• The Audio Function implementation shall assign a Channel ID to each channel entering the Audio Function via
the Output Pin of an Input Terminal so that Host software has a means to trace individual channels as they
traverse the Audio Function.
• Each Feature Unit, Effect Unit, or SRC Unit shall preserve the Channel IDs from its incoming Cluster to the
outgoing Cluster.
• The Selector Unit shall preserve the Channel IDs from its currently selected incoming Cluster to the outgoing
Cluster.
• A Mixer Unit, Processing Unit, or Extension Unit typically creates ‘new’ content in their outgoing Cluster, by
applying some algorithm on any or all incoming channels. In this case, the Audio Function implementation
should assign a new Channel ID to each channel in the outgoing Cluster. However, if there is a reason to
indicate that the outgoing Cluster contains channel content predominantly originating from a specific
incoming channel, then the Channel ID of that incoming channel may appear in the outgoing Cluster. It is left
to the implementation to carefully consider how to populate the Channel IDs in the outgoing Cluster of these
Unit types.

Subsequent sections provide more guidance on how to manage Channel IDs for all defined Terminal and Unit
types.

3.13.3 INPUT TERMINAL

The Input Terminal (IT) is used to interface between the Audio Function’s ‘outside world’ and other Units in the
Audio Function. It serves as a receptacle for audio information flowing into the Audio Function. Its purpose is to
represent a source of incoming audio data after this data has been properly extracted from the original audio
stream into the separate logical channels that are embedded in this stream (the decoding process). The logical
channels are grouped into a Cluster and leave the Input Terminal through its single Output Pin.

The Input Terminal that represents an audio stream entering the Audio Function by means of a USB OUT Endpoint,
shall have a dedicated AudioStreaming OUT Interface that contains this Endpoint and there shall be a one-to-one
relationship between that AudioStreaming Interface and its associated Input Terminal.

Release 4.0 April 2023 39

Universal Serial Bus Device Class Definition for Audio Devices

The Input Terminal may represent inputs to the Audio Function other than USB OUT Endpoints. A Line-In
connector on an Audio Device is an example of such a non-USB input. A digital input connector, such as S/PDIF, is
another example.

The Input Terminal Descriptor contains a field that either holds a direct reference to its associated AudioStreaming
Interface or contains a list of Connector Entity IDs referencing its associated Connector Entities. The Host needs to
use both the AudioStreaming Interface and Endpoint Descriptors or the Connector Entity Descriptors, in
conjunction with the Input Terminal Descriptor to get a full understanding of the characteristics and capabilities of
the Input Terminal. Stream-related parameters are stored in the AudioStreaming Interface or Connector Entity
Descriptors. AudioControl-related parameters are stored in the Input Terminal Descriptor. Stream-related
AudioControls are in the AudioStreaming Interface or the Connector Entity. Audio control-related AudioControls
are in the Input Terminal.

The conversion process from incoming, possibly encoded, audio streams to logical audio channels always involves
some form of decoding engine. The decoding types range from rather trivial decoding schemes like converting
interleaved stereo 16-bit PCM data into Front Left and Front Right logical channels to very sophisticated schemes
like converting an MPEG-2 7.1 encoded audio stream into Front Left, Front Left of Center, Front Center, Front Right
of Center, Front Right, Back Left, Back Right and Low Frequency Effects logical channels. The decoding engine is
considered part of the Entity that receives the encoded audio data streams (like a USB AudioStreaming Interface).
The type of decoding is therefore implied by the value in the wFormat field, located in the class-specific
AudioStreaming Self Descriptor. The associated Input Terminal deals with the logical channels after they have been
decoded.

If there is an AudioStreaming Interface associated with the Input Terminal, then the Cluster Configuration on the
Output Pin of the Input Terminal shall be determined through selection of the appropriate Alternate Setting of the
associated AudioStreaming Interface. The Input Terminal shall have a Read-Only (r) Cluster Control, indicating
which Cluster Configuration is currently in use. There shall be a one-to-one relationship between the Alternate
Setting of the Interface and the corresponding Cluster Configuration as indicated by the CUR value of the Cluster
Control. In other words, the reported CUR Cluster Control value shall be identical to the currently selected
Alternate Setting in the associated AudioStreaming Interface. Also, the Cluster Configuration shall be compatible
with the audio data format of the currently selected Alternate Setting of the AudioStreaming Interface. If Alternate
Setting zero is currently selected, then the output Cluster shall be Inactive, and the reported CUR value of the
Cluster Control shall be zero.

If there are one or more Connector Entities or internal transducers associated with the Input Terminal, then the
Input Terminal shall have a Read-Write (rw) Cluster Control, used to select the Cluster Configuration on the Output
Pin of the Input Terminal. It is the responsibility of the Host Software, potentially via user intervention, to choose a
Cluster Configuration that is appropriate for the type of device that is currently plugged into the external
Connector(s).

If the Input Terminal represents an embedded (set of) transducer(s), then the Cluster Control can be used to
control the behavior of the transducer(s), such that they produce an audio stream that is compatible with the
selected Cluster Configuration.

In both cases, the Input Terminal shall expose a Read-Only (r) Cluster Active Control that indicates whether the
output Cluster is currently Active or Inactive.

The Audio Function shall assign unique Channel IDs to all channels in the output Cluster of the Input Terminal so
that Host software has a means to trace individual channels as they traverse the Audio Function.

Release 4.0 April 2023 40

Universal Serial Bus Device Class Definition for Audio Devices

The Input Terminal has a single Clock Input Pin. The clock signal present at that Pin is used as the sampling clock for
all underlying hardware that is represented by this Input Terminal. There is a field in the Input Terminal Descriptor
that uniquely identifies the Clock Entity to which the Input Terminal is connected. If there is no need for the Audio
Function to expose any clock information related to the Input Terminal, the Clock Input Pin of the Input Terminal
may be left unconnected. In this case, it is assumed that the Input Terminal is internally connected to a clock that is
always valid.

The Input Terminal optionally provides the Voltage Control. It is used to set the voltage level of the power supply
that provides either phantom power or bias voltage to the Input Terminal’s associated microphone.

Depending on the state of the physical audio source the Input Terminal represents, the current Cluster on the
Output Pin may be Inactive. For example, an AudioStreaming OUT Interface may currently be set to Alternate
Setting 0 (non-streaming) and therefore, the Input Terminal that represents this Interface will have an Inactive
output Cluster. Likewise, if the clock on the Input Terminal’s Clock Input Pin is currently invalid, then the output
Cluster shall be Inactive.

If the Input Terminal is an explicit member of a Power Domain, switching the Power Domain to any Power State
other than PS0 or PS1 shall render the Input Terminal non-functional, and its output Cluster shall be Inactive.

In some cases, the Audio Function needs to indicate to Host software that the Input Terminal is functionally or
physically related to one or more other Terminals or Entities. This can be expressed by using the Group construct.
A typical example of such a relationship is one or more Input Terminals, representing a set of microphones, and an
Output Terminal, representing the earpieces of a headset. They can be grouped together in a single Group that has
these Terminals as its Members.

The symbol for the Input Terminal is depicted in the following figure:
Figure 3-6: Input Terminal Icon

3.13.4 OUTPUT TERMINAL

The Output Terminal (OT) is used to interface between Units inside the Audio Function and the ‘outside world’. It
serves as an outlet for audio information, flowing out of the Audio Function. Its purpose is to represent a sink of
outgoing audio data before this data is properly packed from the original separate logical channels into the
outgoing audio stream (the encoding process). The Cluster enters the Output Terminal through a single Input Pin.
The Cluster Configuration on the Input Pin may change dynamically, depending on certain actions that occur on
the upstream connection. This results in a change in the physical Cluster Configuration within the physical interface
the Output Terminal represents.

An Output Terminal that represents an audio stream leaving the Audio Function by means of a USB IN Endpoint,
shall have a dedicated AudioStreaming IN Interface that contains this Endpoint and there shall be a one-to-one
relationship between that AudioStreaming Interface and its associated Output Terminal. In this case, when the
Cluster on the Output Terminal’s Input Pin changes, or if the clock to the Output Terminal becomes invalid, then
the Audio Function shall switch the Active Alternate Setting of the AudioStreaming Interface to Alternate Setting 0

Release 4.0 April 2023 41

Universal Serial Bus Device Class Definition for Audio Devices

and update the Valid Alternate Settings Control to reflect which Alternate Settings are compatible with the new
Cluster Configuration. The Host may then take appropriate action to start the stream.

If the Output Terminal is an explicit member of a Power Domain, switching the Power Domain to any Power State
other than PS0 or PS1 shall render the Output Terminal non-functional, and the Audio Function shall switch the
Active Alternate Setting of the AudioStreaming Interface to Alternate Setting 0 and update the Valid Alternate
Settings Control to reflect that none of the Alternate Settings are currently valid.

The Output Terminal may represent outputs from the Audio Function other than USB IN Endpoints. A speaker built
into an Audio Device, or a Line-Out connector is an example of such a non-USB output. For an Output Terminal
that does not represent an AudioStreaming Interface, handling of Cluster Configuration changes at the Output
Terminal’s Input Pin is left to the Audio Function implementation.

The Output Terminal Descriptor contains a field that either holds a direct reference to its associated
AudioStreaming Interface or contains a list of Connector Entity IDs referencing its associated Connector Entities or
internal transducers. The Host needs to use both the AudioStreaming Interface and Endpoint Descriptors or the
Connector Entity Descriptors, in conjunction with the Output Terminal Descriptor to get a full understanding of the
characteristics and capabilities of the Output Terminal. Stream-related parameters are stored in the
AudioStreaming Interface or Connector Entity Descriptors. AudioControl-related parameters are stored in the
Output Terminal Descriptor. Stream-related AudioControls are in the AudioStreaming Interface or the Connector
Entity. Audio control-related AudioControls are in the Output Terminal.

The conversion process from incoming logical audio channels to possibly encoded audio streams always involves
some form of encoding engine. The encoding engine is considered part of the Entity that transmits the encoded
audio data streams (like the AudioStreaming Interface). The type of encoding is therefore implied by the value in
the wFormat field, located in the class-specific AudioStreaming Self Descriptor. The associated Output Terminal
deals with the logical channels before encoding.

The Output Terminal has a single Clock Input Pin. The clock signal present at that Pin is used as the sampling clock
for all underlying hardware that is represented by this Output Terminal. There is a field in the Output Terminal
Descriptor that uniquely identifies the Clock Entity to which the Output Terminal is connected. If there is no need
for the Audio Function to expose any clock information related to the Output Terminal, the Clock Input Pin of the
Output Terminal may be left unconnected. In this case, it is assumed that the Output Terminal is internally
connected to a clock that is always valid.

In some cases, the Audio Function needs to indicate to Host software that the Output Terminal is functionally or
physically related to one or more other Terminals or Entities. This can be expressed by using the Group construct.
A typical example of such a relationship is one or more Input Terminals, representing a set of microphones, and an
Output Terminal, representing the earpieces of a headset. They can be grouped together in a single Group that has
these Terminals as its Members.

The symbol for the Output Terminal is depicted in the following figure:

Release 4.0 April 2023 42

Universal Serial Bus Device Class Definition for Audio Devices

Figure 3-7: Output Terminal Icon

Audio OT

Clock

3.13.4.1 EN 50332 SUPPORT

This specification supports both the EN 50332-2:2013 and the EN 50332-3:2017 standards. Both are aimed to limit
the amount of exposure to high SPL levels, typically when using headphones or headsets for extended periods of
time. The EN 50332-2 standard describes a method of passively reporting acoustic or electrical output levels
whereas the EN 50332:3 standard provides a more dynamic dosage reporting mechanism where the Device
actively calculates actual dosage levels on an ongoing basis. For details, refer to the appropriate specifications [EN
50332-2:2013] and [EN 50332-2:2017]. A Device is allowed to support either EN 50332-2 or EN 50332-3
specifications. However, if the Device supports EN 50332-3, it is recommended the Device also supports EN 50332-
2.

3.13.4.1.1 EN 50332-2
When the Device supports the EN 50332-2:2013 specification, it exposes either an EN 50332-2 Acoustic Level
Segment or an EN 50332-2 Voltage Level Segment as part of its Terminal Companion Descriptor for each Output
Terminal that supports the EN 50332-2 specification. (For details, see Section 4.5.3.4.3.2.1, “EN 50332-2 Acoustic
level Segment” and Section 4.5.3.4.3.2.2, “EN 50332-2 Voltage level Segment.”) Host software can then use this
information to gauge and potentially limit dosage and exposure levels to the user.

3.13.4.1.2 EN 50332-3
When the Device supports the EN 50332-3:2017 specification, it exposes a Momentary Exposure Level (MEL)
Control in the Output Terminal that supports the specification. The MEL Control periodically reports back to the
Host the current Exposure Level, according to the following rules:

Every second, the USB Audio Device shall compute the Momentary Exposure Level based on all of the audio
samples feeding the digital-to-analog converters within that second. The MEL Control shall generate an interrupt
message within 2 interrupt endpoint Service Intervals of the new MEL value becoming available. The intent is for
the Host to read the MEL Control in response to each interrupt. The Host is then responsible for utilizing the MEL
values to perform additional calculations and threshold comparisons, culminating in the display of appropriate
warnings and/or the reduction of the output level in accordance with EN 50332-3:2017 and related standards.

To compute Momentary Exposure Level (MEL) for headphones or a headphone jack, a Device shall perform the
following computations every second:

1. Apply an A-weighting filter to the stereo digital audio signal that is sent to the digital-to-analog converter.
2. To the result of step 1, apply a transfer function representing the digital-to-analog converter and
amplifier’s frequency response and output gain, resulting in a stereo audio signal expressed in volts (V).
• If the output is a headphone speaker, apply a transfer function to the result representing the
headphone speaker’s frequency response and sensitivity, assuming a Head And Torso Simulator
(HATS) diffuse field measurement.

Release 4.0 April 2023 43

Universal Serial Bus Device Class Definition for Audio Devices

• If the output is a headphone jack, assume a default sensitivity (as derived from the relevant
tables in the EN 50332-3 specification) for an unknown set of attached headphones by
multiplying by 40/3 Pa/V.
This will result in a stereo audio signal expressed in pascals (Pa).
3. To each audio channel of the result of step 2, square all the samples within this 1-second interval, sum
them, and divide by the number of samples per audio channel within the 1-second interval, producing
two “mean square” numbers expressed in pascals squared (Pa2).
4. Using the results of step 3, add the two “mean square” numbers, divide by the square of 20 µPa, take the
base-10 logarithm, multiply by 10 to yield dB, and finally subtract 3 dB (as per the EN 50332-3
specification). This produces a single MEL value expressed in dB(A).

The following figure further illustrates the procedure outlined above:

Figure 3-8: MEL Procedure

3.13.5 MIXER UNIT

The Mixer Unit (MU) transforms two or more logical input channels into one or more logical output channels. The
input channels are grouped into one or more Clusters. Each Cluster enters the Mixer Unit through one of the 𝑝
Input Pins. The logical output channels are grouped into one Cluster and leave the Mixer Unit through a single
Output Pin.

The Pin channels on each Input Pin 𝑖 of the Mixer Unit are numbered from one to 𝑁𝑖 . The PCC value 𝑁𝑖 for each of
the Input Pins is inherited from the first upstream Entity that defines a PCC value on its Output Pin.

Every input channel can be mixed into all the output channels. If 𝑁 is the total number of Input Pin channels (𝑁 =
∑𝑝𝑖=1 𝑁𝑖 ) and 𝑀 is the number of Output Pin channels the Mixer supports, then there is a two-dimensional array
(𝑁 × 𝑀) of Mixer Controls in the Mixer Unit. Some Mixer Controls may be Read-Only and have fixed values. For
example, a ‘missing’ connection may be advertised as a Read-Only Mixer Control with fixed value - dB.

On each Input Pin 𝑖, the incoming Cluster logical channels, numbered from one to 𝑛𝑖 , are mapped one-to-one onto
the Pin channels on that Input Pin, i.e. for each Input Pin 𝑖, Cluster channel 1 is mapped onto Pin channel 1, Cluster
channel 2 is mapped onto Pin channel 2, and so on, up to Cluster channel 𝑛𝑖 which is mapped onto Pin channel 𝑛𝑖 .
Note that the PCC values on each of the Mixer Unit’s Input Pins (𝑁𝑖 ) may be different from the actual number of
channels in the incoming Clusters (𝑛𝑖 ). However, for each Input Pin 𝑖, 𝑁𝑖 ≥ 𝑛𝑖 shall always be true.

Release 4.0 April 2023 44

Universal Serial Bus Device Class Definition for Audio Devices

Mixer Controls residing on currently unused Pin channels shall always remain accessible and retain their last
setting for as long as the Mixer Unit remains in a Power State that requires this.

The Cluster Control is typically implemented as Read (r) and exposes a single Cluster Configuration. However, the
Mixer Unit can expose multiple Cluster Configurations on its Output Pin and implement the Cluster Control as
Read-Write (rw). A Cluster Control value of zero indicates that the output Cluster Configuration is inherited from
the current Cluster on Input Pin one (the dominant Input Pin). The Read-Only Cluster Active Control shall always be
implemented and indicates whether the output Cluster is currently Active or not.

Since the Mixer Unit likely redefines the channels in its output Cluster as a processed result of any of the incoming
channels, it would be appropriate to treat these channels as independent of all other channels in the Audio
Function and therefore assign them unique IDs in their respective wChannelID fields in the output Cluster.
However, this is not a requirement, and it is left to the implementation to assign Channel ID values as accurately as
possible. For example, a Mixer Unit that has two inputs, each carrying a stereo Cluster, where the Cluster on Input
Pin one is the main audio stream and the Cluster on Input Pin two is mixed into that main audio stream, it may be
appropriate to propagate the Channel IDs from the Cluster on Input Pin one to the output Cluster.

The Mixer Unit can also be used to repurpose or redefine the channel relationships in a Cluster. By creating a Mixer
Unit with a single Input Pin, and setting up the proper Mixer Controls, it can freely define its Output Cluster and its
Descriptor to serve the intended purpose. For example, an incoming stereo Cluster that is defined as Front Left,
Front Right, can be redefined to become Headphone Left, Headphone Right, by setting the Mixer Controls such
that the Front Left and Front Right channels get fully mixed into output channels one and two respectively and by
setting the wRelationship fields of the output Cluster Descriptor to Headphone Left and Headphone Right,
respectively. As another example, the wPurpose fields of certain channels could be modified to reflect the actual
purpose of the output Cluster of the Mixer Unit. This could be used to create a Cluster specifically designated for
ultrasonic purposes, starting from an incoming Cluster that contains full bandwidth information, including
ultrasonic information.

If the Mixer Unit is an explicit member of a Power Domain, switching the Power Domain to any Power State other
than PS0 or PS1 shall render the Unit non-functional, and its output Cluster shall be Inactive.

The symbol for the Mixer Unit can be found in the following figure:
Figure 3-9: Mixer Unit Icon

3.13.6 SELECTOR UNIT

The Selector Unit (SU) has 𝑝 Input Pins and a single Output Pin. It incorporates a Selector Control that allows to
select one Cluster from the 𝑝 Clusters on the Input Pins, and routes that Cluster unaltered to the single Output Pin.
It represents a multi-channel source selector, capable of selecting between 𝑝 sources where 𝑝 shall be larger than
one. Note that each incoming Cluster may have different characteristics in terms of the number of channels, their
Purpose, Relationship, Grouping, or Channel IDs. To determine the current Cluster on the Output Pin, Host
software needs to use the currently selected Input Pin and track the audio path upstream on that Input Pin until it

Release 4.0 April 2023 45

Universal Serial Bus Device Class Definition for Audio Devices

finds an Entity’s Output Pin that advertises an Output Cluster definition and its Active State (the Selector Unit itself
does not contain a Cluster Active Control).

Each of the Input Pins inherits the PCC value 𝑁𝑝 from the first upstream Entity that defines a PCC value on its
Output Pin. The Selector Unit does not advertise a PCC value on its output. It is the maximum of all PCC values on
its Input Pins.

The Selector Unit also has the optional capability to disconnect the output from all its inputs. In this case the
Cluster on the Output Pin does not contain audio and is therefore an Inactive Cluster. If the disconnect option is
supported, the range of the Selector Control CUR Attribute shall include zero.

If the Selector Unit is an explicit member of a Power Domain, switching the Power Domain to any Power State
other than PS0 or PS1 shall render the Unit non-functional, and its output Cluster shall be Inactive.

The symbol for the Selector Unit can be found in the following figure:
Figure 3-10: Selector Unit Icon

nc

3.13.7 FEATURE UNIT

The Feature Unit (FU) is essentially an 𝑁-channel Unit that provides basic manipulation of a (cascaded) set of
single-parameter audio processes on the incoming logical channels. For each Pin channel, the Feature Unit
optionally provides AudioControls for the following features:

• Bypass
• Mute
• Gain
• Tone Control (Bass, Mid, Treble)
• Graphic Equalizer
• Automatic Gain Control
• Delay
• Bass Boost
• Loudness
• Input Gain Pad
• Phase Inverter

In addition, the Feature Unit optionally provides the above AudioControls but now influencing all channels at once.
In this way, Cluster-wide AudioControls can be implemented. The Cluster-wide AudioControls are cascaded after
the individual channel AudioControls. This setup is especially useful in multi-channel systems where the individual
channel AudioControls may be used for channel balancing and the Cluster-wide AudioControls may be used for
overall settings.

A Feature Unit shall implement at least one AudioControl.

Release 4.0 April 2023 46

Universal Serial Bus Device Class Definition for Audio Devices

The Pin channels in the Feature Unit are numbered from one to 𝑁. The PCC value 𝑁 is inherited from the first
upstream Entity that defines a PCC value on its Output Pin. The Primary channel has channel number zero and is
always virtually present.

The Feature Unit Descriptor reports which AudioControls are present for every Pin channel in the Feature Unit,
including the Primary channel. All Pin channels in the Feature Unit are fully independent. There exists no cross
coupling among channels within the Feature Unit.

The Feature Unit never alters the incoming Cluster in terms of the number of logical channels in the Cluster, 𝑛, nor
their Purpose, Relationship, Grouping, or Channel IDs. In other words, the Feature Unit does not in any way
redefine the incoming Cluster description and there are always as many logical output Cluster channels as there
are input Cluster channels. The Cluster enters the Feature Unit through a single Input Pin and leaves the Unit
through a single Output Pin.

The incoming logical Cluster channels are mapped one-to-one onto the Pin channels of the Feature Unit, i.e.,
Cluster channel 1 is mapped onto Pin channel 1, Cluster channel 2 is mapped onto Pin channel 2, and so on, up to
Cluster channel 𝑛 which is mapped onto Pin channel 𝑛. Note that the number of Pin channels (𝑁) in the Feature
Unit may be different from the actual number of logical channels in the incoming Cluster (𝑛). However, 𝑁 ≥ 𝑛 shall
always be true.

AudioControls residing on currently unused Pin channels shall always remain accessible and retain their last setting
for as long as the Feature Unit remains in a Power State that requires this.

If the optional Bypass Control is present on a Pin channel, then engaging the bypass function shall result in passing
the corresponding logical channel of the incoming Cluster unaltered to that same logical channel of the outgoing
Cluster.

If the optional Bypass Control is present on the Primary channel, then engaging the bypass function shall result in
passing the incoming Cluster unaltered to the output of the Unit.

If the Feature Unit is an explicit member of a Power Domain, switching the Power Domain to any Power State
other than PS0 or PS1 shall render the Unit non-functional, and its output Cluster shall be Inactive.

The symbol for the Feature Unit is depicted in the following figure:
Figure 3-11: Feature Unit Icon

3.13.8 SAMPLING RATE CONVERTER UNIT

The Sampling Rate Converter (SRC) Unit (RU) is included here as an optional way to indicate where exactly within
the Audio Function sampling rate conversion takes place. In many cases, it is unnecessary to indicate this point and
any SRC Unit may be omitted from the topology without materially affecting the information presented to the
Host. The primary reason to include the SRC Unit is to accurately report any latencies incurred by the Sampling
Rate Conversion process.

Release 4.0 April 2023 47

Universal Serial Bus Device Class Definition for Audio Devices

The SRC Unit provides a bridge function between different Clock Domains within the Audio Function. The SRC Unit
does not provide AudioControls that impact the SRC functionality of the Unit. It takes the audio on all the logical
channels in the input Cluster belonging to a certain Clock Domain and converts them into the same logical
channels in the output Cluster but now belonging to another Clock Domain.

The SRC Unit never alters the incoming Cluster in terms of the number of logical channels in the Cluster, 𝑛, nor
their Purpose, Relationship, Grouping, or Channel IDs. In other words, the SRC Unit does not in any way redefine
the incoming Cluster description and there are always as many logical output Cluster channels as there are input
Cluster channels. The Cluster enters the SRC Unit through a single Input Pin and leaves the Unit through a single
Output Pin.

The Pin channels in the SRC Unit are numbered from one to 𝑁. The PCC value 𝑁 is inherited from the first
upstream Entity that defines a PCC value on its Output Pin. There is no Primary channel.

The incoming logical Cluster channels are mapped one-to-one onto the Pin channels of the SRC Unit, i.e., Cluster
channel 1 is mapped onto Pin channel 1, Cluster channel 2 is mapped onto Pin channel 2, and so on, up to Cluster
channel 𝑛 which is mapped onto Pin channel 𝑛. Note that the number of Pin channels (𝑁) in the SRC Unit may be
different from the actual number of logical channels in the incoming Cluster (𝑛). However, 𝑁 ≥ 𝑛 shall always be
true.

The SRC Unit has two Clock Input Pins. One Clock Input Pin is associated with the single Input Pin of the SRC Unit.
The other Clock Input Pin is associated with the single Output Pin of the SRC Unit. The clock signals present at
those two Clock Input Pins identify the two Clock Domains between which the SRC Unit is converting. Note that it
is allowed to have both Clock Input Pins connected to clock signals belonging to the same Clock Domain. It is also
allowed to leave one or both Clock Input Pins unconnected if there is no need for the Audio Function to expose
clock information related to the unconnected side of the SRC Unit.

If the SRC Unit is an explicit member of a Power Domain, switching the Power Domain to any Power State other
than PS0 or PS1 shall render the Unit non-functional, and its output Cluster shall be Inactive.

The symbol for the SRC Unit is depicted in the following figure:
Figure 3-12: Sampling Rate Converter Unit Icon

Audio Audio
Clock1 Clock2

3.13.9 EFFECT UNIT

The Effect Unit (EU) is a multi-channel processing unit that provides advanced manipulation of a multi-parameter
AudioControl on the incoming logical channels on a per-channel basis. For each logical channel, the Effect Unit
shall provide one of the following AudioControls:

• Parametric Equalizer Section

• Reverberation
• Modulation Delay
• Dynamic Range Compressor

Release 4.0 April 2023 48

Universal Serial Bus Device Class Definition for Audio Devices

In addition, the Effect Unit optionally provides one of the above AudioControls but now influencing all channels of
the Cluster at once. In this way, a Cluster-wide AudioControl may be implemented. The Cluster-wide AudioControl
is cascaded after the individual channel AudioControls. This setup is especially useful in multi-channel systems
where the individual channel AudioControls may be used for channel balancing and the Cluster-wide AudioControl
may be used for overall settings.

The Pin channels in the Effect Unit are numbered from one to 𝑁. The PCC value 𝑁 is inherited from the first
upstream Entity that defines a PCC value on its Output Pin. The Primary channel has channel number zero and is
always virtually present.

The Effect Unit Descriptor reports which AudioControls are present for every Pin channel in the Effect Unit,
including the Primary channel. All Pin channels in the Effect Unit are fully independent. There exists no cross
coupling among channels within the Effect Unit.

The Effect Unit never alters the incoming Cluster in terms of the number of logical channels in the Cluster, 𝑛, nor
their Purpose, Relationship, Grouping, or Channel IDs. In other words, the Effect Unit does not in any way redefine
the incoming Cluster description and there are always as many logical output Cluster channels as there are input
Cluster channels. The Cluster enters the Effect Unit through a single Input Pin and leaves the Unit through a single
Output Pin.

The incoming logical Cluster channels are mapped one-to-one onto the Pin channels of the Effect Unit, i.e., Cluster
channel 1 is mapped onto Pin channel 1, Cluster channel 2 is mapped onto Pin channel 2, and so on, up to Cluster
channel 𝑛 which is mapped onto Pin channel 𝑛. Note that the number of Pin channels (𝑁) in the Effect Unit may be
different from the actual number of logical channels in the incoming Cluster (𝑛). However, 𝑁 ≥ 𝑛 shall always be
true.

AudioControls residing on currently unused Pin channels shall always remain accessible and retain their last setting
for as long as the Effect Unit remains in a Power State that requires this.

If the optional Bypass Control is present on a Pin channel, then engaging the bypass function shall result in passing
the corresponding logical channel of the incoming Cluster unaltered to that same logical channel of the outgoing
Cluster.

If the optional Bypass Control is present on the Primary channel, then engaging the bypass function shall result in
passing the incoming Cluster unaltered to the output of the Unit.

If the Effect Unit is an explicit member of a Power Domain, switching the Power Domain to any Power State other
than PS0 or PS1 shall render the Unit non-functional, and its output Cluster shall be Inactive.

3.13.9.1 PARAMETRIC EQUALIZER SECTION EFFECT UNIT

The Parametric Equalizer Section (PE) Effect Unit is used to manipulate and equalize the frequency characteristics
of the original audio information around a certain center frequency. To build a parametric equalizer, a number of
these PEQS Effect Units may need to be cascaded to obtain the desired functionality. The parameters that can be
manipulated to obtain the desired equalizing effect are:

• Center Frequency: the frequency around which the audio spectrum is manipulated. Expressed in Hz.
• Q Factor: a measure for the range of frequencies around the center frequency that are influenced. Expressed
as a ratio.
• Gain: the amount of gain or attenuation at the center frequency. Expressed in dB.

Release 4.0 April 2023 49

Universal Serial Bus Device Class Definition for Audio Devices

The algorithm to produce the desired equalization effect can be manipulated on a per-channel basis. The Primary
channel concept allows equalization for all channels simultaneously.

The symbol for the PE Processing Unit can be found in the following figure:
Figure 3-13: PEQS Effect Unit Icon

3.13.9.2 REVERBERATION EFFECT UNIT

The Reverberation (RV) Effect Unit is used to add room acoustics effects to the original audio information. These
effects may range from small room reverberation effects to simulation of a large concert hall reverberation. A set
of AudioControls can be manipulated to obtain the desired reverberation effects.

• Reverb Type: Room1 (small room), Room2 (medium room), Room3 (large room), Hall1 (medium concert hall),
Hall2 (large concert hall), Plate, Delay, and Panning Delay. This is a macro Control that chooses among
potentially different algorithms to obtain the desired reverberation effect. Changing this Control has no
impact on the value of the other AudioControls below. However, Host software may choose to reprogram the
Reverb Time Control (as an example) to a default value for the selected Reverb Type when the Reverb Type
Control is set to a different Reverb Type.
• Reverb Level: sets the amount of reverberant sound versus the original sound. Expressed as a ratio.
• Reverb Time: sets the time over which the reverberation will continue. Expressed in s.
• Reverb Delay Feedback: used with Reverb Types Delay and Delay Panning. Sets the way in which delay
repeats. Expressed as a ratio.
• Reverb Pre-Delay: sets the delay time between original sound and initial reverb reflection. Expressed in ms.
• Reverb Density: sets the density of the reverb reflections.
• Reverb Hi-Freq Roll-Off: sets the cut-off frequency of a low pass filter on the reflections. Expressed in Hz.

It is entirely left to the designer how a certain reverberation effect is obtained. It is not the intention of this
specification to precisely define all the parameters that influence the reverberation experience (for instance in a
multi-channel system, it is possible to create very similar reverberation impressions, using different algorithms and
parameter settings on all channels).

The algorithm to produce the desired equalization effect can be manipulated on a per-channel basis. The Primary
channel concept allows equalization for all channels simultaneously.

The symbol for the Reverberation Effect Unit can be found in the following figure:

Release 4.0 April 2023 50

Universal Serial Bus Device Class Definition for Audio Devices

Figure 3-14: Reverberation Effect Unit Icon

3.13.9.3 MODULATION DELAY EFFECT UNIT

The Modulation Delay (MD) Effect Unit is used to add modulation (like chorus) effects to the original audio
information. A number of parameters can be manipulated to obtain the desired modulation effects.

• Modulation Delay Balance: controls the ratio of the original sound to that of the effected sound. Expressed as
a ratio.
• Modulation Delay Rate: sets the speed (frequency) of the modulator. Expressed in Hz.
• Modulation Delay Depth: sets the depth at which the sound is modulated. Expressed in ms.
• Modulation Delay Time: sets the delay that is added to the modulated sound before adding it to the original
sound. Expressed in ms.
• Modulation Delay Feedback Level: controls the amount of the modulated sound that is routed back to the
input of the modulator unit. Expressed as a ratio.

It is entirely left to the designer how a certain modulation effect is obtained.

The algorithm to produce the desired equalization effect can be manipulated on a per-channel basis. The Primary
channel concept allows equalization for all channels simultaneously.

The symbol for the Modulation Delay Effect Unit can be found in the following figure:
Figure 3-15: Modulation Delay Effect Unit Icon

3.13.9.4 DYNAMIC RANGE COMPRESSOR/EXPANDER EFFECT UNIT

The Dynamic Range Compressor/Expander (DR) Effect Unit is used to intelligently compress, expand, or limit the
dynamic range of the original audio information. Several parameters can be manipulated to influence the desired
effect.

Note: Two Dynamic Range Compressor/Expander Effect Units may be used together for companding.

Release 4.0 April 2023 51

Universal Serial Bus Device Class Definition for Audio Devices

Figure 3-16: Dynamic Range Compressor/Expander Transfer Characteristic

Active Input Range dB Output

20

15

10

5
-25 -20 -15 -10 -5 dB

5 10 Input
-5

-10

-15

R=3 -20

R=2
-25
R=3/2
R=1 -30

R=2/3 -35

-40

-45

• Ratio R: determines the slope of the static input-to-output transfer characteristic in the effect’s active input
range. The effect is defined in terms of the ratio R, which is the inverse of the derivative of the output power
PO as a function of the input power PI when PO and PI are expressed in dB.
𝑃
𝜕𝐿𝑜𝑔(𝑃𝑂 )
−1 𝑅
𝑅 =
𝜕𝐿𝑜𝑔(𝑃𝐼 /𝑃𝑅 )
PR is the reference level, and it is made equal to the so-called line level. All levels are expressed relative to the
line level (0 dB), which is usually 15-20 dB below the maximum level. Compression is obtained when R > 1,
R = 1 does not affect the signal and R < 1 gives rise to expansion.
• Maximum Amplitude: the upper boundary of the active input range, relative to the line level (0 dB). Expressed
in dB.
• Threshold level: the lower boundary of the active input level, relative to the line level (0 dB).
• Attack Time: determines the response of the effect as a function of time to a step in the input level. Expressed
in ms.
• Release Time: relates to the recovery time of the gain of the compressor after audio is no longer within the
boundaries between Threshold and Maximum Amplitude. Expressed in ms.
• Make-up Gain: set to compensate for the gain loss in the effect. Expressed in dB.

It is entirely left to the designer how a certain dynamic range effect is obtained.

Release 4.0 April 2023 52

Universal Serial Bus Device Class Definition for Audio Devices

The algorithm to produce the desired equalization effect can be manipulated on a per-channel basis. The Primary
channel concept allows equalization for all channels simultaneously.

The symbol for the Dynamic Range Compressor/Expander Effect Unit can be found in the following figure:
Figure 3-17: Dynamic Range Compressor/Expander Effect Unit Icon

3.13.10 PROCESSING UNIT

The Processing Unit (PU) represents a functional block inside the Audio Function that transforms one or more
logical input channels into one or more logical output channels. The input channels are grouped into one or more
Clusters. Each Cluster enters the Processing Unit through one of the 𝑝 Input Pins. The logical output channels are
grouped into one Cluster and leave the Processing Unit through a single Output Pin.

The Pin channels on each Input Pin 𝑖 of the Processing Unit are numbered from one to 𝑁𝑖 . The PCC value 𝑁𝑖 for
each of the Input Pins is inherited from the first upstream Entity that defines a PCC value on its Output Pin.

On each Input Pin 𝑖, the incoming Cluster logical channels, numbered from one to 𝑛𝑖 , are mapped one-to-one onto
the Pin channels on that Input Pin, i.e. for each Input Pin 𝑖, Cluster channel 1 is mapped onto Pin channel 1, Cluster
channel 2 is mapped onto Pin channel 2, and so on, up to Cluster channel 𝑛𝑖 which is mapped onto Pin channel 𝑛𝑖 .
Note that the number of Pin channels on each of the Processing Unit’s Input Pins (𝑁𝑖 ) may be different from the
actual number of channels in the incoming Clusters (𝑛𝑖 ). However, for each Input Pin 𝑖, 𝑁𝑖 ≥ 𝑛𝑖 shall always be
true.

AudioControls residing on currently unused channels shall always remain accessible and retain their last setting for
as long as the Processing Unit remains in a Power State that requires this.

This specification defines several standard transforms (algorithms) that are considered necessary to support
additional Audio functionality; these transforms are not covered by the other Unit types but are commonplace
enough to be included in this specification so that a generic driver can provide control for it.

A Processing Unit shall implement at least one of the defined algorithms.

If the optional Bypass Control is present, then engaging the bypass function shall result in passing the incoming
Cluster on Input Pin one (the dominant Input Pin) unaltered to the output of the Unit. If it is necessary to be able
to bypass the Processing Unit’s functionality but providing an output Cluster different from the input Cluster on
Input Pin one, then an explicit bypass topology using a Selector Unit should be implemented.

The Processing Unit can expose multiple Cluster Configurations on its Output Pin. The Cluster Control is typically
implemented as Read-Write (rw) and is used by Host software to select the desired Cluster Configuration on the
Output Pin. In other words, the selection of the Cluster Configuration controls the operational mode of the
Processing Unit. A Cluster Control value of zero indicates that the output Cluster Configuration is inherited from
the current Cluster on Input Pin one (the dominant Input Pin). The Read-Only Cluster Active Control Shall always be
implemented and indicates whether the output Cluster is currently Active or not.

Release 4.0 April 2023 53

Universal Serial Bus Device Class Definition for Audio Devices

Generally, the Processing Unit redefines the channels in its output Cluster as a processed result of any or all of the
incoming channels. It is therefore appropriate to treat these channels as independent of all other channels in the
Audio Function and assign them unique IDs in their respective wChannelID fields in the output Cluster. However, in
subsequent sections, some guidelines are provided on how to manage Channel IDs for the different types of
Processing Units, defined by this specification.

If the Processing Unit is an explicit member of a Power Domain, then switching the Power Domain to any Power
State other than PS0 or PS1 shall render the Unit non-functional, and its output Cluster shall be Inactive.

3.13.10.1 UP/DOWN-MIX PROCESSING UNIT

The Up/Down-mix (UD) Processing Unit provides facilities to derive m output audio channels from n input audio
channels. The algorithms and transforms applied to accomplish this are not defined by this specification and may
be proprietary. The input channels are grouped into one or more Clusters. Each Cluster enters the Processing Unit
through one of the 𝑝 Input Pins. The logical output channels are grouped into one Cluster and leave the Up/Down-
mix Processing Unit through a single Output Pin.

The Up/Down-mix Processing Unit may support multiple modes of operation. The logical input channels in the
incoming Clusters are defined by Entities in the upstream audio path to which the Input Pins of the Up/Down-mix
Processing Unit are connected. The Up/Down-mix Processing Unit Descriptor reports which up/down-mixing
modes the Unit supports through its waClusterDescrID() array. Each element of the waClusterDescrID () array
indicates which output channels in the output Cluster are effectively present in a particular mode. Mode selection
is accomplished by selecting an output Cluster through the Cluster Control.

As an example, consider the case where an Up/Down-mix Processing Unit is connected to the Input Terminal,
producing Dolby AC-3 5.1 decoded audio. The input Cluster to the Up/Down-mix Processing Unit therefore
contains Front Left, Front Right, Front Center, Surround Array Left, Surround Array Right (Left Surround and Right
Surround in CEA-861.2 parlance), and LFE logical channels.

Suppose the Audio Function’s hardware is limited to reproducing only dual channel audio. Then the Up/Down-mix
Processing Unit could use some (sophisticated) algorithms to down-mix the available spatial audio information into
two (‘enriched’) channels so that the maximum spatial effects can be experienced, using only two channels. It is
left to the implementation to use the appropriate down-mix algorithm depending on the physical nature of the
Output Terminal to which the Up/Down-mix Processing Unit is eventually routed. For instance, a different down-
mix algorithm may be needed whether the ‘enriched’ stereo stream is sent to a pair of speakers or to a headset.
However, this knowledge already resides within the Audio Function and deciding which down-mix algorithm to use
does not need Host intervention.

As a second interesting example, suppose the hardware is capable of servicing eight discrete audio channels (for
example, a full-fledged MPEG-2 7.1 system). Now the Up/Down-mix Processing Unit could use certain techniques
to derive meaningful content for the extra audio channels (Front Left of Center, Front Right of Center) that are
present in the output Cluster and are missing in the input channel Cluster (AC-3 5.1). This is a typical example of an
up-mix situation.

Since the Up/Down-mix Processing Unit redefines the channels in its output Cluster as a processed result of any or
all of the incoming channels, it is appropriate to treat these channels as independent of all other channels in the
Audio Function and therefore assign them unique IDs in their respective wChannelID fields in the output Cluster.

The symbol for the Up/Down-mix Processing Unit is depicted in the following figure:

Release 4.0 April 2023 54

Universal Serial Bus Device Class Definition for Audio Devices

Figure 3-18: Up/Down-mix Processing Unit Icon

3.13.10.2 CHANNEL REMAP PROCESSING UNIT

The Channel Remap (CR) Processing Unit provides facilities to derive m output audio channels from n input audio
channels. It creates an output Cluster that contains a subset of the input channels, effectively acting as a simple
channel selection mechanism with no audio processing involved. The input channels are grouped into one or more
Clusters. Each Cluster enters the Processing Unit through one of the 𝑝 Input Pins. The logical output channels are
grouped into one Cluster and leave the Channel Remap Processing Unit through a single Output Pin.

The logical input channels in the incoming Clusters are defined by Entities in the upstream audio path to which the
Input Pins of the Channel Remap Processing Unit are connected. The Channel Remap Processing Unit may support
multiple channel remapping modes. The Channel Remap Processing Unit Descriptor reports which channel
remapping modes the Unit supports through its waClusterDescrID() array. Each element of the waClusterDescrID()
array indicates which output channels in the output Cluster are effectively present in a particular remapping mode.
Mode selection is accomplished by selecting an output Cluster through the Cluster Control.

Since the Channel Remap Processing Unit does not redefine the channels in its output Cluster but merely selects
channels from the incoming channels, it shall preserve the Channel IDs associated with those incoming channels as
they are bundled to create the outgoing Cluster.

The symbol for the Channel Remap Processing Unit is depicted in the following figure:
Figure 3-19: Channel Remap Processing Unit Icon

3.13.10.3 STEREO EXTENDER PROCESSING UNIT

The Stereo Extender (ST) Processing Unit operates on Front Left and Front Right channels only. It processes the
Front Left and Right from the incoming Cluster to expand the sound field and to make it appear to originate from
outside the Front Left/Right speaker locations. Extended stereo effects may be achieved via various methods. All
other channels in the Cluster are passed through without modification. The algorithms and transforms applied to
accomplish the stereo widening are not defined by this specification and may be proprietary. The perceived width
of the sound field can be controlled via the Width Control.

Since the Stereo Extender Processing Unit does not redefine any of the channels in its output Cluster, it shall
preserve the Channel IDs associated with those incoming channels.

The symbol for the Stereo Extender Unit is depicted in the following figure:

Release 4.0 April 2023 55

Universal Serial Bus Device Class Definition for Audio Devices

Figure 3-20: Stereo Extender Processing Unit Icon

3.13.10.4 MULTI-FUNCTION PROCESSING UNIT

The Multi-Function (MF) Processing Unit groups different but related algorithmic blocks that together provide a
certain functionality. This specification defines several algorithms that are considered useful and commonplace
enough to be included in this specification.

The mandatory Read-Only (r) AlgoPresent Control returns a bitmap indicating what types of algorithms are
performed inside the Multi-Function Processing Unit. The following algorithms are currently supported:

• Algorithm Undefined
• Beam Forming Algorithm
• Acoustic Echo Cancellation Algorithm
• Active Noise Cancellation Algorithm
• Blind Source Separation Algorithm
• Noise Suppression/Reduction

The exact implementation of these algorithms and how they may interact is implementation dependent.

Note 1: If there is a need to expose to the Host how the algorithms are interconnected, a designer may
choose to model the assembly of algorithms using multiple Multi-Function Processing Units, each
containing just one or a subset of algorithms and explicitly connecting them together.

Note 2: Most of the algorithms mentioned above involve some form of signal processing that cannot be
assumed to be linear and time invariant.

Note that support for an algorithm may change dynamically (due to Audio Function resource reallocation, for
example). In this case, the AlgoPresent Control shall reflect the new situation and generate an interrupt to inform
the Host of the change.
The optional AlgoEnable Control allows to selectively enable or disable the implemented algorithms.

During normal operation, the Multi-Function Processing Unit transforms one or more logical input channels into
one or more logical output channels. The input channels are grouped into one or more Clusters. Each Cluster
enters the Processing Unit through one of the 𝑝 Input Pins. The logical output channels are grouped into one
Cluster and leave the Processing Unit through a single Output Pin.

Since the Multi-Function Processing Unit redefines the channels in its output Cluster as a processed result of any or
all the incoming channels, it is appropriate to treat these channels as independent of all other channels in the
Audio Function and therefore assign them unique IDs in their respective wChannelID fields in the output Cluster.

A Multi-Function Processing Unit shall implement the Bypass Control so that a generic audio driver that does not
understand what functionality is implemented in the Multi-Function Processing Unit will be capable of removing it
from the signal path.

Release 4.0 April 2023 56

Universal Serial Bus Device Class Definition for Audio Devices

The symbol for the Multi-Function Processing Unit is depicted in the following figure:
Figure 3-21 Multi-Function Processing Unit Icon

3.13.11 EXTENSION UNIT

The Extension Unit (XU) is the method provided by this specification to easily add vendor-specific building blocks to
the specification. The functionality and behavior of the Extension Unit is identified by a Globally Unique Identifier
(GUID) that is included in the Extension Unit Descriptor.

Note: This GUID is not used to identify instances of a Device. Rather, the same GUID is used in all
implementations that incorporate this Extension Unit with this functionality and behavior, requiring
the same vendor-defined software to operate.

The Extension Unit provides vendor-defined functionality inside the Audio Function that transforms one or more
logical input channels into one or more logical output channels. The input channels are grouped into one or more
Clusters. Each Cluster enters the Extension Unit through one of the 𝑝 Input Pins. The logical output channels are
grouped into one Cluster and leave the Extension Unit through a single Output Pin.

The Pin channels on each Input Pin 𝑖 of the Extension Unit are numbered from one to 𝑁𝑖 . The PCC value 𝑁𝑖 for each
of the Input Pins is inherited from the first upstream Entity that defines a PCC value on its Output Pin.

On each Input Pin 𝑖, the incoming Cluster logical channels, numbered from one to 𝑛𝑖 , are mapped one-to-one onto
the Pin channels on that Input Pin, i.e. for each Input Pin 𝑖, Cluster channel 1 is mapped onto Pin channel 1, Cluster
channel 2 is mapped onto Pin channel 2, and so on, up to Cluster channel 𝑛𝑖 which is mapped onto Pin channel 𝑛𝑖 .
Note that the number of Pin channels on each of the Extension Unit’s Input Pins (𝑁𝑖 ) may be different from the
actual number of channels in the incoming Clusters (𝑛𝑖 ). However, for each Input Pin 𝑖, 𝑁𝑖 ≥ 𝑛𝑖 shall always be
true.

AudioControls residing on currently unused Pin channels shall always remain accessible and retain their last setting
for as long as the Extension Unit remains in a Power State that requires this.
If the Bypass Control is present, then engaging the bypass function shall result in passing the incoming Cluster on
Input Pin one (the dominant Input Pin) unaltered to the output of the Unit. If it is necessary to be able to bypass
the Extension Unit’s functionality but providing an output Cluster different from the input Cluster on Input Pin one,
then an explicit bypass topology using a Selector Unit should be implemented.

An Extension Unit shall implement the Bypass Control so that a generic audio driver that does not understand
what functionality is implemented in the Extension Unit will be capable of removing it from the signal path.

The Extension Unit can expose multiple Cluster Configurations on its Output Pin. The Cluster Control will typically
be implemented as Read (r) and informs Host software which Cluster Configuration is currently active. In this case,
the currently active Cluster is determined by the vendor-defined internal operation of the Extension Unit.

However, the Cluster Control may be implemented as Read-Write (rw) and may be used by Host software to select
the desired Cluster Configuration on the Output Pin. In other words, the selection of the Cluster Configuration

Release 4.0 April 2023 57

Universal Serial Bus Device Class Definition for Audio Devices

drives the vendor-defined internal operation of the Extension Unit based on the Cluster Configuration that the
Host Software requires at the Output Pin of the Extension Unit. A Cluster Control value of zero indicates that the
output Cluster is inherited from the current Cluster on Input Pin one (the dominant Input Pin). The Read-Only
Cluster Active Control Shall always be implemented and indicates whether the output Cluster is currently Active or
not.

Since the Extension Unit likely redefines the channels in its output Cluster as a processed result of any or all the
incoming channels, it would be appropriate to treat these channels as independent of all other channels in the
Audio Function and therefore assign them unique IDs in their respective wChannelID fields in the output Cluster.

If the Extension Unit is an explicit member of a Power Domain, then switching the Power Domain to any Power
State other than PS0 or PS1 shall render the Unit non-functional, and its output Cluster shall be Inactive.

The symbol for the Extension Unit can be found in the following figure:
Figure 3-22: Extension Unit Icon

XU

3.13.12 CLOCK ENTITIES

Clock Entities are special in the sense that they do not directly manipulate logical audio streams. Instead, they
provide the functionality needed to manipulate sampling clock signals and clock routing for the different Input and
Output Terminals and Sampling Rate Converter Units within the Audio Function. A Terminal inside the Audio
Function shall only be connected to one Clock Entity. The clock signal present at the Clock Input Pin of a Terminal
or SRC Unit determines the sampling frequency at which the underlying hardware is operating. If there is no need
for the Audio Function to expose any clock information related to a Terminal or (part of) an SRC Unit, the Clock
Input Pin of the Terminal or SRC Unit may be left unconnected.

3.13.12.1 CLOCK SOURCE

A Clock Source Entity provides a sampling clock signal on its single Clock Output Pin that is frequency-locked to the
Reference Clock of the Clock Domain of which the Clock Source Entity is part. The actual output frequency of the
Clock Source Entity can be different from the frequency of the Clock Domain’s Reference Clock. Multiple Clock
Source Entities can be part of the same Clock Domain, all sharing the same Reference Clock. Although they all may
have a different output frequency, their output clock signals are all frequency-locked to one another since they are
all derived from the same Reference Clock. By interacting with the Sampling Frequency Control inside the each
Clock Source Entity, Host software can retrieve or alter the current output sampling frequency of the Clock Source
Entity. Note that the Reference Clock frequency of the Clock Domain that is used by a Clock Source Entity is not
exposed to the Host.

All different sampling clocks used inside the Audio Function shall be represented by separate Clock Source Entities.
Even if the clock is generated ‘inside a Terminal’, that clock needs to be represented by a Clock Source Entity. As an
example, a sampling clock could be recovered based on the amount of audio samples coming into the Audio
Function over a USB OUT adaptive Endpoint. Alternatively, a sampling clock may be derived from the S/PDIF signal
coming into the Audio Function on an external connector.

Release 4.0 April 2023 58

Universal Serial Bus Device Class Definition for Audio Devices

Note: In the case of an adaptive isochronous data Endpoint that support only a discrete number of sampling
frequencies, the Endpoint shall at least tolerate 1000 PPM inaccuracy on the reported Sampling
Frequency Control values to accommodate sample clock inaccuracies.

The Clock Source Entity descriptor contains a field that indicates the Clock Domain of which the Clock Source Entity
is part. Furthermore, since Input and Output Terminals only have a Clock Input Pin, a clock signal shall never be
generated from a Terminal directly.

The output of a Clock Source Entity does not have to be always valid. For instance, if a Clock Source Entity
represents an external sampling clock input on the Audio Function, the output of that Clock Source may not be
valid when there is nothing connected to the external clock input. The Clock Source can always be queried for the
validity of its output signal.

The symbol for the Clock Source Entity can be found in the following figure:
Figure 3-23: Clock Source Icon

3.13.12.2 CLOCK SELECTOR

A Clock Selector Entity provides the functionality to select among different available sampling clock signals. It shall
have two or more Clock Input Pins from which one is routed to the single Clock Output Pin.

Switching between Clock Inputs may be Host controlled (the Clock Selector’s Selector Control is programmable via
the appropriate Push Command) or the Audio Function may switch Clock Inputs due to some external event. A
Clock Selector may support both control methods. The Selector Control can notify the Host of the change by
generating an interrupt.

The symbol for the Clock Selector Entity can be found in the following figure:
Figure 3-24: Clock Selector Icon

3.13.13 CONNECTOR ENTITIES

Connector Entities are not explicitly exposed in the Audio Function topology. A Connector, or rather the signal(s) it
carries is represented inside the Audio Function via the associated Input or Output Terminal. The Connector Entity
provides optional AudioControls that advertise the insertion state of the Connector, and potentially also
information about the nature and characteristics of the external device that is currently plugged into the
Connector.

3.13.14 POWER DOMAINS AND POWER DOMAIN ENTITIES

Power Domains and their associated Power Domain Entities are not explicitly exposed in the Audio Function
topology. Rather, a Power Domain bundles Entities whose Power States are managed together. The associated

Release 4.0 April 2023 59

Universal Serial Bus Device Class Definition for Audio Devices

Power Domain Entity provides AudioControls to impact the power behavior of the Power Domain through
Commands, addressed to the Power Domain Entity.

3.13.15 CHANNELGROUPS, ENTITYGROUPS, AND COMMITGROUPS

ChannelGroups, EntityGroups and CommitGroups are not explicitly exposed in the Audio Function topology. A
ChannelGroup shall have only channels as its Members. An EntityGroup shall have only Entities as its Members. A
CommitGroup shall have only AudioControls as its Members.

3.14 OPERATIONAL MODEL

A Device may support multiple configurations. Within each configuration there may be multiple interfaces, each
possibly having Alternate Settings. These interfaces may pertain to different functions that co-reside in the same
composite Device. Even several independent Audio Functions may exist in the same Device. Interfaces belonging to
the same Audio Function are grouped into an Audio Interface Association (AIA). If the Device contains multiple
independent Audio Functions, there shall be multiple Audio Interface Associations, each providing full access to
their associated Audio Function.

As an example of a composite Device, consider a computer display equipped with a built-in stereo speaker system.
Such a Device could be configured to have one interface dealing with configuration and control of the monitor part
of the Device (HID Class), while an Association of two other interfaces deals with its audio aspects. One of those,
the AudioControl Interface, is used to control the inner workings of the function ( Control etc.) whereas the other,
the AudioStreaming Interface, handles the data traffic, sent to the monitor’s audio subsystem.

The AudioStreaming Interface could be configured to operate in mono mode (Alternate Setting x) in which only a
single channel data stream is sent to the Audio Function. The receiving Input Terminal would then output a mono
cluster that could feed into an Up/Down-mix Processing Unit that duplicates this mono audio stream into two
logical channels at its Output Pin, and those could then be reproduced on both speakers. From an interface point
of view, such a setup requires one isochronous Endpoint in Alternate Setting x of the AudioStreaming Interface to
receive the mono audio data stream, in addition to the mandatory control Endpoint and optional interrupt
Endpoint in the AudioControl Interface.

The same system could be used to play back stereo audio. In this case, the stereo AudioStreaming Interface is
selected (Alternate Setting y). This Interface also consists of a single isochronous Endpoint, now receiving a data
stream that interleaves Front Left and Front Right channel samples. The receiving Input Terminal then splits the
stream into a Front Left and Front Right logical channel and outputs a stereo Cluster that feeds into the Up/Down-
mix Processing Unit. Rather than duplicating the mono channel, the Processing Unit now simply passes the
incoming stereo Cluster unaltered to its Output Pin. From an interface point of view, this setup requires one
isochronous Endpoint in Alternate Setting y of the AudioStreaming Interface to receive the stereo audio data
stream. The AudioControl Interface Alternate Setting remains unchanged.

If the above AudioStreaming Interface were an asynchronous sink, one extra isochronous Feedback Endpoint
would also be necessary.

As stated earlier, Audio functionality is located at the interface level in the Device Class hierarchy. The following
sections describe the Audio Interface Association, containing a single AudioControl Interface and optional
AudioStreaming Interfaces, together with their associated Endpoints that are used for Audio Function control and
for audio data stream transfer.

Release 4.0 April 2023 60

Universal Serial Bus Device Class Definition for Audio Devices

3.14.1 AUDIOCONTROL INTERFACE

To control the functional behavior of a particular Audio Function, the Host manipulates the Entities inside the
Audio Function. To make these objects accessible, the Audio Function shall expose a single AudioControl Interface.
This Interface may contain the following Endpoints:

• A control Endpoint for manipulating Entity Control settings and retrieving the state of the Audio Function. This
Endpoint is mandatory, and the default Endpoint 0 is used for this purpose.
• An interrupt Endpoint. The Endpoint is optional but shall be implemented if any of the AudioControls inside
the Device have the need to generate and interrupt to notify the Host of a change in the Audio Function’s
behavior.

The AudioControl Interface is the only entry point to access the internals of the Audio Function. All Commands that
are concerned with the manipulation of AudioControls within the Audio Function’s Entities shall be directed to the
AudioControl Interface of the Audio Function. Likewise, all Descriptors related to the internals of the Audio
Function are part of the class-specific AudioControl Interface Descriptor.

The AudioControl Interface of an Audio Function shall only support a single Alternate Setting (Alternate Setting 0).

3.14.1.1 CONTROL ENDPOINT

The audio interface class uses Endpoint 0 (the default pipe) as the standard way to control the Audio Function
using class-specific Commands. These Commands are always directed to one of the Entities that make up the
Audio Function. The format and contents of these Commands are detailed further in this document.

3.14.1.2 INTERRUPT ENDPOINT

A USB AudioControl Interface may support an optional interrupt Endpoint to inform the Host about dynamic
changes that occur on the different addressable Entities inside the Audio Function. The interrupt Endpoint is used
by the entire Audio Interface Association to convey change information to the Host. It is considered part of the
AudioControl Interface because this is the anchor interface for the AIA.

3.14.2 AUDIOSTREAMING INTERFACE

AudioStreaming Interfaces are used to interchange digital audio data streams between the Host and the Audio
Function. They are optional. An Audio Function may have zero or more AudioStreaming Interfaces associated with
it, each possibly carrying data of a different nature and format. Each AudioStreaming Interface shall have at most
one isochronous data Endpoint. This construction guarantees a one-to-one relationship between the
AudioStreaming Interface and the single audio data stream, related to the Endpoint. In some cases, the
isochronous data Endpoint is accompanied by an associated isochronous explicit feedback Endpoint for
synchronization purposes. The isochronous data Endpoint and its associated feedback Endpoint shall follow the
Endpoint numbering scheme as set forth in the USB Core Specifications.

An AudioStreaming Interface may have Alternate Settings that can be used to change certain characteristics of the
Interface and its underlying Endpoint. A typical use of Alternate Settings is to provide a way to change the
subframe size and/or number of channels on an active AudioStreaming Interface. Whenever an AudioStreaming
Interface requires an isochronous data Endpoint, it shall at least provide the default Alternate Setting (Alternate
Setting 0) with zero bandwidth requirements (no isochronous data Endpoint defined) and one additional Alternate
Setting that contains the actual isochronous data Endpoint. All non-zero Alternate Settings of an AudioStreaming
Interface shall use the same data Endpoint and explicit feedback Endpoint, if present. More specifically, the data
Endpoint shall use the same Endpoint number for all non-zero Alternate Settings of the Interface. Any Alternate

Release 4.0 April 2023 61

Universal Serial Bus Device Class Definition for Audio Devices

Setting which has an explicit feedback Endpoint shall use the same Endpoint number in all non-zero Alternate
Settings.

The class-specific AudioStreaming Interface Descriptor contains two fields (wStartDelayUnits and wStartDelay
field) that together indicate how much time it takes this Interface to reliably produce valid outgoing data (i.e., valid
audio sample data and valid packet sizes – see Section 7.2.1.2.1, “Service Interval Packet Size Calculation”) or
effectively consume incoming data. This time is measured using as a reference the first IN or OUT PID that occurs
at the start of an audio stream.

For proper streaming operation, it is highly recommended that the Host ignores all samples received between
issuing the first IN PID and the expiration of the start delay time (as indicated by the wStartDelay field value) on
input. Likewise, on output, the Host should send silence for at least the indicated start delay time after issuing the
first OUT PID to avoid any undesired artifacts.

Switching from one active Alternate Setting to another active Alternate Setting on an AudioStreaming Interface
shall never be performed directly. Instead, the Host software needs to first switch the Interface to its (mandatory)
inactive Alternate Setting 0 and then, in a second step, switch the AudioStreaming Interface to the newly desired
active Alternate Setting. The Device shall generate a Request Error on any standard SET_INTERFACE request that
attempts to switch the Interface from one active Alternate Setting to a different active Alternate Setting.

For every defined AudioStreaming interface, there shall be a corresponding Input or Output Terminal defined in
the Audio Function. For the Host to fully understand the nature and behavior of the connection, it needs to
consider the Interface- and Endpoint-related Descriptors as well as the Terminal-related Descriptor.

Note: Interfaces and Endpoints are USB Core Specification concepts and therefore use a Host-centric
terminology. Terminals are Audio Device Class concepts and therefore use Audio Function-centric
terminology. OUT Interfaces and Endpoints correspond to Input Terminals, and IN Interfaces and
Endpoints correspond to Output Terminals.

3.14.2.1 ISOCHRONOUS AUDIO DATA STREAM ENDPOINT

In general, the data streams that are handled by an isochronous audio data Endpoint do not necessarily map
directly to the logical channels that exist within the Audio Function. As an example, consider the case where
multiple logical audio channels are compressed into a single data stream (AC-3, WMA …). The format of such a
data stream may be entirely different from the original format of the logical channels (for example, 640 kbits/s AC-
3 5.1 audio as opposed to 6-channel 16-bit 44.1 kHz audio). Therefore, to describe the data transfer at the
Endpoint level correctly, the notion of logical channel is replaced by the notion of audio data stream. It is the
responsibility of the AudioStreaming Interface which contains the OUT Endpoint to convert between the audio
data stream and the embedded logical channels before handing the data over to the Input Terminal. In many
cases, this conversion process involves some form of decoding. Likewise, the AudioStreaming Interface which
contains the IN Endpoint shall convert logical channels from the Output Terminal into an audio data stream, often
using some form of encoding.

Requests to control properties that exist within an Audio Function, such as or Mute cannot be sent to the
Endpoint in an AudioStreaming Interface. An AudioStreaming Interface operates on audio data streams and is
unaware of the number of logical channels it eventually serves. Instead, these Commands shall be directed to the
proper Audio Function Entities via the AudioControl Interface.

As already mentioned, an AudioStreaming Interface may have zero or one isochronous audio data Endpoint. If
multiple synchronous audio channels need to be communicated between Host and Audio Function, they shall be

Release 4.0 April 2023 62

Universal Serial Bus Device Class Definition for Audio Devices

clustered into one physical Cluster by interleaving the individual audio data, and the result can be directed to the
single Endpoint.

If an Audio Function needs more than one Cluster to operate, each Cluster is directed to the Endpoint of a separate
AudioStreaming Interface, belonging to the same Audio Interface Association (all servicing the same Audio
Function).

3.14.2.2 ISOCHRONOUS FEEDBACK ENDPOINT

For adaptive audio source Endpoints and asynchronous audio sink Endpoints, an explicit synchronization
mechanism is needed to maintain synchronization during transfers. For details about synchronization at different
Endpoint speeds, see the applicable USB Core Specifications.

3.14.2.3 AUDIO DATA FORMAT

The format used to transport audio data over the USB is entirely determined by the value in the wFormat field of
the class-specific AudioStreaming Self Descriptor. Some additional fields in this Descriptor further describe the
format. For details about the defined Format Types and associated data formats, see Section 7, “Audio Data
Formats.”

3.14.3 AUDIOCONTROLS
Inside an Entity, functionality is described through AudioControls. Each AudioControl provides access to a specific
audio characteristic, such as , Bass, etc.

Each Entity shall only contain a specific set of AudioControls, permitted for use by that Entity as defined by this
specification.

It is important to note that all interactions with Entities in the Audio Function are performed via AudioControls. All
accessible parameters inside these Entities are modeled using the concept of the AudioControl and its associated
Attributes.

AudioControls are either associated with a particular Pin channel inside an Entity or influence the behavior of the
Entity as a whole.

Each AudioControl has a set of Attributes that can be manipulated or that present information about the behavior
of the AudioControl. An AudioControl has the following Attributes:

• Current Attribute (CUR) – Mandatory.

• Next Attribute (NEXT) – Optional.
• Capabilities Attribute (CAP) – Mandatory.
• Range Attribute (RANGE) – Optional.

Attributes may be implemented as Read[-Only] (r), Read-Write (rw) and even Write[-Only] (w).

An Attribute of an AudioControl provides the finest level of addressable control granularity within the Audio
Function. See Section 5.3.2, “AudioControl Commands” for information on how to access a particular
AudioControl.

3.14.3.1 CLUSTER-WIDE AUDIOCONTROLS

AudioControls are often implemented on a per-Pin-channel basis. However, to control the settings of an
AudioControl as a whole, the concept of a Cluster-wide AudioControl has been defined. Changing the value of a

Release 4.0 April 2023 63

Universal Serial Bus Device Class Definition for Audio Devices

Cluster-wide AudioControl applies that change to all Pin channels simultaneously. Note that a Cluster-wide
AudioControl shall always be implemented separate from and independent of the per-Pin-channel AudioControls.
Changing the setting of a Cluster-wide AudioControl shall not affect the settings of any of the individual Pin
channel AudioControl settings. The following figure illustrates the concept.
Figure 3-25: N-channel AudioControl with Cluster-wide AudioControl

Pin Ch 1

Pin Ch 2

Pin Ch n

Pin Ch 0

Per-Pin channel Cluster-wide

AudioControls AudioControl

3.14.3.2 SYNCHRONIZING MULTIPLE AUDIOCONTROLS

In some cases, multiple AudioControls need to be manipulated simultaneously to achieve a desired effect or to
avoid unwanted transitory side effects. To accomplish this, every AudioControl may have a Next Attribute that is
used to preload a certain value for the AudioControl. Only when a Commit Command is issued (to the Audio
Function as a whole or to a CommitGroup) will all AudioControls in the Audio Function or those in the
CommitGroup take on as their Current Attribute setting the value that was preloaded in their Next Attribute. This
effectively provides a means to have several AudioControls (including those present in the AudioStreaming
Interfaces and all other Entities) take on new values simultaneously. Note further that any AudioControls in the
entire Audio Function may be synchronized using this technique by properly interleaving sets of Push(NEXT)
Commands with Commit Command sequences.

For more detailed information on AudioControl and Attribute manipulation, see Section 5, “Commands &
Requests.”

3.14.4 CLOCK MODEL

Clock Entities provide a way to accurately describe the use and distribution of sampling clock signals throughout
the Audio Function. Sampling frequencies inside the Audio Function may only be influenced by directly interacting
with the Sampling Frequency Control inside a Clock Source Entity. The Sampling Frequency Control RANGE
Attribute provides the necessary information for Host software to determine what sampling frequencies the
AudioControl (and thus the Clock Source Entity) supports.

A side effect of changing the sampling frequency could be that certain AudioStreaming Interfaces may need to
switch to a different Alternate Setting to support the bandwidth needed for the new sampling frequency. This
specification does not allow an AudioStreaming Interface to switch from one Alternate Setting to another on its
own except to change to Alternate Setting zero, which is the idle setting. Instead, when the Audio Function detects
that it can no longer support a certain Alternate Setting on an AudioStreaming Interface, it shall switch to Alternate
Setting zero on that Interface and report the change to Host software through the Active Alternate Setting Control

Release 4.0 April 2023 64

Universal Serial Bus Device Class Definition for Audio Devices

interrupt. The Host can then query the Interface for new valid Alternate Settings for the Interface through the Get
Valid Alternate Settings Control Command and make an appropriate selection.

Note: To keep the number of Alternate Settings in an AudioStreaming Interface to a minimum, it is not
recommended to provide a separate Alternate Setting for every supported sampling frequency. A few
Active Alternate Settings (low bandwidth, medium bandwidth, high bandwidth) may be enough to
provide reasonable bandwidth control.

Audio streams can be bridged from one Clock Domain to another using the Sampling Rate Converter Unit.

3.14.5 CONNECTOR MODEL

Each Connector has its own Connector Entity and is identified by a unique Connector ID. The Connector Entity
optionally has an embedded Insertion Detect Control that indicates through its CUR Attribute whether there is
currently an external device plugged into the Connector that carries information related to the Terminal to which
the Connector is associated.

It is important to note that a Connector and its associated Connector Entity in the context of this specification may
represent only part of a physical external connector on the Device. For example, an external physical connector
that can accept both headphones and headsets would be represented by two distinct Connectors and Connector
Entity pairs, where one Connector and its Connector Entity would be associated with an Input Terminal (for the
microphone part) and another Connector and its Connector Entity would be associated with an Output Terminal
(for the speaker part). Each corresponding Connector Entity could have its own independent Insertion Detect
Control. One would detect the insertion of the microphone part of a headset, whereas the other would detect the
insertion of the speaker part of either a headphone or a headset.

The Insertion Detect Control change generates an interrupt whenever the Audio Function autonomously detects a
change in insertion state.

A Connector Entity is always associated with one or more Input or Output Terminals through which the signals,
carried over the Connector enter or leave the Audio Function. Multiple Connector Entities may be associated with
the same Terminal. This indicates that there is a functional relationship among the Connectors, and they should be
considered conceptually as a whole but with an implementation that requires multiple physical plugs or
receptacles at the same time. A typical example of this is a set of three 3.5 mm receptacles that are used to
connect a 5.1 surround-capable speaker set to the Audio Device. One Connector carries Front Left/Right signals,
the second carries Surround Left/Right signals, while the third one carries Center/LFE signals typically.

3.14.6 POWER DOMAIN MODEL

Managed Power Domain support is optional for an Audio Function. The Audio Function has zero or more Power
Domains, each represented by a Power Domain Entity that is identified by a unique Power Domain ID. Each Power
Domain can be individually manipulated by the Host to optimize overall power consumption. If implemented, the
Audio Function contains a set of Power Domain Entities with their embedded AudioControls that allows the Host
to set Power States for each individual Power Domain inside the Audio Function. The Audio Function contains as
many Power Domain Entities as there are Power Domains and access to a specific Power Domain Control is based
on the Power Domain Entity ID. A Power Domain contains one or more member Entities. Power Domain Entities
shall never be a member of a Power Domain. An Entity shall not be a member of more than one Power Domain.

A Power Domain shall support five Power States PS0 to PS4.

Power State PS0 is the fully operational state and shall be the default Power State for all Power Domains.

Release 4.0 April 2023 65

Universal Serial Bus Device Class Definition for Audio Devices

Power State PS1 is a functional state where power consumption shall be no higher than in Power State PS0. Audio
fidelity may be reduced in this Power State. All AudioControls and bypass functionality residing in Entities that are
members of the Power Domain shall remain operational and memory content such as downloaded algorithms shall
be preserved.

Power State PS2 is a state where audio streaming over USB Endpoints shall not occur, and power consumption
shall be no higher than in Power State PS1. All AudioControls and bypass functionality residing in Entities that are
members of the Power Domain shall remain operational and memory content such as downloaded algorithms or
buffered audio shall be preserved.

Power State PS3 is a state where audio streaming on the Output Pins of the member Entities shall not occur unless
the optional Bypass functionality is engaged. Power consumption shall be no higher than power consumption in
Power State PS2. All AudioControls and bypass functionality residing in member Entities shall remain operational.
Memory content is not guaranteed to be preserved. Any Entity that is affected by the loss of memory may revert
to unspecified values for its AudioControls when transitioning to a higher Power State.

Power State PS4 is a non-functional state where audio streaming on the Output Pins of the member Entities shall
not occur, and power consumption shall be no higher than power consumption in Power State PS3. AudioControls
and bypass functionality residing in member Entities shall become non-operational and may lose state. Memory
content is likely to be lost. Any Entity that is affected by the loss of memory may revert to unspecified values for its
AudioControls when transitioning to a higher Power State. This allows the Audio Function to completely remove
power from the Power Domain. Note that Power Domain Entities themselves shall never be part of a Power
Domain and their AudioControls shall always be operational.

Note: The Audio Function shall ignore any traffic on its USB streaming Endpoints when their associated
Terminals are in Power States PS2 or below.

An Audio Function is not allowed to change the Power State of any of its Power Domains autonomously. A change
in Power State shall always be initiated through an explicit Command from the Host. Power States shall be retained
across Function or Device suspend. All transitions from any Power State PSx to any other Power State PSy are
allowed.

Note: Power States operate independently from the Function or Device suspend state. For example,
bringing all Power Domains in an Audio Function into Power State PS4 does not automatically put the
USB Device into USB suspend.

An Audio Function shall always honor a Command to change the Power State of a Power Domain.

Actual power consumption levels for each Power State are not advertised. The only requirement is that a higher
numbered Power State consume no more power than all lower numbered Power States, potentially at the expense
of higher exit times. Entry time is defined as the approximate time it takes to get from the fully operational Power
State PS0 to a lower Power State. Exit time is defined as the approximate time it takes to get from a lower Power
State back to the fully operational Power State PS0. The Power Domain Entity Descriptor indicates the entry and
exit times for Power States PS1 to PS4. Intermediate entry and exit times (from PSn to PSm where both n and m
are non-zero) are not listed but shall never be larger than the corresponding PS0 to PSn times (for entry times) or
PSn to PS0 times (for exit times).

The Audio Function is best placed to manage the details of its resources and their power consumption under
various conditions. These details are therefore not exposed to the Host. Rather, the Host manages the Power
States in each of the Power Domains to indicate to the Audio Function which parts of the Audio Function it intends
to use.

Release 4.0 April 2023 66

Universal Serial Bus Device Class Definition for Audio Devices

The Power State Control shall be used by Host software to Command a Power State change or return the current
Power State.

3.14.6.1 INTERDEPENDENCY AMONG POWER STATES, BYPASS FUNCTIONALITY AND

OUTPUT CLUSTERS
The Cluster present on the Output Pin of an Entity depends, among other things, on the Power State of the Entity,
the state of the Bypass Control, and the state of the Cluster Control.

The following table summarizes the required behavior.

Table 3-1: Output Cluster Configuration and Cluster Content Behavior

Bypass Control Not Present or Disabled Enabled

Cluster Control n>0 0 Don’t Care
PS0 Cluster Configuration Cluster Configuration
Power State Ctrl

set to n. inherited from Input Pin 1. Cluster Configuration and

PS1
Cluster Content Cluster Content generated Cluster Content inherited
PS2 generated internally internally from Input Pin 1
PS3
Inactive Inactive
PS4 Inactive

3.14.7 ADDITIONAL POWER CONSIDERATIONS AND REQUIREMENTS

It is strongly recommended that an Audio 4.0 compliant High Speed Device support LPM/L1. When supported, the
implementation shall follow the requirements defined in the USB 2.0 LPM/L1 ECN and LPM Errata. LPM/L1 is not
applicable to Full Speed or SuperSpeed Devices. LPM/L1 enables both the Device and Host to save USB subsystem
power by entering “sleep” during L1 power state. Sleep power is typically similar to Suspend power.

3.14.7.1 USB AUDIO 4.0 LPM/L1 IMPLEMENTATION EXAMPLE

Isochronous transfers for legacy USB audio (UAC1 or UAC2) typically occur every 1 ms for Full Speed USB. The bus
is busy during a large part of the service interval. Experience has shown that legacy USB audio cannot replace the
3.5 mm phone jack in mobile phones and other portable, battery powered products, as power consumption is too
high during audio playback use case. An alternative implementation is High Speed USB and UAC2, where audio is
transferred in short bursts. There are some power savings due to the bus being idle during most of the service
interval. However, the bus cannot enter Suspend L2 for further power savings since L2 entry/exit latency is ~10 ms
or more.

The USB 2.0 Link Power Management (LPM) specification defines a new L1 power state (LPM/L1) for High Speed
USB, with directed L1 entry and short L1 exit latency. The combination of High Speed USB bursting and the L1
power state offers significant power savings opportunities for USB Audio 4.0 devices and hosts. The figure below
shows typical events for an LPM/L1 capable USB 4.0 Audio device.

Release 4.0 April 2023 67

Universal Serial Bus Device Class Definition for Audio Devices

Figure 3-26: Typical behavior of Audio Device that implements LPM/L1

Note: There is repeated LPM/L1 exit (Resume), data burst, and LPM/L1 entry sequences. Actual timing is
Host and implementation dependent. Typical Host implementations enable L1 power state ~50 % of
the time with a 1 ms Service Interval and 85-90 % with a 4 ms Service Interval. Note that SOF(s) after
the data burst are optional and Host implementation dependent.

3.14.7.2 LPM/L1 BACKGROUND INFO

From a USB point of view, LPM/L1 is directed Suspend, derived from SuperSpeed U1/U2 power states. LPM/L1 fast
entry uses an extended LPM Token, while LPM/L1 fast exit uses short duration Host Initiated Resume Signaling.
Active data transfer times for typical audio streams are typically less than 20 µs per each ms of Service Interval.
Even considering LPM/L1 entry and exit overhead, significant power savings are possible with 1 ms Service Interval.
Audio 4.0 Devices may offer 2 ms or 4 ms Service Intervals for even more power savings. This is especially useful
when Audio 4.0 Devices are used with battery powered Hosts supporting audio-only playback use cases.

3.14.7.3 DEVICE LPM/L1 IMPLEMENTATION REQUIREMENTS

An LPM/L1 capable Audio 4.0 Device shall present support for LPM/L1 in the extended USB BOS descriptor. The
Device shall present a descriptor corresponding to 75 µs or 100 µs Host Initiated Resume Delay (HIRD). An Audio
4.0 Device shall accept the LPM/L1 entry Commands (i.e., return an ACK), even if the Device’s isochronous RX
Endpoint buffer has not yet been emptied or the isochronous TX Endpoint already contains samples for the next
audio burst. During the L1 power state, the Audio 4.0 Device USB subsystem shall enter the L1 low-power mode
aka ‘sleep’ with lower power consumption than during ‘idle’ i.e., no active data transfers.

3.14.7.4 HOST LPM/L1 BEHAVIOR NOTES

A USB Host that supports LPM/L1 is required to follow the requirements defined in the USB 2.0 LPM/L1 ECN and
LPM Errata. The Host may decide to not use LPM/L1 even if the Audio 4.0 Device supports LPM/L1. When LPM/L1
is enabled, LPM/L1 capable Hosts may, and typically will (but are not required to) Command LPM/L1 entry after
servicing the Audio 4.0 Device Endpoint(s) in each Service Interval. When in L1, the Host is responsible for, and can
be trusted to resume the Device in time to service the next Service Interval.

3.14.7.5 USB AND AUDIO SYNCHRONIZATION

For isochronous Endpoints, the Host will send one or more SOF tokens before resuming transactions to the Device
Endpoint(s). The Host may (but is not required to) send one or more SOF tokens before Commanding LPM/L1 entry

Release 4.0 April 2023 68

Universal Serial Bus Device Class Definition for Audio Devices

after servicing the Endpoint(s). As indicated in Figure 3-26, a highly optimized host can send an LPM/L1 Token
immediately after the Data burst.

An LPM/L1 capable Audio 4.0 Device can use sequential SOF tokens to synchronize its internal clock with the USB
clock. If the Host does not send enough consecutive SOF tokens, and/or the Device needs additional SOF tokens in
order to resynchronize its internal clock with the USB clock, the Device may NAK the Command to put the link into
an L1 state after the Host has serviced the endpoint(s) in the current Service Interval; this ensures that the link
stays in L0 until the next Service Interval. The Device shall NAK the LPM/L1 token no more than once every 64 ms
to ensure that LPM/L1 usage does save power.

3.14.8 GROUP MODEL

Group support is optional for an Audio Function. This specification defines three group types as indicated in
Section 3.10, “Groups.”

3.14.8.1 CHANNELGROUP
ChannelGroups are an integral part of a Cluster and groups together closely related audio channels in the Cluster.
The ChannelGroup is a lightweight construct, consisting of a list of Channel IDs of the related channels.

3.14.8.2 ENTITYGROUP
EntityGroups group together Entities that have a relationship to or are associated with one another in some form.
For example, some Input Terminals that represent the various microphones of a sophisticated gaming headset,
and an Output Terminal that represents the earpieces (headphone part) of that headset can be grouped together
to indicate that they are part of the same device. The EntityGroup is a lightweight construct, consisting of an
EntityGroup Descriptor that simply enumerates the Members of the EntityGroup by listing their respective Entity
IDs. An EntityGroup does not have any associated AudioControls. Connector Entities are typically not included as
Members of an EntityGroup. The Connectors associated with a Terminal are referenced directly from within the
Terminal Descriptor.

3.14.8.3 COMMITGROUPS
CommitGroups allow the Audio Function to indicate to Host software which AudioControls are best updated using
the Commit Capability to avoid unwanted artifacts or provide the best user experience. As a generic example,
multiple AudioControls that impact a single feature, such as those exposed by Effect Units or Processing Units, are
best set to their desired values at the same point in time. In most cases, it would be even desirable to
synchronously apply the desired changes across multiple channels as well.

The CommitGroup is a lightweight construct, exposed within the Audio Function through its CommitGroup
Descriptor. A CommitGroup Descriptor lists all AudioControls that are best manipulated in a synchronous fashion
to achieve smooth and artifact-free operation of a certain feature, effect or process. If supported, the Descriptor ID
may be used in the Commit Command to only affect the AudioControls in that CommitGroup.

All AudioControls that are Members of a CommitGroup shall support the NEXT Attribute. Multiple CommitGroups
may be exposed within the Audio Function. AudioControls may be members of multiple CommitGroups.

3.14.9 BINDING BETWEEN BUTTONS AND AUDIOCONTROLS

A Device that contains an Audio Function may also have one or more buttons/sliders/knobs that are intended to
control certain aspects of the Audio Function inside the Device. An example is a Gain Control on a multimedia

Release 4.0 April 2023 69

Universal Serial Bus Device Class Definition for Audio Devices

speaker. Since an Audio Function may potentially contain many AudioControls of the same type, there is a need to
bind a physical control (button, knob, slider, jog, …) to a particular AudioControl inside the Audio Function.

This specification provides two mutually exclusive methods to create this binding:

• The button is implemented using the Human Interface Device Class (Universal Serial Bus Device Class
Definition for Human Interface Devices (HID)). In this case, it is the responsibility of the Host to establish the
binding between the HID event and the action issued to the Audio Function(s)
• The button is an integral part of the AudioControl. In this case, the AudioControl shall notify the Host of any
change through the interrupt mechanism.

It is prohibited to implement both methods for the same physical button. However, it is allowed to use the first
method for some of the buttons and the second method for the remaining buttons. It is strongly discouraged to
implement buttons that use neither of the above-mentioned methods, i.e., buttons that are invisible to Host
software and have a local effect only.

3.14.9.1 BUTTON IS A HID CONTROL

In this case, the button is separate from the Audio Function and is implemented as part of a HID interface within
the Device. Any change of state for the button is communicated to Host software via HID reports. It is then up to
Host software to interpret the button state change and determine the appropriate action to be taken toward the
Audio Function. Therefore, the binding responsibility resides entirely within the application or Operating System
software. Although this method provides extensive flexibility, it also puts the burden of providing the correct
binding on the software, making it sometimes hard to create generic application or OS software that generates the
proper (manufacturer intended) binding.

3.14.9.2 BUTTON IS AN INTEGRAL PART OF THE AUDIOCONTROL

In this case, the button directly interacts with the actual AudioControl. The change of state of the AudioControl
resulting from the button manipulation is reported to Host software through the AudioControl interrupt
mechanism. Consequently, the binding between the button and the AudioControl is very direct and entirely
dictated by the design of the Device. This method provides a very clear and straightforward binding between the
button and the behavior of the Audio Function but prevents the Host from allocating the button for other
purposes.

Release 4.0 April 2023 70

Universal Serial Bus Device Class Definition for Audio Devices

4 DESCRIPTORS

The following sections describe the standard and class-specific USB Descriptors for the Audio Interface Class.

4.1 STANDARD DESCRIPTORS

Because Audio functionality is always residing at the interface level, all relevant fields in the standard Descriptors
shall indicate that class information is to be found at the interface level so that enumeration software looks down
at the interface level to determine the Interface Class and to also ensure that IAD-aware enumeration software
gets loaded.

4.2 CLASS-SPECIFIC DESCRIPTORS

Note: In this 4.0 version of the Specification, High Capability Descriptors as defined in version 3.0 of the
Audio Device Class-specification are renamed to Extended Descriptors. Likewise, the Get High
Capability Descriptor Command is renamed to Get Extended Descriptor Command.

This specification uses Extended Descriptors to express most of its class-specific Descriptors. Extended Descriptors
are never part of the configuration Descriptor hierarchy, returned by the Get Configuration Command. Only
traditional layout Descriptors shall be included in this hierarchy. An Extended Descriptor shall therefore always be
referenced by a traditional class-specific Descriptor that includes the Extended Descriptor’s unique ID as one of its
fields.

In the remainder of this specification, the “Extended” designation may be omitted when referring to a class-
specific Extended Descriptor. Furthermore, the designation “class-specific” may be omitted when the context
makes it obvious that a class-specific Descriptor is referenced.

4.2.1 TRADITIONAL CLASS-SPECIFIC DESCRIPTOR CHARACTERISTICS

Traditional class-specific Descriptors as defined in this specification all follow a common layout. The first three
fields of any traditional class-specific Descriptor are common to all traditional Descriptors and are followed by a
layout that is specific to the type and subtype of the descriptor.

The bLength field contains the total length of the Descriptor, in bytes.

The bDescriptorType field in part follows the bit allocation scheme of the bmRequestType field and identifies the
Descriptor as being a class-specific Descriptor. Bit D7 of this field is reserved and shall be set to zero. Bits D6..5 are
used to indicate that this is a class-specific Descriptor (D6..5 = 0b01). Bits D4..0 are used to encode the Descriptor
type.

The bDescriptorSubtype field further qualifies the exact nature of the Descriptor.
Table 4-1: Traditional Class-specific Descriptor Layout

Offset Field Size Value Description

0 bLength 1 Number Size of this Descriptor, in bytes: 3+n.
1 bDescriptorType 1 0b001xxxxx Descriptor type.
2 bDescriptorSubtype 1 Constant Descriptor subtype.
3 --- n --- Descriptor type- and subtype-specific
Descriptor layout.

Release 4.0 April 2023 71

Universal Serial Bus Device Class Definition for Audio Devices

4.2.2 EXTENDED CLASS-SPECIFIC DESCRIPTOR CHARACTERISTICS

Most class-specific Descriptors of an Audio Function are not retrieved during enumeration time. Rather, they can
only be retrieved using the Get Extended Descriptor Command as defined in Section 5.3.5.2, “Extended Descriptor
”.

Note: The only class-specific Descriptors that are retrieved during enumeration are the traditional class-
specific Descriptors of Type CS_INTERFACE and Subtype AC_GENERIC or AS_GENERIC.

An Extended Descriptor is always referenced by another Descriptor via its Audio Function-wide unique ID.

An Extended Descriptor may be up to 64 KiB in length. Also, it shall generate an interrupt of source type
EXTENDED_DESCRIPTOR whenever the Audio Function changes the Descriptor during normal operation (dynamic
Descriptor).

Extended Descriptors all follow a common layout. The first five fields of any Extended Descriptor are common to all
Extended Descriptors and are followed by a layout that is specific to the type and subtype of the Descriptor. The
common fields are:

• The wLength field contains the total length of the Descriptor, in bytes. Descriptor lengths up to a maximum of
65,535 bytes are supported.
• The wDescriptorType field in part follows the bit allocation scheme of the bmRequestType field as defined by
the standard USB Request and identifies the Descriptor as being a class-specific Descriptor. Bits D15..7 of this
field are reserved. Bits D6..5 are used to indicate that this is a class-specific Descriptor (D6..5 = 0b01). Bits
D4..0 are used to encode the Descriptor type.
• The wDescriptorSubtype field further qualifies the exact nature of the Descriptor.
• The wDescriptorID field contains a value that uniquely identifies the Extended Descriptor within the Audio
Function. The value zero is reserved and shall not be used as a valid Descriptor ID.
• The wStrDescriptorID field contains the ID of a class-specific String Descriptor that provides additional
descriptive information about the subject of this Descriptor. For example, a class-specific Clock Source
Descriptor would use this field to provide a human-readable name for the Clock Source. This field shall be set
to zero if there is no String Descriptor associated with the class-specific Extended Descriptor.
Table 4-2: Class-specific Descriptor Layout

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes: 10+n.
2 wDescriptorType 2 Constant Descriptor type.
4 wDescriptorSubtype 2 Constant Descriptor subtype.
6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number ID of a class-specific String Descriptor that
provides additional information about the
subject of this Descriptor.
10 --- n --- Descriptor type- and subtype-specific
Descriptor layout.

Release 4.0 April 2023 72

Universal Serial Bus Device Class Definition for Audio Devices

4.2.3 COMMON FIELDS IN SOME CLASS-SPECIFIC DESCRIPTORS

Some fields appear in more than one class-specific Descriptor. Instead of repeating the description of that field for
all Descriptor instances, the description is indicated once in this Section and applies to all Descriptor instances in
which the field appears.

4.2.3.1 wxxxID FIELD

Each Entity (Clock, Unit, Terminal, and Power Domain) within the Audio Function is assigned a unique identification
number: The Clock Entity ID, Unit ID, Terminal ID, or Power Domain ID.

All Entity Descriptors have a common field at offset ten that contains the unique ID for the Entity. The value zero
(0x0000) is used to identify the Interface itself.

Besides uniquely identifying all addressable Entities in an Audio Function, the IDs (except for the Power Domain ID)
also serve to describe the topology of the Audio Function, i.e., the wSourceID field of a Unit or Terminal Descriptor
indicates to which other Unit or Terminal this Unit or Terminal is connected. Likewise, the wCSourceID field in a
Terminal or SRC Descriptor indicates to which Clock Entity this Terminal or SRC is connected. Furthermore, the
Entity IDs are also used to indicate to which Power Domain each Entity belongs.

4.2.3.2 wxxxDescrID FIELD

Most class-specific Descriptors have one or more 2-byte fields with a name of the form “wxxxDescrID” that
contains the ID of another class-specific Descriptor that either serves as an extension to the first Descriptor or
refers to a String Descriptor that provides additional human-readable information. A value of zero in this field
indicates that there is no extension Descriptor or String Descriptor associated with that first Descriptor.

4.2.3.3 dOptControls FIELD

The dOptControls field (or daOptControls() array elements) contains a bitmap, indicating which of the predefined
optional AudioControls for that Entity are present. If a certain AudioControl is not present, then the bit shall be set
to 0b0. If an AudioControl is present, then the bit shall be set to 0b1. Each implemented AudioControl provides all
necessary information for the Host to determine how to interact with it via the mandatory Capability (CAP)
Attribute. AudioControls that are mandated by this specification are not included in the dOptControls field.

4.3 AUDIO FUNCTION DESCRIPTOR SET

The standard interface association mechanism is used to describe an Audio Interface Association (AIA). All
Interfaces belonging to the same AIA shall be identified by means of the standard Interface Association Descriptor
(IAD).

Each AIA shall consist of the mandatory AudioControl Interface that shall be the first in the AIA (having the lowest
interface number). All AudioStreaming Interfaces shall be contiguously numbered and immediately follow the
AudioControl Interface in the AIA.

Note: For more information on Interface Association, refer to USB Interface Association Descriptor Device
Class Code and Use Model White Paper, available on the USB web site.

The collection of the Interface Association Descriptor (IAD) and the full set of Standard Descriptors and class-
specific Descriptors that together comprise the entire Audio Function is called the Audio Function Descriptor Set
(AFDS).

Release 4.0 April 2023 73

Universal Serial Bus Device Class Definition for Audio Devices

4.3.1 AUDIO FUNCTION DESCRIPTOR SET LAYOUTS

In the Audio Device Class-specification Revision 2.0 (and lower), all Descriptors are of the traditional type as
defined in the USB Core Specifications. All class-specific Descriptors are interspersed with the standard Descriptors
to provide a full view of the Audio Function. The AFDS 2.0 is always retrieved from the Device at enumeration time
as part of a Configuration Descriptor bundle. Note that in the Configuration Descriptor bundle, the IAD does not
need to immediately precede the AudioControl Interface Descriptor to which the IAD refers. It is sufficient that the
IAD is positioned earlier in the bundle than the AudioControl Interface Descriptor to which the IAD refers.
Figure 4-1: Audio 2.0 Function Descriptor Set layout

Interface Association

More AudioStreaming Interfaces

Interface #n Interface #(n+1)

AUDIO_CONTROL AUDIO_STREAMING
Alternate Setting #0 Alternate Setting #0

CS_Interface Interface #(n+1)

AUDIO_STREAMING
Alternate Setting #1

Interrupt Endpoint #n * CS_Interface

Standard
Data Endpoint #p *

Traditional
Class-Specific CS_Endpoint

Feedback *
Endpoint #q

Optional More Alternate Settings

* In the case of SuperSpeed and SuperSpeed+, this Descriptor may actually consist of
multiple standard Descriptors

Starting with Audio Device Class-specification Revision 3.0, the concept of the Extended Descriptor is introduced
and some of the class-specific Descriptors are defined as Extended Descriptors.
Most of the class-specific Descriptors are still of the traditional type and interspersed with the standard
Descriptors. The AFDS 3.0 is always retrieved from the Device at enumeration time as part of a Configuration
Descriptor bundle. The Host then uses the Get Extended Descriptor Command to retrieve the Extended Descriptors
whenever it encounters an Extended Descriptor ID in one of the traditional class-specific Descriptors. As such, the
set of Extended Descriptors is not considered to be part of the AFDS.

Release 4.0 April 2023 74

Universal Serial Bus Device Class Definition for Audio Devices

Figure 4-2: Audio 3.0 Function Descriptor Set layout

Interface Association

More AudioStreaming Interfaces

Interface #n Interface #(n+1)

AUDIO_CONTROL AUDIO_STREAMING
Alternate Setting #0 Alternate Setting #0 CS_Interface
CLUSTER #x

CS_Interface Interface #(n+1)

CS_Interface
AUDIO_STREAMING
Alternate Setting #1 EX_TERMINAL

CS_Interface
Interrupt Endpoint #n * CS_Interface CONNECTORS

Standard
Data Endpoint #p *

Traditional CS_Interface
Class-Specific CS_Endpoint CLUSTER #y

Extended Feedback *
Class-Specific Endpoint #q

Optional More Alternate Settings

* In the case of SuperSpeed and SuperSpeed+, this Descriptor may actually consist of
multiple standard Descriptors

With this version 4.0 of the Specification, the use of Extended Descriptors has been further generalized. All class-
specific information is now made available exclusively through Extended Descriptors as illustrated in Figure 4-3.
The only traditional class-specific Descriptors that are interspersed with the standard Descriptors are Descriptors
of type xx_GENERIC. The AFDS 4.0 is retrieved from the Device at enumeration time as part of a Configuration
Descriptor bundle. Alternatively, an AFDS can be retrieved from the Device as a Higher Revision Level AFDS
through other means as described in Section 8, “Backwards Compatibility Considerations”.

The class-specific Descriptors of type xx_GENERIC simply contain a list of Descriptor IDs. Each ID references an
Extended Descriptor in the Extended Descriptor Store. The Host should consult the Descriptor Store to retrieve all
class-specific information regarding the Audio Function. It uses the Get Extended Descriptor Command to retrieve
each individual Descriptor in the Store. The Extended Descriptor Store is not considered to be part of the AFDS.

Release 4.0 April 2023 75

Universal Serial Bus Device Class Definition for Audio Devices

Figure 4-3: Audio 4.0 Function Descriptor Set layout

Interface Association

More AudioStreaming Interfaces

Interface #n Interface #(n+1) CS Extended Descriptor Store

AUDIO_CONTROL AUDIO_STREAMING
Alternate Setting #0 Alternate Setting #0 AC Interface

EXT_INTERFACE
CS_INTERFACE Interface #(n+1) AC_SELF
AC_GENERIC AUDIO_STREAMING
Alternate Setting #1
EXT_INTERFACE EXT_INTERFACE
Interrupt Endpoint #n * IT #x IT #y
CS_INTERFACE
AS_GENERIC
EXT_INTERFACE EXT_INTERFACE
Standard Data Endpoint #p * OT #x OT #y

EXT_INTERFACE EXT_INTERFACE
Traditional
Feedback Endpoint * FU #x FU #y
#q
Class-Specific

Extended
Class-Specific
AS Interface
Optional More Alternate Settings EXT_INTERFACE EXT_INTERFACE
AS_SELF VALID_FREQ
* In the case of SuperSpeed and SuperSpeed+, this Descriptor may actually consist of
multiple standard Descriptors

Cluster

EXT_INTERFACE EXT_INTERFACE
CLUSTER #x CLUSTER #y

String

EXT_STRING EXT_STRING
STRING #x STRING #y

4.4 CLUSTER DESCRIPTOR

A Cluster is a grouping of audio channels that share the same characteristics such as sampling frequency, bit
resolution, etc. The Cluster Descriptor is used to characterize the Cluster. At the highest level, it is structured as
follows:

Release 4.0 April 2023 76

Universal Serial Bus Device Class Definition for Audio Devices

Figure 4-4: Cluster Descriptor

Header

Common Block

Channel 1 Block

Channel n Block

The Cluster Descriptor consists of a fixed Header, followed by an optional Common Block, followed by as many
Channel Blocks as there are channels in the Cluster.

4.4.1 CLUSTER DESCRIPTOR HEADER

The Header starts with the wLength field that contains the total number of bytes in the Cluster Descriptor,
including the Cluster Descriptor Header and all included Cluster Descriptor Blocks.

The wNrChannels field indicates the number of audio channels present in the Cluster.

If for any reason, the Cluster is currently Inactive (not carrying any audio data), then its Cluster Descriptor shall not
be used for any purpose.
Note: The Cluster Descriptor shall not generate an interrupt when the Cluster changes between Active and
Inactive since the Cluster Descriptor itself does not change.
Table 4-3: Cluster Descriptor Header

Offset Field Size Value Description

0 wLength 2 Number Size of the entire Cluster Descriptor,
including the Header and all of its building
blocks, in bytes.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.
4 wDescriptorSubtype 2 Constant CLUSTER Descriptor subtype.
6 wDescriptorID 2 Number Unique ID of this Cluster Descriptor.
8 wStrDescriptorID 2 Number Unique ID of a String Descriptor.

Release 4.0 April 2023 77

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

10 wNrChannels 2 Number Number of channels present in the Cluster:
n.

4.4.2 CLUSTER DESCRIPTOR BLOCK

The Cluster Descriptor Header is followed by one or more Cluster Descriptor Blocks. There is an optional Common
Block, followed by as many mandatory Channel Blocks as there are channels in the Cluster. Each Block consists of
one or more Segments.

The Common Block consists of Segments that contain relevant information about characteristics of the Cluster as a
whole. All Common Block Segments are optional.

A Channel Block consists of Segments that contain relevant information about that channel’s characteristics. At a
minimum, each Channel Block shall contain either an Information Segment or an Ambisonic Segment and a single
End Block Segment. All other Segments are optional.

Figure 4-5 further illustrates the above concepts.

Figure 4-5: Cluster Descriptor Block

Block

Segment 1

Segment 2

Segment m

End Segment

4.4.2.1 SEGMENTS
There are two types of Segments. Common Block Segments contain pertinent information about the Cluster as a
whole. Channel Block Segments contain pertinent information about certain aspects of a particular channel. Both
Segment types share the same layout.

The wSegmentType field describes the Segment Type (Common Block or Channel Block) and the type of content
contained in the Segment.

The layout for a Segment is always as follows:

Table 4-4: Cluster Descriptor Segment

Offset Field Size Value Description

0 wLength 2 Number Length of the Segment, in bytes: 4+n.

Release 4.0 April 2023 78

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

2 wSegmentType 2 Constant Describes the Segment Type and the type of
content in the Segment.
4 Segment-specific n Segment-specific content.

4.4.2.1.1 END BLOCK SEGMENT

Each Block is terminated by an End Block Segment. The End Block Segment marks the end of the variable length
Block. The End Block Segment does not have a Segment-specific section and is structured as follows:
Table 4-5: End Block Segment

Offset Field Size Value Description

0 wLength 2 Number Length of the Segment, in bytes: 4.
2 wSegmentType 2 Constant CLUSTER_END_BLOCK.

4.4.2.1.2 COMMON BLOCK SEGMENTS

This specification currently does not define any Common Block Segments.

4.4.2.1.3 CHANNEL BLOCK SEGMENTS

The following Channel Block Segment types are defined:

• Information
• Ambisonic
• Channel Description

Values for the Channel Block Segment types can be found in Appendix A.13, “Cluster Descriptor Segment Types.”

4.4.2.1.3.1 INFORMATION SEGMENT

The Information Segment contains relevant information for a particular channel in the Cluster. It is mutually
exclusive with the Ambisonic Segment for that same channel.
The wChPurpose field indicates the primary Purpose of the channel. Currently defined Purposes for a channel are:

• Generic Audio: contains audio primarily used for direct capture or reproduction.
• Voice: intended to be interpreted by humans.
• Speech: intended to be interpreted or generated by machine.
• Ambient: contains audio other than the primary channels.
• Reference: contains final processed audio. For example, a reference for AEC processing.
• Ultrasonic: contains signals with spectral content above audible limits (typically >20 kHz).
• Vibrokinetic: contains very low frequency information, typically used to actuate vibrators or motion
actuators.
• Sense: contains current/voltage/temperature/etc. real-time data.
• Non-Audio: indicates that the channel carries non-audio information. Examples of non-audio information
are real-time pressure sensing data or amplifier gain feedback data, etc.
• Silence: indicates that the channel contains inaudible content, i.e., silence.

Release 4.0 April 2023 79

Universal Serial Bus Device Class Definition for Audio Devices

The values for the Purposes listed above can be found in Appendix A.14, “Channel Purpose Definitions.”

Note that great care should be taken when indicating the primary Purpose for an audio channel. For example, if
unprocessed audio is subsequently processed specifically to be used by applications that use human voice as input,
such a channel may be marked Voice. Also, a channel that is marked for a particular primary Purpose may be used
for other purposes, compatible with the primary Purpose, as well.

The wChRelationship field describes the relationship of this channel with respect to the other channels in the
Cluster. Currently defined relationships for a channel are described in the table below.

Array
Pattern_X
Pattern _Y
Pattern _A
Pattern _B
Pattern _M
Pattern _S

Table 4-6: Channel Relationships

Channel Relationship Acronym Description

RELATIONSHIP_UNDEFINED UND Relationship undefined or unknown

Mono M A mono channel

Generic

Left L A generic Left channel. Not intended for headphones

Right R A generic Right channel. Not intended for headphones
Array AR A channel that is part of an array configuration

Release 4.0 April 2023 80

Universal Serial Bus Device Class Definition for Audio Devices

Channel Relationship Acronym Description

Headset_Mic HM A Headset microphone
Headset_Mic Left HML A Headset microphone to the left of the mouth
Headset_Mic Right HMR A Headset microphone to the right of the mouth
Headset_Mic Center HMC A Headset microphone in a central position w.r.t. the mouth
Body_Mic BM A microphone worn on the body
Body_Mic Left BML A microphone worn on the left side of the body
Body_Mic Right BMR A microphone worn on the right side of the body
Body_Mic Center BMC A microphone worn centrally on the body
Input-Related

Lapel_Mic LM A microphone worn on the lapel

Lapel_Mic Left LML A microphone worn on the left lapel
Lapel_Mic Right LMR A microphone worn on the right lapel
Lapel_Mic Center LMC A microphone worn centrally at lapel height
Pattern_X PX The X channel of an X/Y microphone
Pattern _Y PY The Y channel of an X/Y microphone
Pattern _A PA The A channel of an A/B microphone
Pattern _B PB The B channel of an A/B microphone
Pattern _M PM The M channel of an M/S microphone
Pattern _S PS The S channel of an M/S microphone
Front Left FL Refer to Figure 4-6. Not intended for headphones
Front Right FR Refer to Figure 4-6. Not intended for headphones
Front Center FC Refer to Figure 4-6
Front Left of Center FLC Refer to Figure 4-6
Front Right of Center FRC Refer to Figure 4-6
Front Wide Left FWL Refer to Figure 4-6
Input/Output-Related

Front Wide Right FWR Refer to Figure 4-6

Side Left SL Refer to Figure 4-6
Side Right SR Refer to Figure 4-6
Surround Array Left SAL Refer to Figure 4-6
Surround Array Right SAR Refer to Figure 4-6
Back Left BL Refer to Figure 4-6
Back Right BR Refer to Figure 4-6
Back Center BC Refer to Figure 4-6
Back Left of Center BLC Refer to Figure 4-6
Back Right of Center BRC Refer to Figure 4-6
Back Wide Left BWL Refer to Figure 4-6

Release 4.0 April 2023 81

Universal Serial Bus Device Class Definition for Audio Devices

Channel Relationship Acronym Description

Back Wide Right BWR Refer to Figure 4-6
Top Center TC Refer to Figure 4-6
Top Front Left TFL Refer to Figure 4-6
Top Front Right TFR Refer to Figure 4-6
Top Front Center TFC Refer to Figure 4-6
Top Front Left of Center TFLC Refer to Figure 4-6
Top Front Right of Center TFRC Refer to Figure 4-6
Top Front Wide Left TFWL Refer to Figure 4-6
Top Front Wide Right TFWR Refer to Figure 4-6
Top Side Left TSL Refer to Figure 4-6
Top Side Right TSR Refer to Figure 4-6
Top Surround Array Left TSAL Refer to Figure 4-6
Top Surround Array Right TSAR Refer to Figure 4-6
Top Back Left TBL Refer to Figure 4-6
Top Back Right TBR Refer to Figure 4-6
Top Back Center TBC Refer to Figure 4-6
Top Back Left of Center TBLC Refer to Figure 4-6
Top Back Right of Center TBRC Refer to Figure 4-6
Top Back Wide Left TBWL Refer to Figure 4-6
Top Back Wide Right TBWR Refer to Figure 4-6
Bottom Center BOC Refer to Figure 4-6
Bottom Front Left BFL Refer to Figure 4-6
Bottom Front Right BFR Refer to Figure 4-6
Bottom Front Center BFC Refer to Figure 4-6
Bottom Front Left of Center BFLC Refer to Figure 4-6
Bottom Front Right of Center BFRC Refer to Figure 4-6
Bottom Front Wide Left BFWL Refer to Figure 4-6
Bottom Front Wide Right BFWR Refer to Figure 4-6
Bottom Side Left BSL Refer to Figure 4-6
Bottom Side Right BSR Refer to Figure 4-6
Bottom Surround Array Left BSAL Refer to Figure 4-6
Bottom Surround Array Right BSAR Refer to Figure 4-6
Bottom Back Left BBL Refer to Figure 4-6
Bottom Back Right BBR Refer to Figure 4-6
Bottom Back Center BBC Refer to Figure 4-6

Release 4.0 April 2023 82

Universal Serial Bus Device Class Definition for Audio Devices

Channel Relationship Acronym Description

Bottom Back Left of Center BBLC Refer to Figure 4-6
Bottom Back Right of Center BBRC Refer to Figure 4-6
Bottom Back Wide Left BBWL Refer to Figure 4-6
Bottom Back Wide Right BBWR Refer to Figure 4-6
Low Frequency Effects LFE Refer to Figure 4-6
Low Frequency Effects Left LFEL Refer to Figure 4-6
Low Frequency Effects Right LFER Refer to Figure 4-6
Headphone Left HPL Refer to Figure 4-6
Headphone Right HPR Refer to Figure 4-6

The following figure presents a spatial “view from above” of the different Channel Relationships.

Release 4.0 April 2023 83

Universal Serial Bus Device Class Definition for Audio Devices

Figure 4-6: 3D Representation of the Channel Relationships

Front
Top
TFWL TFL TFLC TFC TFRC TFR TFWR

FWL FL FLC FC FRC FR FWR

BFWL BFL BFLC BFC BFRC BFR BFWR

Bottom

TSAR
TSAL

SAR
SAL

BSAR
BSAL

LFEL LFE LFER

TSAR
TSAL

SAR
SAL

BSAR
BSAL

TC

TSR
Right
TSL

HPR

BSR
HPL
BSL
Left

SR
SL

BOC

BSAR
BSAL

SAR

TSAR
SAL
TSAL

BSAR
BSAL

SAR
SAL

TSAR
TSAL

BBWL BBL BBLC BBC BBRC BBR BBWR

BWL BL BLC BC BRC BR BWR

TBWL TBL TBLC TBC TBRC TBR TBWR

Back

Constant names, acronyms, and their associated values are listed in Appendix A.15, “Channel Relationship
Definitions.”

The wChannelID field is used to uniquely identify the channel within the scope of the entire Audio Function. No
two channels in the entire Audio Function shall use the same value for this field. Setting the wChannelID field to
zero is prohibited. Implementations shall populate this field so that a Host can retrieve meaningful signal routing
information from within the Audio Function. Although it is in some cases difficult to establish definite rules on how
to propagate Channel IDs when Clusters pass through certain types of Units, it is left to the implementation to
decide how to manage Channel IDs in those cases.

The wChGroupID field is used to create a link among a subset of channels in the Cluster by specifying the same
value in the wChGroupID field for those channels. This field is useful when multiple sets of related channels are

Release 4.0 April 2023 84

Universal Serial Bus Device Class Definition for Audio Devices

present in the same Cluster. For example, a single Cluster may carry one ChannelGroup of generic Left and Right
stereo channels and another ChannelGroup of generic Left and Right stereo channels. The first generic Left and
Right channels would share the same ChannelGroup ID value, indicating that they belong to a ChannelGroup, and
the second generic Left and Right stereo channels would share another ChannelGroup ID value, indicating they
belong to a different ChannelGroup. The scope of the wChGroupID field is limited to the Cluster to which the
channels belong. A value of zero in this field indicates that the channel is not part of a ChannelGroup. Note also
that this specification does not prohibit ChannelGroups that contain only a single channel. However, the
ChannelGroup ID should not be confused with the concept of the Channel ID.

Each channel in a ChannelGroup shall have a

The wConID field identifies the Connector (through its Connector Entity) to which the channel is associated. If the
channel is not associated with a Connector, then the wConID field shall be set to zero. See Section 4.5.3.14,
“Connector Entity Descriptor” for more details.
Table 4-7: Information Segment

Offset Field Size Value Description

0 wLength 2 Number Length of the Channel Information
Segment, in bytes: 14.
2 wSegmentType 2 Constant CHANNEL_INFORMATION.
4 wChPurpose 2 Constant Intended purpose of the channel.
6 wChRelationship 2 Constant Describes the relationship of this channel
with the other channels. Refer to Appendix
A.15, “Channel Relationship Definitions.”
8 wChannelID 2 Number ID used to uniquely identify a channel.
10 wChGroupID 2 Number ID used to group channels together.
12 wConID 2 Number The Connector ID of the Connector to which
the channel is associated.

4.4.2.1.3.2 AMBISONIC SEGMENT

The Ambisonic Segment contains relevant information for a particular Ambisonic channel in the Cluster. It is
mutually exclusive with the Information Segment for that same channel.
The wCompOrdering field indicates the convention used for ordering of the spherical harmonics. See Appendix
A.16, “Ambisonic Component Ordering Convention Types” for the supported component ordering conventions. All
channels in a ChannelGroup (see below) shall indicate the same component ordering convention.

The wACN field contains the Ambisonic Channel Number.

The wAmbNorm field indicates the type of normalization used for the channel. See Appendix A.17, “Ambisonic
Normalization Types” for the supported normalization types. All channels in a ChannelGroup (see below) shall
indicate the same normalization type.

The wChannelID field is used to uniquely identify the channel within the scope of the entire Audio Function. No
two channels in the entire Audio Function shall use the same non-zero value for this field. Setting the wChannelID
field to zero is prohibited. Implementations shall populate this field so that a Host can retrieve meaningful signal
routing information from within the Audio Function. Although it is in some cases difficult to establish definite rules

Release 4.0 April 2023 85

Universal Serial Bus Device Class Definition for Audio Devices

on how to propagate Channel IDs when Clusters pass through certain types of Units, it is left to the
implementation to decide how to manage Channel IDs in those cases.

The wChGroupID field is used to create a link among a subset of channels in the Cluster by specifying the same
value in the wChGroupID field for those channels. This field is useful when multiple sets of related channels are
present in the same Cluster. For example, a single Cluster may carry a ChannelGroup of 5.1 channels and another
ChannelGroup of microphone channels. The 5.1 channels would share the same ChannelGroup ID value, indicating
that they belong to a (5.1) ChannelGroup, and the microphone channels would share another ChannelGroup ID
value, indicating they belong to a different (microphone) ChannelGroup. The scope of the wChGroupID field is
limited to the Cluster to which the channels belong. A value of zero in this field indicates that the channel is not
part of a ChannelGroup. Note also that this specification does not prohibit ChannelGroup that contain only a single
channel. However, the ChannelGroup ID should not be confused with the concept of the Channel ID.

The wConID field identifies the Connector (through its Connector Entity) to which the channel is associated. If the
channel is not associated with a Connector, then the wConID field shall be set to zero. See Section 4.5.3.14,
“Connector Entity Descriptor” for more details.
Table 4-8: Ambisonic Segment

Offset Field Size Value Description

0 wLength 2 Number Length of the Segment, in bytes: 16.
2 wSegmentType 2 Constant CHANNEL_AMBISONIC.
4 wCompOrdering 2 Number The Component Ordering convention for
the Ambisonic Spherical Harmonics. Refer
to Appendix A.16, “Ambisonic Component
Ordering Convention Types.”
6 wACN 2 Number Ambisonic Channel Number.
8 wAmbNorm 2 Number Applied channel normalization.
Refer to Appendix A.17, “Ambisonic
Normalization Types.”
10 wChannelID 2 Number ID used to uniquely identify a channel.
12 wChGroupID 2 Number ID used to group channels together.
14 wConID 2 Number The Connector ID of the Connector to which
the channel is associated.

4.4.2.1.3.3 CHANNEL DESCRIPTION SEGMENT

The Channel Description Segment contains the ID of a String Descriptor in the wChStrDescrID field that provides
more information about the channel.
Table 4-9: Channel Description Segment

Offset Field Size Value Description

0 wLength 2 Number Length of the Segment, in bytes: 6.
2 wSegmentType 2 Constant CHANNEL_DESCRIPTION.
4 wChStrDescriptorID 2 Number ID of a String Descriptor that further
describes the channel.

Release 4.0 April 2023 86

Universal Serial Bus Device Class Definition for Audio Devices

4.4.3 EXAMPLE CLUSTER DESCRIPTOR

A Cluster that carries 5.1 discrete channels may have the following Cluster Descriptor:
Table 4-10: Cluster Descriptor Example

Offset Field Size Value Description

0 wLength 2 0x3E Total length of the Descriptor, in bytes: 120.
2 wDescriptorType 2 0x21 EXT_INTERFACE Descriptor type.
4 wDescriptorSubtype 2 0x11 CLUSTER Descriptor subtype.
Header

6 wDescriptorID 2 Implementation- Unique ID of this Cluster Descriptor.

dependent

8 wStrDescriptorID 2 Implementation- Unique ID of a String Descriptor.

dependent

10 wNrChannels 2 6 Number of channels present in the Cluster.

12 wLength 2 0x08 Length of the Information Segment, in bytes.
Information Segment

14 wSegmentType 2 0x20 CHANNEL_INFORMATION.

16 wChPurpose 2 0x00 Generic Audio.
18 wChRelationship 2 0x0B Front Left.
20 wChannelID 2 0x0020 Channel ID.
22 wChGroupID 2 0x0001 ID used to group channels together.
24 wConID 2 0x0000 No associated Connector Entity.
26 wLength 2 0x03 Length of the End Block Segment, in bytes.
End

28 wSegmentType 2 0xFF END_SEGMENT.

30 wLength 2 0x08 Length of the Information Segment, in bytes.
32 wSegmentType 2 0x20 CHANNEL_INFORMATION.
Information

34 wChPurpose 2 0x00 Generic Audio.

36 wChRelationship 2 0x0C Front Right.
38 wChannelID 2 0x0021 Channel ID.
40 wChGroupID 2 0x0001 ID used to group channels together.
42 wConID 2 0x0000 No associated Connector Entity.
44 wLength 2 0x03 Length of the End Block Segment, in bytes.
End

46 wSegmentType 2 0xFF END_SEGMENT.

48 wLength 2 0x08 Length of the Information Segment.
50 wSegmentType 2 0x20 CHANNEL_INFORMATION.
Information

52 wChPurpose 2 0x00 Generic Audio.

54 wChRelationship 2 0x0D Front Center.
56 wChannelID 2 0x0022 Channel ID.
58 wChGroupID 2 0x0001 ID used to group channels together.
60 wConID 2 0x0000 No associated Connector Entity.

Release 4.0 April 2023 87

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

62 wLength 2 0x03 Length of the End Block Segment, in bytes.
End

64 wSegmentType 2 0xFF END_SEGMENT.

66 wLength 2 0x08 Length of the Information Segment.
68 wSegmentType 2 0x20 CHANNEL_INFORMATION.
Information

70 wChPurpose 2 0x00 Generic Audio.

72 wChRelationship 2 0x14 Surround Array Left.
74 wChannelID 2 0x0023 Channel ID.
76 wChGroupID 2 0x0001 ID used to group channels together.
78 wConID 2 0x0000 No associated Connector Entity.
80 wLength 2 0x03 Length of the End Block Segment.
End

82 wSegmentType 2 0xFF END_SEGMENT.

84 wLength 2 0x08 Length of the Information Segment.
86 wSegmentType 2 0x20 CHANNEL_INFORMATION.
Information

88 wChPurpose 2 0x00 Generic Audio.

90 wChRelationship 2 0x15 Surround Array Right.
92 wChannelID 2 0x0024 Channel ID.
94 wChGroupID 2 0x0001 ID used to group channels together.
96 wConID 2 0x0000 No associated Connector Entity.
98 wLength 2 0x03 Length of the End Block Segment.
End

100 wSegmentType 2 0xFF END_SEGMENT.

102 wLength 2 0x08 Length of the Information Segment.
104 wSegmentType 2 0x20 CHANNEL_INFORMATION.
Information

106 wChPurpose 2 0x00 Generic Audio.

108 wChRelationship 2 0x43 Low Frequency Effects.
110 wChannelID 2 0x0025 Channel ID.
112 wChGroupID 2 0x0001 ID used to group channels together.
114 wConID 2 0x0000 No associated Connector Entity.
116 wLength 2 0x03 Length of the End Block Segment.
End

118 wSegmentType 2 0xFF END_SEGMENT.

4.4.4 CEA-861.2 CHANNEL MAPPING

The Advanced Audio Extensions specification CEA-861.2 allocates speaker spatial locations in a different order than
what is defined in this USB Audio Definition. Also, there are channel relationships defined that do not appear in the
USB Audio Definition. To provide consistency in implementations that use both the CEA-861.2 allocations and the
USB Audio channel relationships, a mapping scheme between the two is included here to which such
implementations shall adhere.

Release 4.0 April 2023 88

Universal Serial Bus Device Class Definition for Audio Devices

The mapping between the USB-defined channel relationships and the CEA speaker allocations is included in the
table in Appendix A.15, “Channel Relationship Definitions.”

4.4.5 PHYSICAL VERSUS LOGICAL CLUSTER

This specification makes a distinction between a logical and physical Cluster. Hence, there are also two types of
Cluster Descriptors defined:

• Logical Cluster Descriptor

• Physical Cluster Descriptor

The layout of these Descriptors is identical as described in Section above.

Both the logical and physical Cluster Descriptors are not independent Descriptors as such. They are always
referenced by other Descriptors. The referencing Descriptors always include a wClusterDescrID field that contains
the unique ID of the Cluster Descriptor they reference.

The following Descriptors reference a logical Cluster Descriptor:

• Input Terminal Descriptor

• Mixer Unit Descriptor
• Processing Unit Descriptor
• Extension Unit Descriptor

The class-specific AudioStreaming Interface Descriptor in each Alternate Setting of an AudioStreaming Interface
(except for Alternate Setting 0) reference a physical Cluster Descriptor.

Connectors Descriptors also reference a physical Cluster Descriptor to provide information about the physical
channels that travel over the various connectors, associated with a Terminal.

4.5 AUDIOCONTROL INTERFACE DESCRIPTORS

The AudioControl (AC) Interface Descriptors contain all relevant information to fully characterize the
corresponding Audio Function. The standard Interface Descriptor characterizes the Interface itself, whereas the
class-specific Interface Descriptors provide pertinent information concerning the internals of the Audio Function. It
specifies revision level information and lists the capabilities of each Entity within the Audio Function.

4.5.1 STANDARD AC INTERFACE DESCRIPTOR

The standard AC Interface Descriptor is identical to the standard interface Descriptor defined in the USB Core
Specifications, except that some fields now have dedicated values.
Table 4-11: Standard AC Interface Descriptor

Offset Field Size Value Description

0 bLength 1 Number Size of this Descriptor, in bytes: 9.
1 bDescriptorType 1 Constant INTERFACE Descriptor type.
2 bInterfaceNumber 1 Number Number of Interface. A zero-based value
identifying the index in the array of
concurrent interfaces supported by this
configuration.

Release 4.0 April 2023 89

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

3 bAlternateSetting 1 Number Value used to select an Alternate Setting for
the Interface identified in the prior field.
Shall be set to 0.
4 bNumEndpoints 1 Number Number of Endpoints used by this interface
(excluding Endpoint 0). This number is
either 0 or 1 if the optional interrupt
Endpoint is present.
5 bInterfaceClass 1 AUDIO Audio Interface Class code (assigned by the
USB). See Appendix A.5, “Audio Interface
Class Code.”
6 bInterfaceSubClass 1 AUDIO Audio Interface Subclass code. Assigned by
CONTROL this specification. See Appendix A.6, “Audio
Interface Subclass Codes.”
7 bInterfaceProtocol 1 IP_VERSION_0 Interface Protocol code. Indicates the
4_00 current version of the specification. See
Appendix A.7, “Audio Interface Protocol
Codes”
8 iInterface 1 Index Index of a String Descriptor that describes
this interface. Zero indicates that no string
Descriptor is associated with this
Descriptor.

4.5.2 CLASS-SPECIFIC AC INTERFACE DESCRIPTOR

The class-specific AC Interface Descriptor is a traditional Descriptor that contains the list of Extended Descriptor
IDs, used to fully describe the AudioControl Interface and the Audio building blocks it encapsulates. All class-
specific Clock Descriptor IDs, all Unit Descriptors ID, all Terminal Descriptor IDs, and all Power Domain Entity
Descriptor IDs shall be in that list. Cluster Descriptor IDs, Terminal Companion Descriptor IDs, and Connector
Descriptor IDs shall not be included in the list as they are indirectly referenced by other Extended Descriptors in
the list.

The order in which the Entity Descriptor IDs are reported is not important because every Descriptor can be
identified through its bDescriptorType and bDescriptorSubtype field.

The following table defines the class-specific AC Interface Descriptor.

Table 4-12: Class-specific AC Interface Descriptor

Offset Field Size Value Description

0 bLength 1 Number Size of this Descriptor, in bytes: 4+n*2.
1 bDescriptorType 1 Constant CS_INTERFACE Descriptor type.
2 bDescriptorSubtype 1 Constant AC_GENERIC Descriptor subtype.
3 bNrDescriptorIDs 1 Number Number of Descriptor IDs in the list: n.
4 waDescriptorID(1) 2 Bitmap First Descriptor ID.
… … … … …
4+(n-1)*2 waDescriptorID(n) 2 Bitmap Last Descriptor ID.

Release 4.0 April 2023 90

Universal Serial Bus Device Class Definition for Audio Devices

If there is a need to list more than 125 Descriptor IDs, then more Descriptors of subtype AC_GENERIC may be
included.

4.5.3 CLASS-SPECIFIC AC INTERFACE BUILDING BLOCK DESCRIPTORS

There are three types of Clock Entities, seven types of Units, two types of Terminals, and one type of Power
Domain Entity. Their respective Extended Descriptors are described in the following sections.

4.5.3.1 AC SELF DESCRIPTOR

The Self Descriptor provides information to the Host that is related to the functional aspects of the AudioControl
Interface itself. It indicates the presence of optional AudioControls at the Interface level.

The presence of the optional Latency Controls is advertised here in the dOptControls field of the class-specific
AudioControl Interface Descriptor and not repeated in every Terminal and Unit Descriptor. If implemented, every
Terminal and Unit within the Device shall expose a Latency Control.

Bit D1 in the dOptControls field indicates whether the Commit Capability supports Commits by CommitGroup
(D1 = 0b1) or only Function-wide Commits (D1 = 0b0).

The following table defines the Self Descriptor.

Table 4-13: Self Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes: 14.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.
4 wDescriptorSubtype 2 Constant AC_SELF Descriptor subtype.
6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.
10 dOptControls 4 Bitmap D0: Latency Controls present.
D1: Commit Capability supports
update by CommitGroup.
D31..1: Reserved.

4.5.3.2 INPUT TERMINAL DESCRIPTOR

The Input Terminal Descriptor provides information to the Host that is related to the functional aspects of the
Input Terminal.

The Input Terminal is uniquely identified by the value in the wTerminalID field. This value shall be passed in the
wEntityID field of each Command that is directed to the Terminal.

The wCSourceID field contains a constant indicating to which Clock Entity the Clock Input Pin of this Input Terminal
is connected. If the Clock Input Pin is not connected to a Clock Entity, the wCSourceID field shall be set to zero.

The wPCC field contains the Pin Channel Count value for the connection that originates at the Output Pin of the
Input Terminal.

The waClusterDescrID() array contains the IDs of the Cluster Descriptors that characterize the Clusters that may be
exposed on the Output Pin of the Input Terminal. The wNrClusterDescrIDs field contains the number of elements

Release 4.0 April 2023 91

Universal Serial Bus Device Class Definition for Audio Devices

in that array and shall always be greater than zero. For a detailed description of the Cluster Descriptor, see Section
4.4, “Cluster Descriptor”.

In case the Input Terminal represents an AudioStreaming Interface, then the Cluster Control shall be present
(dOptControls bit D0 = 0b1) and be implemented as Read-Only. The current Cluster Configuration shall be
determined by the currently selected Alternate Setting of the AudioStreaming Interface, and the Cluster Control
CUR value shall be set to the ordinal number of the currently selected Alternate Setting of the AudioStreaming
Interface (including zero when Alternate Setting zero is selected). The wNrClusterDescrIDs field shall be set to the
number of Alternate Settings the AudioStreaming Interface supports (excluding Alternate Setting zero) and the
waClusterDescrID() array shall list in order, the IDs of the Cluster Descriptors that correspond to each of the active
Alternate Settings of the AudioStreaming Interface.

In case the Input Terminal represents either an external connection via one or more Connector Entities or an
internal transducer, then, if more than one Cluster Configuration is available, the Cluster Control shall be present
(dOptControls bit D0 = 0b1) and be implemented as Read-Write. If only one Cluster Configuration is available, the
Cluster Control shall not be present (dOptControls bit D0 = 0b0).

The Cluster Active Control shall always be present (and is therefore not part of the dOptControls field).

The wTermCompDescrID field contains the unique ID of the Terminal Companion Descriptor that is associated with
this Terminal. This field shall be set to zero (no Terminal Companion Descriptor available) when the Input Terminal
represents an AudioStreaming OUT Interface. For a detailed description of the Terminal Companion Descriptor, see
Section 4.5.3.4, “Terminal Companion Descriptor.”

If an AudioStreaming Interface is associated with this Terminal, then Variant 1 of the Descriptor applies and the
wDescriptorVariant field shall be set to VARIANT_INTERFACE. The bInterfaceNumber field shall contain the
interface number of that AudioStreaming Interface.

If one or more Connector Entities are associated with the Terminal, then Variant 2 of the Descriptor applies and
the wDescriptorVariant field shall be set to VARIANT_ENTITIES. The wNrAssocEntityIDs field shall contain the
number of elements in the following waAssocEntityID() array that contains the unique IDs of those Connector
Entities.

If there are no Entities associated with this Terminal, then the wDescriptorVariant field shall be set to
VARIANT_NONE and no Descriptor Variant shall be present.

The OCN:ICN:IPN triplet used to access an AudioControl within the Input Terminal shall be set to 0:0:0.
The following table presents an outline of the Input Terminal Descriptor.
Table 4-14: Input Terminal Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes:
• VARIANT_NONE: 26+p*2.
• VARIANT_INTERFACE: 27+p*2.
• VARIANT_ENTITIES: 28+p*2+q*2.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.
4 wDescriptorSubtype 2 Constant INPUT_TERMINAL Descriptor subtype.
6 wDescriptorID 2 Number Unique ID for this Descriptor.

Release 4.0 April 2023 92

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.
10 wTerminalID 2 Number Value uniquely identifying the Terminal
within the Audio Function. This value is used
in all Commands to address this Terminal.
12 wCSourceID 2 Number ID of the Clock Entity to which this Input
Terminal is connected.
14 dOptControls 4 Bitmap D0: Cluster Control.
D1: Voltage Control
D2: Overload Control.
D3: Underflow Control.
D4: Overflow Control.
D31..5: Reserved.
18 wPCC 2 Number The Pin Channel Count for the connection
originating at the Output Pin of this Input
Terminal.
20 wNrClusterDescrIDs 2 Number Number of elements in the
waClusterDescrID() array: p.
22 waClusterDescrID(1) 2 Number ID of the first Cluster Descriptor for this Input
Terminal.
… … … … …
20+p*2 waClusterDescrID(p) 2 Number ID of the last Cluster Descriptor for this Input
Terminal.
22+p*2 wTermCompDescrID 2 Number ID of the Terminal Companion Descriptor for
this Input Terminal. Shall be set to zero if no
Terminal Companion Descriptor is present.
24+p*2 wDescriptorVariant 2 Number Variant number of the following Descriptor
Variant. Shall be either VARIANT_NONE,
VARIANT_INTERFACE or VARIANT_ENTITIES.

Variant VARIANT_INTERFACE applies when the Input Terminal is associated with an AudioStreaming Interface.

Offset Field Size Value Description

26+p*2 bInterfaceNumber 1 Number Interface number of the associated
AudioStreaming Interface.

Variant VARIANT_ENTITIES applies when the Input Terminal is associated with one or more Connector Entities.

Offset Field Size Value Description

26+p*2 wNrAssocEntityIDs 2 Number Number of elements in the
baAssocEntityID() array: q.
26+p*2+2 waAssocEntityID(1) 2 Number ID of the first Entity associated with this
Input Terminal.
… … … … …
26+p*2+q*2 waAssocEntityID(q) 2 Number ID of the last Entity associated with this Input
Terminal.

Release 4.0 April 2023 93

Universal Serial Bus Device Class Definition for Audio Devices

4.5.3.3 OUTPUT TERMINAL DESCRIPTOR

The Output Terminal Descriptor provides information to the Host that is related to the functional aspects of the
Output Terminal.

The Output Terminal is uniquely identified by the value in the wTerminalID field. This value shall be passed in the
wEntityID field of each Command that is directed to the Terminal.

The wSourceID field is used to describe the connectivity for this Terminal. It contains the ID of the Unit or Terminal
to which this Output Terminal is connected via its Input Pin. The Cluster Descriptor, describing the logical channels
entering the Output Terminal is not repeated here. It is up to the Host software to trace the connection ‘upstream’
to locate the Cluster Descriptor pertaining to this Cluster.

The wCSourceID field contains a constant indicating to which Clock Entity the Clock Input Pin of this Output
Terminal is connected. If the Clock Input Pin is not connected to a Clock Entity, the wCSourceID field shall be set to
zero.

The PCC value for the Input Pin of the Output Terminal is not included here and is inherited from the first upstream
Entity that defines a PCC value on its Output Pin.

The wTermCompDescrID field contains the unique ID of the Terminal Companion Descriptor that is associated with
this Terminal. This field shall be set to zero (no Terminal Companion Descriptor available) when the Output
Terminal represents an AudioStreaming IN Interface. For a detailed description of the Terminal Companion
Descriptor, see Section 4.5.3.4, “Terminal Companion Descriptor.”

If an AudioStreaming Interface is associated with this Terminal, then the AudioStreaming Variant of the Descriptor
applies and the wDescriptorVariant field shall be set to VARIANT_INTERFACE. The bInterfaceNumber field shall
contain the interface number of that AudioStreaming Interface.

If one or more Connector Entities are associated with the Terminal, then the Entities Variant of the Descriptor
applies and the wDescriptorVariant field shall be set to VARIANT_ENTITIES. The wNrAssocEntityIDs field shall
contain the number of elements in the following waAssocEntityID() array that contains the unique IDs of those
Connector Entities.

If there are no Entities associated with this Terminal, then the wDescriptorVariant field shall be set to
VARIANT_NONE and no Descriptor Variant shall be present.

The OCN:ICN:IPN triplet used to access an AudioControl within the Output Terminal shall be set to 0:0:0.
The following table presents an outline of the Output Terminal Descriptor.
Table 4-15: Output Terminal Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes:
• VARIANT_NONE: 24.
• VARIANT_INTERFACE: 25.
• VARIANT_ENTITIES: 24+q*2.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.
4 wDescriptorSubtype 2 Constant OUTPUT_TERMINAL Descriptor subtype.
6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.

Release 4.0 April 2023 94

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

10 wTerminalID 2 Number Value uniquely identifying the Terminal
within the Audio Function. This value is
used in all Commands to address this
Terminal.
12 wSourceID 2 Number ID of the Unit or Terminal to which this
Terminal is connected.
14 wCSourceID 2 Number ID of the Clock Entity to which this Output
Terminal is connected.
16 dOptControls 4 Bitmap D0: Overload Control.
D1: Underflow Control.
D2: Overflow Control.
D31..3: Reserved.
20 wTermCompDescrID 2 Number ID of the Terminal Companion Descriptor
for this Output Terminal. Shall be set to
zero if no Terminal Companion Descriptor is
present.
22 wDescriptorVariant 2 Number Variant number of the following Descriptor
Variant. Shall be either VARIANT_NONE,
VARIANT_INTERFACE or VARIANT_ENTITIES.

Variant VARIANT_INTERFACE applies when the Output Terminal is associated with an AudioStreaming Interface.

Offset Field Size Value Description

24 bInterfaceNumber 1 Number Interface number of the associated
AudioStreaming Interface.

Variant VARIANT_ENTITIES applies when the Output Terminal is associated with one or more Connector Entities.

Offset Field Size Value Description

24 wNrAssocEntityIDs 2 Number Number of elements in the
baAssocEntityID() array: q.
26 waAssocEntityID(1) 2 Number ID of the first Entity associated with this
Output Terminal.
… … … … …
24+q*2 waAssocEntityID(q) 2 Number ID of the last Entity associated with this
Output Terminal.

4.5.3.4 TERMINAL COMPANION DESCRIPTOR

The Terminal Companion Descriptor optionally returns additional physical information about the channels that
enter or leave the Terminal.

Release 4.0 April 2023 95

Universal Serial Bus Device Class Definition for Audio Devices

Figure 4-7: Terminal Companion Descriptor

Header

Common Block

Channel 1 Block

Channel n Block

The Terminal Companion Descriptor consists of a fixed Header, followed by an optional Common Block, followed
by as many Channel Blocks as there are channels in the Cluster of the Terminal.

4.5.3.4.1 TERMINAL COMPANION DESCRIPTOR HEADER

The Header starts with the wLength field that contains the number of bytes in the Terminal Companion Descriptor
Header.

The wTotalLength field contains the number of bytes in the entire Terminal Companion Descriptor, including the
Header and all Terminal Companion Descriptor Blocks.

The wNrChannels field indicates the number of audio channels present in the logical Cluster of the Terminal.
Table 4-16: Terminal Companion Descriptor Header

Offset Field Size Value Description

0 wLength 2 Number Length of the Terminal Companion
Descriptor Header, in bytes: 14.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.
4 wDescriptorSubtype 2 Constant TERMINAL_COMPANION Descriptor
subtype.
6 wDescriptorID 2 Number Unique ID of this Terminal Companion
Descriptor.
8 wStrDescriptorID 2 Number Unique ID of String Descriptor.

Release 4.0 April 2023 96

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

10 wTotalLength 2 Number Total length of the Terminal Companion
Descriptor, including the Terminal
Companion Descriptor Header and all
included Terminal Companion Descriptor
Blocks.
12 wNrChannels 2 Number Number of channels present in the Cluster
of the Terminal.

4.5.3.4.2 TERMINAL COMPANION DESCRIPTOR BLOCK

The Terminal Companion Descriptor Header is followed by one or more Terminal Companion Descriptor Blocks.
There is an optional Common Block, followed by as many Channel Blocks as there are channels in the Cluster of the
Terminal. Each Block consists of one or more Segments.

The Common Block consists of Segments that contain relevant information about characteristics of the Terminal as
a whole. All Common Block Segments are optional.

A Channel Block consists of Segments that contain relevant information about that channel’s physical
characteristics. All Channel Block Segments are optional. It is strongly recommended that the same layout is used
for each Channel Block, i.e., the same Segments appear in the same order in each Channel Block. Figure 4-8 further
illustrates the above concepts.
Figure 4-8: Terminal Companion Channel Block

Block

Segment 1

Segment 2

Segment m

End Segment

4.5.3.4.3 SEGMENTS

There are two types of Segments. Common Block Segments contain pertinent information about the Terminal as a
whole. Channel Block Segments contain pertinent information about certain aspects of a particular channel in the
Cluster of the Terminal. Both Segment types share the same layout.

The wSegmentType field describes the Segment Type (Common Block or Channel Block) and the type of content
contained in the Segment.

The layout for a Segment is always as follows:

Release 4.0 April 2023 97

Universal Serial Bus Device Class Definition for Audio Devices

Table 4-17: Terminal Companion Descriptor Segment

Offset Field Size Value Description

0 wLength 2 Number Length of the Segment, in bytes: 4+n.
2 wSegmentType 2 Constant Describes the Segment Type and the type of
content in the Segment.
4 Segment-specific n Segment-specific content.

4.5.3.4.3.1 END BLOCK SEGMENT

Each Block is terminated by an End Block Segment. The End Block Segment marks the end of the variable length
Block. The End Block Segment does not have a Segment-specific section and is structured as follows:
Table 4-18: End Block Segment

Offset Field Size Value Description

0 wLength 2 Number Length of the Segment, in bytes: 4.
2 wSegmentType 2 Constant TERMINAL_COMPANION_END_BLOCK.

4.5.3.4.3.2 COMMON BLOCK SEGMENTS

The following Common Block Segment types are defined:

• EN 50332-2 Acoustic level

• EN 50332-2 Voltage level

Values for the Common Block Segment types can be found in Appendix A.18, “Terminal Companion Segment
Types.”

4.5.3.4.3.2.1 EN 50332-2 ACOUSTIC LEVEL SEGMENT

The EN 50332-2 Acoustic Level Segment contains, in the wOutputlevel field, the measured headphone transducer
output level in units of dB SPL A-weighted when all controls affecting the output are set to maximize the output
level, following the procedures set forth within EN 50332-2:2013. The value may range from 0.00 dB(A) to
+255.9961 dB(A) (0xFFFF) in steps of 1/256 dB(A) or 0.00390625 dB(A) (0x0001).
Only Output Terminals that represent headphone transducers shall be permitted to have this segment.

See [EN 50332-2:2013] for more details.

Table 4-19: EN 50332-2 Acoustic Level Segment

Offset Field Size Value Description

0 wLength 2 Number Length of the Segment, in bytes: 6.
2 wSegmentType 2 Constant EN_50332-2_SPL_LEVEL.
4 wOutputLevel 2 Number Maximum achievable output level
expressed in dB SPL (A-weighted) when
rendering the EN 50332-2:2013 test signal.

4.5.3.4.3.2.2 EN 50332-2 VOLTAGE LEVEL SEGMENT

Release 4.0 April 2023 98

Universal Serial Bus Device Class Definition for Audio Devices

The EN 50332-2 Voltage Level Segment contains, in the wOutputlevel field, the measured headphone jack output
level in units of millivolts rms (mVrms) unweighted, when all controls affecting the output are set to maximize the
output level, following the procedures set forth within EN 50332-2:2013. The value may range from 0.00 mVrms to
+65,635 mVrms (0xFFFF) in steps of 1 mVrms (0x0001).

Only Output Terminals that carry electrical analog audio signals shall be permitted to have this segment:

See [EN 50332-2:2013] for more details.

Table 4-20: EN 50332-2 Voltage Level Segment

Offset Field Size Value Description

0 wLength 2 Number Length of the Segment, in bytes: 6.
2 wSegmentType 2 Constant EN_50332-2_VOLTAGE_LEVEL.
4 wOutputLevel 2 Number Maximum achievable output voltage,
expressed in unweighted mVrms, when
rendering the EN 50332-2:2013 test signal.

4.5.3.4.3.3 CHANNEL BLOCK SEGMENTS

The following Channel Block Segment types are defined:

• Bandwidth
• Magnitude Response
• Magnitude/Phase Response
• Position

Values for the Channel Block Segment types can be found in Appendix A.18, “Terminal Companion Segment
Types.”

4.5.3.4.3.3.1 BANDWIDTH SEGMENT

The Bandwidth Segment contains basic information about the audio bandwidth in the channel. The bandwidth is
specified by providing the lower and upper -3 dB frequency points, in Hz, of the available band in the dMinFreq
and dMaxFreq fields.
Table 4-21: Bandwidth Segment

Offset Field Size Value Description

0 wLength 2 Number Length of the Segment, in bytes: 12.
2 wSegmentType 2 Constant CHANNEL_BANDWIDTH.
4 dMinFreq 4 Number Lower -3 dB frequency point.
8 dMaxFreq 4 Number Upper -3 dB frequency point.

4.5.3.4.3.3.2 MAGNITUDE RESPONSE

The Magnitude Response Segment contains detailed information about the magnitude of the transfer function
(frequency response) of the channel. The magnitude is specified as an array of [frequency point, magnitude] pairs.
The frequency values are specified in Hz and the magnitude values may range from +127.9961 dB (0x7FFF) down

Release 4.0 April 2023 99

Universal Serial Bus Device Class Definition for Audio Devices

to -127.9961 dB (0x8001) in steps of 1/256 dB or 0.00390625 dB (0x0001). In addition, code 0x8000, representing
silence (i.e., - dB), may be used as well.
Table 4-22: Magnitude Segment

Offset Field Size Value Description

0 wLength 2 Number Length of the Segment, in bytes: 10+n*6.
2 wSegmentType 2 Constant CHANNEL_MAGNITUDE_RESPONSE.
4 wNrFreqPoints 2 Number The number of frequency points, n.
6 dFreq(1) 4 Number First frequency value.
10 wMagnitude(1) 2 Number First magnitude value.
… … … …
4+n*6 dFreq(n) 4 Number Last frequency value.
8+n*6 wMagnitude(n) 2 Number Last magnitude value.

4.5.3.4.3.3.3 MAGNITUDE/PHASE RESPONSE

The Magnitude/Phase Response Segment contains detailed information about the magnitude and phase of the
transfer function (frequency response) of the channel. The magnitude/phase is specified as an array of [frequency
point, magnitude, phase] triplets. The frequency values are specified in Hz. The magnitude values may range from
+127.9961 dB (0x7FFF) down to -127.9961 dB (0x8001) in steps of 1/256 dB or 0.00390625 dB (0x0001). In
addition, code 0x8000, representing silence (i.e., - dB), may be used as well. The phase values may range from
+0.99996948242 *  down to - (0x8000) in steps of 1/32768 *  (0x0001).
Table 4-23: Magnitude/Phase Segment

Offset Field Size Value Description

0 wLength 2 Number Length of the Segment, in bytes: 10+n*8.
2 wSegmentType 2 Constant CHANNEL_MAGNITUDE/PHASE_RESPONSE.
4 wNrFreqPoints 2 Number The number of frequency points, n.
6 dFreq(1) 4 Number First frequency value.
10 wMagnitude(1) 2 Number First magnitude value.
12 wPhase(1) 2 Number First phase value.
… … … …
4+n*8 dFreq(n) 4 Number Last frequency value.
6+n*8 wMagnitude(n) 2 Number Last magnitude value.
8+n*8 wPhase(n) 2 Number Last phase value.

4.5.3.4.3.3.4 POSITION_XYZ SEGMENT

The Position_XYZ Segment contains the (X, Y, Z) Cartesian coordinates of the source or sink associated with the
channel. The X, Y, and Z values are expressed in micrometers (µm) and are relative to an unspecified origin at (0, 0,
0). The Audio Function may have out-of-band means to indicate to the Host where the actual origin of the
coordinate system is located on the Device. The value 0xFFFFFFFF has special meaning. It is used to indicate that
the coordinate in a particular dimension is unknown.

Release 4.0 April 2023 100

Universal Serial Bus Device Class Definition for Audio Devices

Table 4-24: Position_XYX Segment

Offset Field Size Value Description

0 wLength 2 Number Length of the Segment, in bytes: 16.
2 wSegmentType 2 Constant CHANNEL_POSITION_XYZ.
4 dX 4 Number X-coordinate of the audio source or sink,
associated with the channel.
8 dY 4 Number Y-coordinate of the audio source or sink,
associated with the channel.
12 dZ 4 Number Z-coordinate of the audio source or sink,
associated with the channel.

4.5.3.4.3.3.5 POSITION_RΘΦ SEGMENT

The Position_ RΘΦ Segment contains the (R, Θ, Φ) spherical coordinates of the source or sink associated with the
channel. The R, Θ, and Φ values are expressed in µm, µrad and µrad respectively, and are relative to an unspecified
origin at (0, 0, 0). The Audio Function may have out-of-band means to indicate to the Host where the actual origin
of the coordinate system is located on the Device. The value 0xFFFFFFFF has special meaning. It is used to indicate
that the coordinate in a particular dimension is unknown.
Table 4-25: Position_RΘΦ Segment

Offset Field Size Value Description

0 wLength 2 Number Length of the Segment, in bytes: 16.
2 wSegmentType 2 Constant CHANNEL_POSITION_ R_THETA_PHI.
4 dR 4 Number R-coordinate of the audio source or sink,
associated with the channel.
8 dΘ 4 Number Θ-coordinate of the audio source or sink,
associated with the channel.
12 dΦ 4 Number Φ-coordinate of the audio source or sink,
associated with the channel.

4.5.3.5 MIXER UNIT DESCRIPTOR

The Mixer Unit is uniquely identified by the value in the wUnitID field of the Mixer Unit Descriptor. This value shall
be passed in the wEntityID field of each Command that is directed to the Mixer Unit.

The wNrInputPins field contains the number of Input Pins (𝑃) of the Mixer Unit. This equals the number of Clusters
that enter the Mixer Unit. The connectivity of the Input Pins is described via the waSourceID() array, containing 𝑃
elements. The index 𝑝 into the array is one-based and directly related to the Input Pin numbers. waSourceID(𝑝)
contains the ID of the Unit or Terminal to which Input Pin 𝑝 is connected.

The Cluster Descriptors, describing the logical channels entering the Mixer Unit are not repeated here. It is up to
the Host software to trace the connections ‘upstream’ to locate the Cluster Descriptors pertaining to the Clusters
and to determine the number of logical channels 𝑛𝑝 in each incoming 𝐶𝑙𝑢𝑠𝑡𝑒𝑟𝑝 .

The PCC value for each of the Input Pins of the Mixer Unit is inherited from the first upstream Entity that defines a
PCC value on its Output Pin.

Release 4.0 April 2023 101

Universal Serial Bus Device Class Definition for Audio Devices

The wPCC field contains the Pin Channel Count value for the connection that originates at the Output Pin of the
Mixer Unit. This also determines the number of Output Pin channels the Mixer Unit supports.

The Input Pin Number (IPN), the Input Channel Number (ICN) and the Output Channel Number (OCN) are used to
access a Mixer Control within the Mixer Unit and are defined as follows:

If 𝑃 is the number of Input Pins, and 𝑁𝑝 is the PCC for Input Pin 𝑝, and 𝑀 is the PCC for the Output Pin, then the
Input Pin Number IPN, the Input Channel Number 𝐼𝐶𝑁𝑝 , and the Output Channel Number 𝑂𝐶𝑁 for the Mixer
Control residing on the crossing of Input Pin 𝑝, Input Pin channel 𝑞 and Output Pin channel 𝑚 is:

𝐼𝑃𝑁 = 𝑝; 𝐼𝐶𝑁𝑝 = 𝑞; and 𝑂𝐶𝑁 = 𝑚

where (1 ≤ 𝑝 ≤ 𝑃), (1 ≤ 𝑞 ≤ 𝑁𝑝 ), (1 ≤ 𝑚 ≤ 𝑀).

Note that if 𝑛𝑝 < 𝑁𝑝 , interaction with an AudioControl that resides on an Input Pin channel outside the range
[0. . 𝑛𝑝 ] is still possible, i.e. the accessibility of an AudioControl shall not depend on the current Cluster
Configuration.

The following figure presents a more graphical explanation.

Release 4.0 April 2023 102

Universal Serial Bus Device Class Definition for Audio Devices

Figure 4-9: Mixer internals

Because the Mixer Unit may redefine the spatial locations of the logical output channels, contained in its output
Cluster, there is a need for a Mixer output Cluster Descriptor.

The waClusterDescrID() array contains the IDs of the Cluster Descriptors that characterize the Clusters that may be
exposed on the Output Pin of the Mixer Unit. A value of zero in an array element indicates that the output Cluster
is inherited from the current Cluster on Input Pin one (the dominant Input Pin). The wNrClusterDescrIDs field
contains the number of elements in that array and shall always be greater than zero. For a detailed description of
the Cluster Descriptor, see Section 4.4, “Cluster Descriptor”. If the Mixer Unit exposes more than one Cluster
Configuration on its Output Pin (wNrClusterDescrIDs > 1), then the Cluster Control shall be present. However, the
Cluster Active Control shall always be present (and is therefore not part of the dOptControls field).

The following table details the structure of the Mixer Unit Descriptor.
Table 4-26: Mixer Unit Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes:
22+p*2+q*2.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.

Release 4.0 April 2023 103

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

4 wDescriptorSubtype 2 Constant MIXER_UNIT Descriptor subtype.
6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.
10 wUnitID 2 Number Value uniquely identifying the Unit within
the Audio Function. This value is used in all
Commands to address this Unit.
12 dOptControls 4 Bitmap D0: Cluster Control.
D1: Underflow Control.
D2: Overflow Control.
D31..3: Reserved.
16 wPCC 2 Number The Pin Channel Count value for the
connection that originates at the Output Pin
of the Mixer Unit.
18 wNrInputPins 2 Number Number of Input Pins of this Unit: p
20 waSourceID(1) 2 Number ID of the Unit or Terminal to which the first
Input Pin of this Mixer Unit is connected.
… … … … …
18+p*2 waSourceID (p) 2 Number ID of the Unit or Terminal to which the last
Input Pin of this Mixer Unit is connected.
20+p*2 wNrClusterDescrIDs 2 Number Number of elements in the
waClusterDescrID() array: p.
22+p*2 waClusterDescrID(1) 2 Number ID of the first Cluster Descriptor for this
Mixer Unit.
… … … … …
20+p*2+q*2 waClusterDescrID(q) 2 Number ID of the last Cluster Descriptor for this
Mixer Unit.

4.5.3.6 SELECTOR UNIT DESCRIPTOR

The Selector Unit is uniquely identified by the value in the wUnitID field of the Selector Unit Descriptor. This value
shall be passed in the wEntityID field of each Command that is directed to the Selector Unit.
The wNrInputPins field contains the number of Input Pins (𝑝) of the Selector Unit. The connectivity of the Input
Pins is described via the waSourceID() array that contains p elements. The index 𝑖 into the array is one-based and
directly related to the Input Pin numbers. waSourceID(𝑖) contains the ID of the Unit or Terminal to which Input Pin
𝑖 is connected.

The Selector Unit does not redefine the Cluster on its Output Pin. Rather, it always inherits the Cluster from its
currently selected Input Pin. If supported, setting the Selector Control to zero (no connect), switches the output
Cluster into the Inactive state.

The Selector Unit Descriptor does not contain a PCC value. The implied PCC value on its Output Pin shall be derived
from the maximum of the inherited PCC values on all its Input Pins.

The OCN:ICN:IPN triplet used to access an AudioControl within the Selector Unit shall be set to 0:0:0.

Release 4.0 April 2023 104

Universal Serial Bus Device Class Definition for Audio Devices

The following table details the structure of the Selector Unit Descriptor.
Table 4-27: Selector Unit Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes: 18+p.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.

4 wDescriptorSubtype 2 Constant SELECTOR_UNIT Descriptor subtype.

6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.
10 wUnitID 2 Number Value uniquely identifying the Unit within
the Audio Function. This value is used in all
Commands to address this Unit.
12 dOptControls 4 Bitmap D31..0: Reserved.
16 wNrInputPins 2 Number Number of Input Pins of this Unit: p
18 waSourceID(1) 2 Number ID of the Unit or Terminal to which the first
Input Pin of this Selector Unit is connected.
… … … … …
16+p waSourceID (p) 2 Number ID of the Unit or Terminal to which the last
Input Pin of this Selector Unit is connected.

4.5.3.7 FEATURE UNIT DESCRIPTOR

The Feature Unit is uniquely identified by the value in the wUnitID field of the Feature Unit Descriptor. This value
shall be passed in the wEntityID field of each Command that is directed to the Feature Unit.

The wSourceID field is used to describe the connectivity for this Feature Unit. It contains the ID of the Unit or
Terminal to which this Feature Unit is connected via its Input Pin. The Cluster Descriptor, describing the 𝑛 logical
channels entering the Feature Unit is not repeated here. It is up to the Host software to trace the connection
‘upstream’ to locate the Cluster Descriptor pertaining to this Cluster. The Feature Unit does not redefine the
output Cluster and thus inherits the output Cluster from the input Cluster.

The Feature Unit Descriptor does not contain a PCC value. The implied PCC value on its Output Pin shall be derived
from the inherited PCC value on its Input Pin.
The OCN = ICN used to access an AudioControl within the Feature Unit is the Pin channel number on which the
AudioControl resides and is in the range [0. . 𝑁] where zero represents the Primary channel and 𝑁 is the number of
Pin channels of the Feature Unit (𝑃𝐶𝐶 = 𝑁). Note that if 𝑛 < 𝑁, interaction with an AudioControl with an OCN =
ICN outside the range [0. . 𝑛] is still possible, i.e., the accessibility of an AudioControl shall not depend on the
current Cluster Configuration. The IPN shall be set to one.

The layout of the Feature Unit Descriptor is detailed in the following table.
Table 4-28: Feature Unit Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes: 18+N*4
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.

Release 4.0 April 2023 105

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

4 wDescriptorSubtype 2 Constant FEATURE_UNIT Descriptor subtype.
6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.
10 wUnitID 2 Number Value uniquely identifying the Unit within
the Audio Function. This value is used in all
Commands to address this Unit.
12 wSourceID 2 Number ID of the Unit or Terminal to which this
Feature Unit is connected.
14 dOptControls(0) 4 Bitmap The AudioControls bitmap for Primary
channel 0:
D0: Bypass Control.
D1: Mute Control.
D2: Gain Control.
D3: Bass Control.
D4: Mid Control.
D5: Treble Control.
D6: Graphic Equalizer Control.
D7: Automatic Gain Control.
D8: Delay Control.
D9: Bass Boost Control.
D10: Loudness Control.
D11: Input Gain Pad Control.
D12: Phase Inverter Control.
D13: Underflow Control.
D14: Overflow Control.
D31..15: Reserved.
18 dOptControls(1) 4 Bitmap The AudioControls bitmap for Pin channel 1.
… … … … …
14+N*4 dOptControls(N) 4 Bitmap The AudioControls bitmap for Pin channel
N.

4.5.3.8 SAMPLING RATE CONVERTER UNIT DESCRIPTOR

The Sampling Rate Converter Unit Descriptor provides information to the Host that is related to the functional
aspects of the SRC Unit.

The SRC Unit is uniquely identified by the value in the wUnitID field. This value shall be passed in the wEntityID
field of each Command that is directed to the Feature Unit.

The wSourceID field is used to describe the connectivity for this SRC Unit. It contains the ID of the Unit or Terminal
to which this SRC Unit is connected via its Input Pin. The Cluster Descriptor, describing the logical channels entering
the SRC Unit is not repeated here. It is up to the Host software to trace the connection ‘upstream’ to locate the
Cluster Descriptor pertaining to this Cluster. The SRC Unit does not redefine the output Cluster and thus inherits
the output Cluster from the input Cluster.

Release 4.0 April 2023 106

Universal Serial Bus Device Class Definition for Audio Devices

The wCSourceInID field contains the ID of the Clock Entity associated with the audio Input Pin. If the audio Input
Pin is not associated with a Clock Entity, then the wCSourceInID shall be set to zero.

The wCSourceOutID field contains the ID of the Clock Entity associated with the audio Output Pin. If the audio
Output Pin is not associated with a Clock Entity, then the wCSourceOutID shall be set to zero.

Note: For the SRC Unit to be useful, at least one of the Clock Input Pins shall be connected to a Clock Entity.

The SRC Unit Descriptor does not contain a PCC value. The implied PCC value on its Output Pin shall be derived
from the inherited PCC value on its Input Pin.

The following table presents an outline of the SRC Unit Descriptor.

Table 4-29: Sampling Rate Converter Unit Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes: 22.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.
4 wDescriptorSubtype 2 Constant SAMPLE_RATE_CONVERTER Descriptor
subtype.
6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.
10 wUnitID 2 Number Value uniquely identifying the Unit within
the Audio Function. This value is used in all
Commands to address this Unit.
12 wSourceID 2 Number ID of the Unit or Terminal to which this SRC
Unit is connected.
14 wCSourceInID 2 Number ID of the Clock Entity to which this SRC Unit
input section is connected.
16 wCSourceOutID 2 Number ID of the Clock Entity to which this SRC Unit
output section is connected.
18 dOptControls 4 Bitmap D0: Underflow Control.
D1: Overflow Control.
D31..2: Reserved.

4.5.3.9 EFFECT UNIT DESCRIPTOR

The Effect Unit is uniquely identified by the value in the wUnitID field of the Effect Unit Descriptor. This value shall
be passed in the wEntityID field of each Command that is directed to the Effect Unit.

The wEffectType field contains a value that fully identifies the Effect Unit. For a list of all supported Effect Unit
Types, see Appendix A.19, “Effect Unit Effect Types.”

The wSourceID field is used to describe the connectivity for this Effect Unit. It contains the ID of the Unit or
Terminal to which this Effect Unit is connected via its Input Pin. The Cluster Descriptor, describing the 𝑛 logical
channels entering the Effect Unit is not repeated here. It is up to the Host software to trace the connection
‘upstream’ to locate the Cluster Descriptor pertaining to this Cluster. The Effect Unit does not redefine the output
Cluster and thus inherits the output Cluster from the input Cluster.

Release 4.0 April 2023 107

Universal Serial Bus Device Class Definition for Audio Devices

The Effect Unit Descriptor does not contain a PCC value. The implied PCC value on its Output Pin shall be derived
from the inherited PCC value on its Input Pin.

The OCN = ICN used to access an AudioControl within the Effect Unit is the Pin channel number on which the
AudioControl resides and is in the range [0. . 𝑁] where zero represents the Primary channel and 𝑁 is the number of
Pin channels of the Effect Unit (𝑃𝐶𝐶 = 𝑁). Note that if 𝑛 < 𝑁, interaction with an AudioControl with an OCN =
ICN outside the range [0. . 𝑛] is still possible, i.e., the accessibility of an AudioControl shall not depend on the
current Cluster Configuration. The IPN shall be set to one.

The following table outlines the Effect Unit Descriptor.

Table 4-30: Effect Unit Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes: 20+N*4.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.
4 wDescriptorSubtype 2 Constant EFFECT_UNIT Descriptor subtype.
6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.
10 wUnitID 2 Number Value uniquely identifying the Unit within
the Audio Function. This value is used in all
Commands to address this Unit.
12 wEffectType 2 Constant Constant identifying the type of effect this
Unit is performing.
14 wSourceID 2 Number ID of the Unit or Terminal to which this
Effect Unit is connected.
16 daOptControls(0) 4 Bitmap The AudioControls bitmap for Primary
channel 0:
D31..0: Effect-specific allocation.
… … … … …
16+N*4 daOptControls(N) 4 Bitmap The AudioControls bitmap for Pin channel
N:
D31..0: Effect-specific allocation.

4.5.3.9.1 PARAMETRIC EQUALIZER SECTION EFFECT UNIT DESCRIPTOR

The wEffectType field of the common Effect Unit Descriptor contains the value PARAM_EQ_SECTION_EFFECT. (See
Appendix A.19, “Effect Unit Effect Types.”

The following table outlines the PEQS Effect Unit Descriptor. It is identical to the common Effect Unit Descriptor,
except for some field values. It is repeated here for clarity.
Table 4-31: Parametric Equalizer Section Effect Unit Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes: 20+N*4.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.
4 wDescriptorSubtype 2 Constant EFFECT_UNIT Descriptor subtype.

Release 4.0 April 2023 108

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.
10 wUnitID 2 Number Value uniquely identifying the Unit within
the Audio Function. This value is used in all
Commands to address this Unit.
12 wEffectType 2 Constant PARAM_EQ_SECTION_EFFECT effect type.
14 wSourceID 2 Number ID of the Unit or Terminal to which this
Effect Unit is connected.
16 daOptControls(0) 4 Bitmap The AudioControls bitmap for Primary
channel 0:
D0: Bypass Control.
D1: Center Frequency Control.
D2: Q Factor Control.
D3: Gain Control.
D4: Underflow Control.
D5: Overflow Control.
D31..6: Reserved.
… … … … …
16+N*4 daOptControls(N) 4 Bitmap The AudioControls bitmap for Pin channel
N.

4.5.3.9.2 REVERBERATION EFFECT UNIT DESCRIPTOR

The wEffectType field of the common Effect Unit Descriptor contains the value REVERBERATION_EFFECT. See
Appendix A.19, “Effect Unit Effect Types.”

The following table outlines the Reverberation Effect Unit Descriptor. It is identical to the common Effect Unit
Descriptor, but extended with some additional fields.

The wNrTypes field indicates how many different Types of Reverberation the Unit supports. The
wTypeStrDescriptorID() array points to a human-readable string that describes the Reverberation effect for each
of the supported Reverberation Types. The wTypeStrDescriptorID() array is optional but when present shall
contain an entry for each supported Reverberation Types. The presence of the wTypeStrDescriptorID() array shall
be derived from the value in the wLength field of the Descriptor.
Table 4-32: Reverberation Effect Unit Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes:
22+N*4+n*2.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.
4 wDescriptorSubtype 2 Constant EFFECT_UNIT Descriptor subtype.
6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.

Release 4.0 April 2023 109

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

10 wUnitID 2 Number Value uniquely identifying the Unit within
the Audio Function. This value is used in all
Commands to address this Unit.
12 wEffectType 2 Constant REVERBERATION_EFFECT effect type.
14 wSourceID 2 Number ID of the Unit or Terminal to which this
Effect Unit is connected.
16 daOptControls(0) 4 Bitmap The AudioControls bitmap for Primary
channel 0:
D0: Bypass Control.
D1: Type Control.
D2: Level Control.
D3: Time Control.
D4: Delay Feedback Control.
D5: Pre-Delay Control.
D6: Density Control.
D7: Hi-Freq Roll-Off Control.
D8: Underflow Control.
D9: Overflow Control.
D31..10: Reserved.
… … … … …
16+N*4 daOptControls(N) 4 Bitmap The AudioControls bitmap for Pin channel
N.
20+N*4 wNrTypes 2 Number The number of Reverberation Types
supported by this Unit.
22+N*4 wTypeStrDescriptorID(1) 2 Number Unique ID for a Type String Descriptor.
… … … … …
20+N*4+n*2 wTypeStrDescriptorID(n) 2 Number Unique ID for a Type String Descriptor.

4.5.3.9.3 MODULATION DELAY EFFECT UNIT DESCRIPTOR

The wEffectType field of the common Effect Unit Descriptor contains the value MOD_DELAY_EFFECT. See
Appendix A.19, “Effect Unit Effect Types.”

The following table outlines the Modulation Delay Effect Unit Descriptor. It is identical to the common Effect Unit
Descriptor, except for some field values. It is repeated here for clarity.
Table 4-33: Modulation Delay Effect Unit Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes: 20+N*4.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.
4 wDescriptorSubtype 2 Constant EFFECT_UNIT Descriptor subtype.
6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.

Release 4.0 April 2023 110

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

10 wUnitID 2 Number Value uniquely identifying the Unit within
the Audio Function. This value is used in all
Commands to address this Unit.
12 wEffectType 2 Constant MOD_DELAY_EFFECT effect type.
14 wSourceID 2 Number ID of the Unit or Terminal to which this
Effect Unit is connected.
16 daOptControls(0) 4 Bitmap The AudioControls bitmap for Primary
channel 0:
D0: Bypass Control.
D1: Balance Control.
D2: Rate Control.
D3: Depth Control.
D4: Time Control.
D5: Feedback Level Control.
D6: Underflow Control.
D7: Overflow Control.
D31..8: Reserved.
… … … … …
16+N*4 daOptControls(N) 4 Bitmap The AudioControls bitmap for Pin channel
N.

4.5.3.9.4 DYNAMIC RANGE COMPRESSOR/EXPANDER EFFECT UNIT DESCRIPTOR

The wEffectType field of the common Effect Unit Descriptor contains the value DYN_RANGE_COMP_EXP_EFFECT.
See Appendix A.19, “Effect Unit Effect Types.”

The following table outlines the Dynamic Range Compressor/Expander Effect Unit Descriptor. It is identical to the
common Effect Unit Descriptor, except for some field values. It is repeated here for clarity.
Table 4-34: Dynamic Range Compressor/Expander Effect Unit Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes: 20+N*4.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.
4 wDescriptorSubtype 2 Constant EFFECT_UNIT Descriptor subtype.
6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.
10 wUnitID 2 Number Value uniquely identifying the Unit within
the Audio Function. This value is used in all
Commands to address this Unit.
12 wEffectType 2 Constant DYN_RANGE_COMP_EXP_EFFECT effect
type.
14 wSourceID 2 Number ID of the Unit or Terminal to which this
Effect Unit is connected.

Release 4.0 April 2023 111

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

16 daOptControls(0) 4 Bitmap The AudioControls bitmap for Primary
channel 0:
D0: Bypass Control.
D1: Compression Ratio Control.
D2: MaxAmpl Control.
D3: Threshold Control.
D4: Attack Time Control.
D5: Release Time Control.
D6: Underflow Control.
D7: Overflow Control.
D31..8: Reserved.
… … … … …
16+N*4 daOptControls(N) 4 Bitmap The AudioControls bitmap for Pin channel
N.

4.5.3.10 PROCESSING UNIT DESCRIPTOR

The Processing Unit is uniquely identified by the value in the wUnitID field of the Processing Unit Descriptor. This
value shall be passed in the wEntityID field of each Command that is directed to the Processing Unit.

The wProcessType field contains a value that fully identifies the Processing Unit. For a list of all supported
Processing Unit Types, see Appendix A.20, “Processing Unit Process Types.”

The wNrInputPins field contains the number of Input Pins (𝑝) of the Processing Unit. The connectivity of the Input
Pins is described via the waSourceID() array that contains p elements. The index 𝑖 into the array is one-based and
directly related to the Input Pin numbers. waSourceID(𝑖) contains the ID of the Unit or Terminal to which Input Pin
𝑖 is connected. The Cluster Descriptors, describing the logical channels entering the Processing Unit are not
repeated here. It is up to the Host software to trace the connections ‘upstream’ to locate the Cluster Descriptors
pertaining to the Clusters.

The PCC value for each of the Input Pins of the Processing Unit is inherited from the first upstream Entity that
defines a PCC value on its Output Pin.

The wPCC field contains the Pin Channel Count value for the connection that originates at the Output Pin of the
Processing Unit. This also determines the number of Output Pin channels the Processing Unit supports.

Because the Processing Unit can freely redefine the output Cluster Configuration, possibly based on the values of
some internal AudioControls, there is a need for an output Cluster Descriptor array.

The waClusterDescrID() array contains the IDs of the Cluster Descriptors that characterize the Clusters that may be
exposed on the Output Pin of the Processing unit. A value of zero in an array element indicates that the output
Cluster is inherited from the current Cluster on Input Pin one (the dominant Input Pin). The wNrClusterDescrIDs
field contains the number of elements in that array and shall always be greater than zero. For a detailed
description of the Cluster Descriptor, see Section 4.4, “Cluster Descriptor”. If the Processing Unit exposes more
than one Cluster Configuration on its Output Pin (wNrClusterDescrIDs > 1), then the Cluster Control shall be
present. However, the Cluster Active Control shall always be present (and is therefore not part of the
dOptControls field).

The following table outlines the Processing Unit Descriptor.

Release 4.0 April 2023 112

Universal Serial Bus Device Class Definition for Audio Devices

Table 4-35: Common Part of the Processing Unit Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes:
24+p*2+q*2.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.
4 wDescriptorSubtype 2 Constant PROCESSING_UNIT Descriptor subtype.
6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.
10 wUnitID 2 Number Value uniquely identifying the Unit within
the Audio Function. This value is used in all
Commands to address this Unit.
12 wProcessType 2 Constant Constant identifying the type of processing
this Unit is performing.
14 dOptControls 4 Bitmap D0: Bypass Control.
D1: Cluster Control.
D2: Underflow Control.
D3: Overflow Control.
D31..4: Process-specific.
18 wPCC 2 Number The Pin Channel Count value for the
connection that originates at the Output Pin
of the Processing Unit.
20 wNrInputPins 2 Number Number of Input Pins of this Unit: p
22 waSourceID(1) 2 Number ID of the Unit or Terminal to which the first
Input Pin of this Processing Unit is
connected.
… … 2 … …
20+p*2 waSourceID (p) 2 Number ID of the Unit or Terminal to which the last
Input Pin of this Processing Unit is
connected.
22+p*2 wNrClusterDescrIDs 2 Number Number of elements in the
waClusterDescrID() array: q.
24+p*2 waClusterDescrID(1) 2 Number Unique ID of the first Cluster Descriptor.
… … … … …
22+p*2+q*2 waClusterDescrID(q) 2 Number Unique ID of the last Cluster Descriptor.

4.5.3.10.1 UP/DOWN-MIX PROCESSING UNIT DESCRIPTOR

The wProcessType field of the common Processing Unit Descriptor contains the value UP/DOWNMIX_PROCESS.
(See Appendix A.20, “Processing Unit Process Types”)

The Cluster Control is used to change the behavior of the Processing Unit by selecting different modes of
operation, resulting in a different output Cluster Configuration. If the Up/Down-mix Processing Unit supports more
than one mode of operation, this AudioControl shall be present.

Release 4.0 April 2023 113

Universal Serial Bus Device Class Definition for Audio Devices

The number of supported modes (n) is identical to the number of output Clusters the Unit supports and is
therefore advertised in the wNrClusterDescrIDs field. The index 𝑖 into this array is one-based and directly related
to the number of the mode described by entry waClusterDescrID(𝑖). It is the value 𝑖 that shall be used as a
parameter for the Set Cluster Command to select the mode 𝑖.

The OCN:ICN:IPN triplet used to access an AudioControl within the Up/Down-mix Processing Unit shall be set to
0:0:0.
Table 4-36: Up/Down-mix Processing Unit Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes:
24+p*2+q*2.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.
4 wDescriptorSubtype 2 Constant PROCESSING_UNIT Descriptor subtype.
6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.
10 wUnitID 2 Number Value uniquely identifying the Unit within
the Audio Function. This value is used in all
Commands to address this Unit.
12 wProcessType 2 Constant UP/DOWNMIX_PROCESS process type.
14 dOptControls 4 Bitmap D0: Bypass Control.
D1: Cluster Control.
D2: Underflow Control.
D3: Overflow Control.
D31..4: Reserved.
18 wPCC 2 Number The Pin Channel Count value for the
connection that originates at the Output Pin
of the Processing Unit.
20 wNrInputPins 2 Number Number of Input Pins of this Unit: p
22 waSourceID(1) 2 Number ID of the Unit or Terminal to which the first
Input Pin of this Processing Unit is
connected.
… … … … …
20+p*2 waSourceID (p) 2 Number ID of the Unit or Terminal to which the last
Input Pin of this Processing Unit is
connected.
22+p*2 wNrClusterDescrIDs 2 Number Number of elements in the
waClusterDescrID() array: q.
24+p*2 waClusterDescrID(1) 2 Number Unique ID of the first Cluster Descriptor.
… … … … …
22+p*2+q*2 waClusterDescrID(q) 2 Number Unique ID of the last Cluster Descriptor.

Release 4.0 April 2023 114

Universal Serial Bus Device Class Definition for Audio Devices

4.5.3.10.2 CHANNEL REMAP PROCESSING UNIT DESCRIPTOR

The wProcessType field of the common Processing Unit Descriptor contains the value CHANNEL_REMAP_PROCESS.
(See Appendix A.20, “Processing Unit Process Types”)

The Cluster Control is used to change the behavior of the Processing Unit by selecting different remapping modes,
resulting in a different output Cluster Configuration. If the Channel Remap Processing Unit supports more than one
remapping mode, this AudioControl shall be present.

The number of supported modes (n) is identical to the number of output Clusters the Unit supports and is
therefore advertised in the wNrClusterDescrIDs field. The index 𝑖 into this array is one-based and directly related
to the number of the mode described by entry waClusterDescrID(𝑖). It is the value 𝑖 that shall be used as a
parameter for the Set Cluster Command to select the mode 𝑖.

The OCN:ICN:IPN used to access an AudioControl within the Channel Remap Processing Unit shall be set to 0:0:0.
Table 4-37: Channel Remap Processing Unit Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes:
24+p*2+q*2.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.
4 wDescriptorSubtype 2 Constant PROCESSING_UNIT Descriptor subtype.
6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.
10 wUnitID 2 Number Value uniquely identifying the Unit within
the Audio Function. This value is used in all
Commands to address this Unit.
12 wProcessType 2 Constant CHANNEL_REMAP_PROCESS process type.
14 dOptControls 4 Bitmap D0: Bypass Control.
D1: Cluster Control.
D2: Underflow Control.
D3: Overflow Control.
D31..4: Reserved.
18 wPCC 2 Number The Pin Channel Count value for the
connection that originates at the Output Pin
of the Processing Unit.
20 wNrInputPins 2 Number Number of Input Pins of this Unit: p
22 waSourceID(1) 2 Number ID of the Unit or Terminal to which the first
Input Pin of this Processing Unit is
connected.
… … … … …
20+p*2 waSourceID (p) 2 Number ID of the Unit or Terminal to which the last
Input Pin of this Processing Unit is
connected.
22+p*2 wNrClusterDescrIDs 2 Number Number of elements in the
waClusterDescrID() array: q.
24+p*2 waClusterDescrID(1) 2 Number Unique ID of the first Cluster Descriptor.

Release 4.0 April 2023 115

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

… … … … …
22+p*2+q*2 waClusterDescrID(q) 2 Number Unique ID of the last Cluster Descriptor.

4.5.3.10.3 STEREO EXTENDER PROCESSING UNIT DESCRIPTOR

The wProcessType field of the common Processing Unit Descriptor contains the value
STEREO_EXTENDER_PROCESS. (See Appendix A.20, “Processing Unit Process Types”)

The Stereo Extender Processing Unit has a single Input Pin. Therefore, the wNrInputs field shall contain the value
1.

The input Cluster to the Stereo Extender Processing Unit shall contain at least Front Left and Front Right logical
input channels. The output Cluster is inherited from the Input Pin and therefore, the wNrClusterDescrIDs field shall
be set to one and the waClusterDescrID(1) field shall be set to zero.

The OCN:ICN:IPN used to access an AudioControl within the Stereo Extender Processing Unit shall be set to 0:0:0.
Table 4-38: Stereo Extender Processing Unit Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes: 26.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.
4 wDescriptorSubtype 2 Constant PROCESSING_UNIT Descriptor subtype.
6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.
10 wUnitID 2 Number Value uniquely identifying the Unit within
the Audio Function. This value is used in all
Commands to address this Unit.
12 wProcessType 2 Constant STEREO_EXTENDER_PROCESS process type.
14 dOptControls 4 Bitmap D0: Bypass Control.
D1: Cluster Control not Implemented.
Shall be set to zero.
D2: Underflow Control.
D3: Overflow Control.
D31..4: Reserved.
18 wPCC 2 Number The Pin Channel Count value for the
connection that originates at the Output Pin
of the Processing Unit.
20 wNrInputPins 2 Number Number of Input Pins of this Unit: 1
22 waSourceID(1) 2 Number ID of the Unit or Terminal to which the
Input Pin of this Processing Unit is
connected.
24 wNrClusterDescrIDs 2 Number Number of elements in the
waClusterDescrID() array: 1.
26 waClusterDescrID(1) 2 Number Shall be set to zero.

Release 4.0 April 2023 116

Universal Serial Bus Device Class Definition for Audio Devices

4.5.3.10.4 MULTI-FUNCTION PROCESSING UNIT DESCRIPTOR

The wProcessType field of the Processing Unit Descriptor contains the value MULTI_FUNCTION_PROCESS. (See
Appendix A.20, “Processing Unit Process Types”)

The Multi-Function Processing Unit may have multiple Input Pins as indicated in the wNrInputPins field.

The OCN:ICN:IPN used to access an AudioControl within the Multi-Function Processing Unit shall be set to 0:0:0.

The Multi-Function Processing Unit can redefine its output Cluster Configurations, depending on which algorithms
are currently active. The Read-Only (r) Cluster Control is used to advertise the current output Cluster
Configuration.

The Bypass Control shall be implemented.

The mandatory Read-Only (r) Algo Present Control returns a bitmap indicating what types of algorithms are
performed inside the Multi-Function Processing Unit. Multiple bits may be set simultaneously.

The optional Algo Enable Control provides a means to selectively enable or disable the implemented algorithms.

The following table outlines the Multi-Function Processing Unit Descriptor.

Table 4-39: Multi-Function Processing Unit Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes:
24+p*2+q*2.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.
4 wDescriptorSubtype 2 Constant PROCESSING_UNIT Descriptor subtype.
6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.
10 wUnitID 2 Number Value uniquely identifying the Unit within
the Audio Function. This value is used in all
Commands to address this Unit.
12 wProcessType 2 Constant MULTI_FUNCTION_PROCESS process type.
14 dOptControls 4 Bitmap D0: Cluster Control.
D1: Algo Enable Control.
D2: Underflow Control.
D3: Overflow Control.
D31..4: Reserved.
18 wPCC 2 Number The Pin Channel Count value for the
connection that originates at the Output Pin
of the Processing Unit.
20 wNrInputPins 2 Number Number of Input Pins of this Unit: p
22 waSourceID(1) 2 Number ID of the Unit or Terminal to which the first
Input Pin of this Processing Unit is
connected.
… … … … …

Release 4.0 April 2023 117

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

20+p*2 waSourceID (p) 2 Number ID of the Unit or Terminal to which the last
Input Pin of this Processing Unit is
connected.
22+p*2 wNrClusterDescrIDs 2 Number Number of elements in the
waClusterDescrID() array: q.
24+p*2 waClusterDescrID(1) 2 Number Unique ID of the first Cluster Descriptor.
… … … … …
22+p*2+q*2 waClusterDescrID(q) 2 Number Unique ID of the last Cluster Descriptor.

4.5.3.11 EXTENSION UNIT DESCRIPTOR

The Extension Unit is uniquely identified by the value in the wUnitID field of the Extension Unit Descriptor. This
value shall be passed in the wEntityID field of each Command that is directed to the Extension Unit.

The Extension Unit Descriptor provides just enough information about the Extension Unit so that a generic Audio
Class driver can be aware of vendor-specific components within the Audio Function. The guidExtensionCode field
shall contain a vendor-specific code in the form of a GUID that further identifies the Extension Unit. Note that the
GUID is used to uniquely identify the functionality and behavior of the Extension Unit. Therefore, the same GUID
shall be used in all implementations that expose the Extension Unit with this functionality and behavior.

For more information about the generation and use of a GUID, see [IETF RFC 4122 GUID].

The wNrInputPins field contains the number of Input Pins (p) of the Extension Unit. The connectivity of the Input
Pins is described via the waSourceID() array that contains p elements. The index 𝑖 into the array is one-based and
directly related to the Input Pin numbers. waSourceID(𝑖) contains the ID of the Unit or Terminal to which Input Pin
𝑖 is connected. The Cluster Descriptors that describe the logical channels that enter the Extension Unit are not
repeated here. It is up to the Host software to trace the connections ‘upstream’ to locate the Cluster Descriptors
pertaining to the Clusters.

The waPCC() array contains the PCC values 𝑁𝑝 the Extension Unit supports on each of its Input Pins, excluding the
Primary channel. However, the actual number of logical channels used on each Input Pin at any given time is solely
determined by the number of channels in the incoming Clusters on each Input Pin.

The Extension Unit can freely redefine its output Cluster Configurations, based on internal settings of the Extension
Unit’s vendor-defined AudioControls.

The Bypass Control shall be implemented.

The waClusterDescrID() array contains the IDs of the Cluster Descriptors that characterize the Clusters that may be
exposed on the Output Pin of the Extension Unit. A value of zero in an array element indicates that the output
Cluster is inherited from the current Cluster on Input Pin one (the dominant Input Pin). The wNrClusterDescrIDs
field contains the number of elements in that array and shall always be greater than zero. For a detailed
description of the Cluster Descriptor, see Section 4.4, “Cluster Descriptor”. If the Extension Unit exposes more than
one Cluster Configuration on its Output Pin (wNrClusterDescrIDs > 1), then the Cluster Control shall be present.
However, the Cluster Active Control shall always be present (and is therefore not part of the dOptControls field).

The OCN:ICN:IPN used to access a class-defined AudioControl within the Extension Unit shall be set to 0:0:0.
The following table outlines the Extension Unit Descriptor.

Release 4.0 April 2023 118

Universal Serial Bus Device Class Definition for Audio Devices

Table 4-40: Extension Unit Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes:
38+p*2+q*2.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.
4 wDescriptorSubtype 2 Constant EXTENSION_UNIT Descriptor subtype.
6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.
10 wUnitID 2 Number Value uniquely identifying the Unit within
the Audio Function. This value is used in all
Commands to address this Unit.
12 guidExtensionCode 16 GUID Vendor-specific code identifying the
Extension Unit.
28 dOptControls 4 Bitmap D0: Cluster Control.
D1: Underflow Control.
D2: Overflow Control.
D31..3: Reserved.
32 wPCC 2 Number The Pin Channel Count value for the
connection that originates at the Output Pin
of the Processing Unit.
34 wNrInputPins 2 Number Number of Input Pins of this Unit: p
36 waSourceID(1) 2 Number ID of the Unit or Terminal to which the first
Input Pin of this Processing Unit is
connected.
… … … … …
34+p*2 waSourceID (p) 2 Number ID of the Unit or Terminal to which the last
Input Pin of this Processing Unit is
connected.
36+p*2 wNrClusterDescrIDs 2 Number Number of elements in the
waClusterDescrID() array: q.
38+p*2 waClusterDescrID(1) 2 Number Unique ID of the first Cluster Descriptor.
… … … … …
36+p*2+q*2 waClusterDescrID(q) 2 Number Unique ID of the last Cluster Descriptor.

4.5.3.12 CLOCK SOURCE DESCRIPTOR

The Clock Source Entity is uniquely identified by the value in the wClockID field of the Clock Source Entity
Descriptor. This value shall be passed in the wEntityID field of each Command that is directed to the Clock Source
Entity.

The Clock Frequency Control shall always be present and may be implemented as either Read (r) or Read-Write
(rw). The supported clock frequencies can be derived from the Range Attribute of the Clock Frequency Control
(fixed rate vs. variable rate). Note that even a Clock Source of Type External may be implemented as Read-Write
(rw) if the Audio Function can influence that external clock through means outside of USB. The actual sampling

Release 4.0 April 2023 119

Universal Serial Bus Device Class Definition for Audio Devices

frequency of the Clock Source can be manipulated through the Clock Frequency Command. In addition, the Clock
Source can be queried for the validity of its current sampling clock signal through a Get Clock Valid Command.

The wAttributes field contains a Clock Type bit field (D0) that indicates whether the Clock Source represents an
external clock (D0 = 0b0) or an internal clock (D0=0b1).

The wClockDomainID field contains the ID of the Clock Domain from which this Clock Source Entity derives its
reference clock.

A value of zero indicates that this Clock Source Entity is independent and free running. Multiple Clock Source
Entities may have a value of zero in this field, which means that they are all independent of one another. (It does
not mean that these Clock Source Entities belong to the same Clock Domain with ID zero.)

A value of 0xFFFF indicates that the Clock Source Entity is synchronized to SOF.

Any other value indicates the ID of the Clock Domain to which this Clock Source Entity is synchronized.

The wReferenceTerminal field contains a reference to a Terminal from which the Clock Source is derived. This is
useful for instance when a Clock Source’s clock signal is derived from the input signal on an S/PDIF connector,
which is represented by an Input Terminal. If the Clock Source is free running or derived from USB SOF (not derived
from a Terminal), this field shall be set to zero.

The OCN:ICN:IPN used to access an AudioControl within the Clock Source Entity shall be set to 0:0:0.

The following table outlines the Clock Source Descriptor.

Table 4-41: Clock Source Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes: 22.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.
4 wDescriptorSubtype 2 Constant CLOCK_SOURCE Descriptor subtype.
6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.
10 wClockID 2 Number Value uniquely identifying the Clock Source
Entity within the Audio Function. This value
is used in all Commands to address this
Entity.
12 wAttributes 2 Bitmap D0: Clock Type:
0: External Clock.
1: Internal Clock.
D15..1: Reserved.
14 wClockDomainID 2 Number 0x0000: Independent.
0x0001..0xFFFE: ID of the reference Clock
Domain.
0xFFFF: Synchronized to SOF.
16 dOptControls 4 Bitmap D31..0: Reserved.
20 wReferenceTerminal 2 Number Terminal ID of the Terminal from which this
Clock Source is derived.

Release 4.0 April 2023 120

Universal Serial Bus Device Class Definition for Audio Devices

4.5.3.13 CLOCK SELECTOR DESCRIPTOR

The Clock Selector Entity is uniquely identified by the value in the wClockID field of the Clock Selector Entity
Descriptor. This value shall be passed in the wEntityID field of each Command that is directed to the Clock Selector
Entity.

The wNrInputPins field contains the number of Clock Input Pins (𝑝) of the Clock Selector Entity. The connectivity of
the Input Pins is described via the waCSourceID() array that contains 𝑝 elements. The index 𝑖 into the array is one-
based and directly related to the Clock Input Pin numbers. waCSourceID(𝑖) contains the ID of the Clock Entity to
which Clock Input Pin 𝑖 is connected.

The OCN:ICN:IPN used to access an AudioControl within the Clock Selector Entity shall be set to 0:0:0.

The following table outlines the Clock Selector Descriptor.

Table 4-42: Clock Selector Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes: 18+p*2.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.
4 wDescriptorSubtype 2 Constant CLOCK_SELECTOR Descriptor subtype.
6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.
10 wClockID 2 Number Value uniquely identifying the Clock
Selector Entity within the Audio Function.
This value is used in all Commands to
address this Entity.
12 dOptControls 4 Bitmap D31..0: Reserved.
16 wNrInputPins 2 Number Number of Input Pins of this Unit: p.
18 waCSourceID(1) 2 Number ID of the Clock Entity to which the first
Clock Input Pin of this Clock Selector Entity
is connected.
… … … … …
16+p*2 waCSourceID (p) 2 Number ID of the Clock Entity to which the last Clock
Input Pin of this Clock Selector Entity is
connected.

4.5.3.14 CONNECTOR ENTITY DESCRIPTOR

The Connector Entity Descriptor returns information about the Attributes of a physical Connector, located on the
enclosure that houses the Audio Function. This Connector is typically user accessible.

The wConID field contains a unique identifier for the Connector. The primary use for this is to indicate that the
same Connector is associated with multiple Terminals. For example, one headset Connector may incorporate the
signals for the stereo headphone of the headset and the signal for the mono microphone. This Connector would
therefore be part of the Output Terminal that represents the stereo headphone but also be part of the Input
Terminal that represents the microphone. This Connector would then be listed in the VARIANT_ENTITIES part of
both the Input Terminal and Output Terminal Descriptor, using the same wConID value in both Descriptors to
indicate the binding.

Release 4.0 April 2023 121

Universal Serial Bus Device Class Definition for Audio Devices

The wConType field contains a value that identifies the physical appearance of the Connector. The constant
definitions for the wConType field can be found in Appendix A.24, “Connector Types”.

The wConAttributes field contains a bitmap that identifies the gender of the Connector (D1..0).

The dConColor field contains either 0x00 in the upper byte and the RGB-coded color of the Connector in the lower
3 bytes or 0x01 in the upper byte and 0x000000 in the lower 3 bytes to indicate color unspecified.

The OCN:ICN:IPN used to access an AudioControl within the Connector Entity shall be set to 0:0:0.

The following table outlines the Connector Entity Descriptor.

Table 4-43: Connector Entity Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes: 24.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.
4 wDescriptorSubtype 2 Constant CONNECTOR Descriptor subtype.
6 wDescriptorID 2 Number Unique ID of this Connector Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.
10 wConID 2 Number Unique ID for the Connector Entity. Can be
used to indicate that the same Connector is
associated with multiple Terminals.
12 dOptControls 4 Bitmap D0: Insertion Control.
D31..1: Reserved.
16 wConType 2 Constant The Connector Type of the Connector.
18 wConAttributes 2 Bitmap D1..0: Gender:
00: Gender Neutral.
01: Plug.
10: Receptacle.
11: Reserved.
D7..2: Reserved.
20 dConColor 4 Number The Connector color of the Connector.

4.5.3.15 POWER DOMAIN ENTITY DESCRIPTOR

Power Domain Entity Descriptors provide information to the Host regarding the existence of one or more Power
Domains within the Audio Function and list the Entities, through their respective Entity ID, that are explicit
members of a particular Power Domain.

There is a Power Domain Entity Descriptor for each Power Domain in the Audio Function. Therefore, the number of
Power Domain Entity Descriptors is an indicator for the number of separately managed Power Domains in the
Audio Function.

The Power Domain and its associated Power Domain Entity are uniquely identified by the value in the
wPowerDomainID field of the Power Domain Entity Descriptor. This value shall be passed in the wEntityID field of
each Command that is directed to the Power Domain Entity.

Release 4.0 April 2023 122

Universal Serial Bus Device Class Definition for Audio Devices

The wEntryTime1..4 fields contain the approximate entry time from Power State PS0 to Power States PS1, PS2,
PS3, and PS4, respectively.

The wExitTime1..4 fields contain the approximate exit time from Power State PS1, PS2, PS3, and PS4 to Power
State PS0, respectively.

The waEntityID() array contains the Entity IDs of the explicit Member Entities. The wNrEntityIDs field contains the
number of elements in that array.

The OCN:ICN:IPN used to access an AudioControl within the Power Domain Entity shall be set to 0:0:0.

The wStrDescriptorID field provides the ID of a String Descriptor to further describe the Power Domain Entity.
Table 4-44: Power Domain Entity Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes: 34+p*2.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.
4 wDescriptorSubtype 2 Constant POWER_DOMAIN Descriptor subtype.
6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.
10 wPowerDomainID 2 Number Value uniquely identifying the Power
Domain Entity within the Audio Function.
This value is used in all Commands to
address this Power Domain Entity.
12 dOptControls 4 Bitmap D31..0: Reserved.
16 wEntryTime1 2 Number Time to enter from PS0 to PS1. Expressed in
50 s increments.
18 wExitTime1 2 Number Time to exit from PS1 to PS0. Expressed in
50 s increments.
20 wEntryTime2 2 Number Time to enter from PS0 to PS2. Expressed in
50 s increments.
22 wExitTime2 2 Number Time to exit from PS2 to PS0. Expressed in
50 s increments.
24 wEntryTime3 2 Number Time to enter from PS0 to PS3. Expressed in
50 s increments.
26 wExitTime3 2 Number Time to exit from PS3 to PS0. Expressed in
50 s increments.
28 wEntryTime4 2 Number Time to enter from PS0 to PS4. Expressed in
50 s increments.
30 wExitTime4 2 Number Time to exit from PS4 to PS0. Expressed in
50 s increments.
32 wNrEntityIDs 2 Number Number of elements in the baEntityID()
array: p.
34 waEntityID(1) 2 Number ID of the first Entity that belongs to this
Power Domain.
… … … …

Release 4.0 April 2023 123

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

32+p*2 waEntityID(p) 2 Number ID of the last Entity that belongs to this
Power Domain.

4.5.3.16 ENTITYGROUP DESCRIPTOR

The waEntityID() array contains the Entity IDs of the Members of the EntityGroup. The wNrEntityIDs field contains
the number of elements in that array.
Table 4-45: EntityGroup Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes: 12+p*2.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.
4 wDescriptorSubtype 2 Constant ENTITY_GROUP Descriptor subtype.
6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.
10 wNrEntityIDs 2 Number Number of elements in the baEntityID()
array: p.
12 waEntityID(1) 2 Number ID of the first Entity that belongs to this
EntityGroup.
… … … …
10+p*2 waEntityID(p) 2 Number ID of the last Entity that belongs to this
EntityGroup.

4.5.3.17 COMMITGROUP DESCRIPTOR

The waAudioControl() array contains all the information necessary to fully identify each AudioControl Member of
the CommitGroup. The wNrAudioControls field contains the number of AudioControls in the array.
Table 4-46: CommitGroup Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes: 12+p*10.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.
4 wDescriptorSubtype 2 Constant COMMIT_GROUP Descriptor subtype.
6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.
10 wNrAudioControls 2 Number Number of AudioControls in the
Commitgroup: p.
12 waAudioControl(1).CS 2 Number Control Selector CS of the first AudioControl
in the CommitGroup.
14 waAudioControl(1).Entity 2 Number ID of the Entity to which the first
AudioControl in the CommitGroup belongs.
16 waAudioControl(1).OCN 2 Number OCN of the first AudioControl in the
CommitGroup.

Release 4.0 April 2023 124

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

18 waAudioControl(1).ICN 2 Number ICN of the first AudioControl in the
CommitGroup.
20 waAudioControl(1).IPN 2 Number IPN of the first AudioControl in the
CommitGroup.
… … … …
12+(p-1)*10 waAudioControl(p).CS 2 Number Control Selector CS of the last AudioControl
in the CommitGroup.
14+(p-1)*10 waAudioControl(p).Entity 2 Number ID of the Entity to which the last
AudioControl in the CommitGroup belongs.
16+(p-1)*10 waAudioControl(p).OCN 2 Number OCN of the last AudioControl in the
CommitGroup.
18+(p-1)*10 waAudioControl(p).ICN 2 Number ICN of the last AudioControl in the
CommitGroup.
20+(p-1)*10 waAudioControl(p).IPN 2 Number IPN of the last AudioControl in the
CommitGroup.

4.6 AUDIOCONTROL ENDPOINT DESCRIPTORS

The following sections describe all possible Endpoint-related Descriptors for the AudioControl Interface.

4.6.1 AUDIOCONTROL CONTROL ENDPOINT DESCRIPTORS

4.6.1.1 STANDARD AUDIOCONTROL CONTROL ENDPOINT DESCRIPTOR

The AudioControl Interface uses the default Endpoint 0. Therefore, there is no dedicated standard AudioControl
Control Endpoint Descriptor.

4.6.1.2 CLASS-SPECIFIC AUDIOCONTROL CONTROL ENDPOINT DESCRIPTOR

There is no dedicated class-specific AudioControl Control Endpoint Descriptor.

4.6.2 AUDIOCONTROL INTERRUPT ENDPOINT DESCRIPTORS

4.6.2.1 STANDARD AUDIOCONTROL INTERRUPT ENDPOINT DESCRIPTOR

The interrupt Endpoint Descriptor is identical to the standard Endpoint Descriptor defined in the USB Core
Specifications. Its fields are set to reflect the interrupt type of the Endpoint. This Endpoint is optional.

The following table outlines the standard AudioControl Interrupt Endpoint Descriptor.
Table 4-47: Standard AudioControl Interrupt Endpoint Descriptor

Offset Field Size Value Description

0 bLength 1 Number Size of this Descriptor, in bytes: 7
1 bDescriptorType 1 Constant ENDPOINT Descriptor type

Release 4.0 April 2023 125

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

2 bEndpointAddress 1 Endpoint The address of the Endpoint on the USB
Device described by this Descriptor. The
address is encoded as follows:
D7: Direction. 1 = IN Endpoint.
D6..4: Reserved.
D3..0: The Endpoint number,
determined by the designer.
3 bmAttributes 1 Bitmap D1..0: Transfer Type
11 = Interrupt.
All other bits are reserved.
4 wMaxPacketSize 2 Number Maximum packet size this Endpoint is
capable of sending or receiving when this
configuration is selected. Used here to pass
6-byte Interrupt Data Message.
6 bInterval 1 Number Interval for polling the Interrupt Endpoint.

4.6.2.2 CLASS-SPECIFIC AUDIOCONTROL INTERRUPT ENDPOINT DESCRIPTOR

There is no class-specific AudioControl interrupt Endpoint Descriptor.

4.7 AUDIOSTREAMING INTERFACE DESCRIPTORS

The AudioStreaming (AS) Interface Descriptors contain all relevant information to characterize the AudioStreaming
Interface in full.

4.7.1 STANDARD AUDIOSTREAMING INTERFACE DESCRIPTOR

The standard AudioStreaming Interface Descriptor is identical to the standard interface Descriptor defined in the
USB Core Specifications, except that some fields now have dedicated values.
Table 4-48: Standard AudioStreaming Interface Descriptor

Offset Field Size Value Description

0 bLength 1 Number Size of this Descriptor, in bytes: 9.
1 bDescriptorType 1 Constant INTERFACE Descriptor type.
2 bInterfaceNumber 1 Number Number of the interface. A zero-based
value identifying the index in the array of
concurrent interfaces supported by this
configuration.
3 bAlternateSetting 1 Number Value used to select an Alternate Setting for
the interface identified in the prior field.
4 bNumEndpoints 1 Number Number of Endpoints used by this interface
(excluding Endpoint 0). Shall be either 0 (no
data Endpoint), 1 (data Endpoint) or 2 (data
and explicit feedback Endpoint).
5 bInterfaceClass 1 AUDIO Audio Interface Class code (assigned by the
USB). See Appendix A.5, “Audio Interface
Class Code.”

Release 4.0 April 2023 126

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

6 bInterfaceSubClass 1 AUDIO Audio Interface Subclass code. Assigned by
STREAMING this specification. See Appendix A.6, “Audio
Interface Subclass Codes.”
7 bInterfaceProtocol 1 IP_VERSION_0 Interface Protocol code. Indicates the
4_00 current version of the specification. See
Appendix A.7, “Audio Interface Protocol
Codes.”
8 iInterface 1 Index Index of a String Descriptor that describes
this interface.

4.7.2 CLASS-SPECIFIC AUDIOSTREAMING INTERFACE DESCRIPTOR

The class-specific AudioStreaming Interface Descriptor is a traditional Descriptor that contains the list of Extended
Descriptor IDs in the waDescriptorID() array, used to fully describe the AudioStreaming Interface.

The order in which the Descriptor IDs are reported is not important because every Descriptor can be identified
through its bDescriptorType and bDescriptorSubtype field.

The following table defines the class-specific AudioStreaming Interface Descriptor.

Table 4-49: Class-specific AudioStreaming Interface Descriptor

Offset Field Size Value Description

0 bLength 1 Number Size of this Descriptor, in bytes: 4+n*2.
1 bDescriptorType 1 Constant CS_INTERFACE Descriptor type.
2 bDescriptorSubtype 1 Constant AS_GENERIC Descriptor subtype.
3 bNrDescriptorIDs 1 Number Number of Descriptor IDs in the list: n.
4 waDescriptorID(1) 2 Bitmap First Descriptor ID.
… … … … …
4+(n-1)*2 waDescriptorID(n) 2 Bitmap Last Descriptor ID.

If there is a need to list more than 125 Descriptor IDs, then more Descriptors of subtype AS_GENERIC may be
included.

4.7.3 CLASS-SPECIFIC AUDIOSTREAMING SELF DESCRIPTOR

The fields in this Descriptor are described in detail in Section 7.4, “Class-specific AudioStreaming Self Descriptor.”
Table 4-50: Class-specific AudioStreaming Self Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor in bytes: 28.

2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.

4 wDescriptorSubtype 2 Constant AS_SELF Descriptor subtype.
6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.

Release 4.0 April 2023 127

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

10 dOptControls 4 Bitmap D0: Active Alternate Setting Control.
D1: Valid Alternate Settings Control.
D31..2: Reserved.
14 wStartDelayUnits 2 Number Indicates the units used for the wStartDelay
field:
0: Undefined.
1: Milliseconds.
2: Decoded PCM samples.
All other values are Reserved.
16 wStartDelay 2 Number Indicates the time it takes this Interface to
reliably produce or consume streamed data.
Units used depend on the value of the
bStartDelayUnits field.
18 wFormat 2 Number The Audio Data Format used to
communicate with this Alternate Setting of
the AudioStreaming Interface. See Section
7, “Audio Data Formats” for further details.
20 wSubslotSize 2 Number The number of bytes occupied by one audio
subslot.
22 wBitResolution 2 Number The number of effectively used bits from
the available bits in an audio subslot.
24 wAuxProtocols 2 Bitmap Bitmap, indicating which Auxiliary Protocols
are required.
26 wControlSize 2 Number Size of the Control Channel Words, in bytes.

4.7.4 CLASS-SPECIFIC AUDIOSTREAMING VALID FREQUENCY RANGE DESCRIPTOR

The AudioStreaming Valid Frequency Range Descriptor provides information to the Host about what sampling
frequency ranges are supported by this Alternate Setting of the AudioStreaming Interface. An Audio Function shall
provide this Descriptor to indicate that this Alternate Setting of the interface is only valid if the selected sampling
frequency is in the range [dMin..dMax]. The values of dMin and dMax are expressed in Hz. If the Alternate Setting
of the interface is valid for any available clock frequency supported by the Audio Function, the Descriptor may be
omitted. Multiple instances of this Descriptor are allowed in the same Alternate Setting of the AudioStreaming
Interface to describe disjoint frequency ranges.
Table 4-51: Class-specific AudioStreaming Valid Frequency Range Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor in bytes: 18.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor type.
4 wDescriptorSubtype 2 Constant AS_VALID_FREQ_RANGE Descriptor
subtype.
6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.

Release 4.0 April 2023 128

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

10 dMin 4 Number The minimum sampling frequency at which
this Alternate Setting of the
AudioStreaming Interface is valid.
14 dMax 4 Number The maximum sampling frequency at which
this Alternate Setting of the
AudioStreaming Interface is valid.

4.8 AUDIOSTREAMING ENDPOINT DESCRIPTORS

The following sections describe all possible Endpoint-related Descriptors for the AudioStreaming Interface.

4.8.1 AUDIOSTREAMING ISOCHRONOUS AUDIO DATA ENDPOINT DESCRIPTORS

The standard and class-specific audio data Endpoint Descriptors provide pertinent information on how audio data
streams are communicated to the Audio Function. In addition, specific Endpoint capabilities and properties are
reported.

4.8.1.1 STANDARD AUDIOSTREAMING ISOCHRONOUS AUDIO DATA ENDPOINT

DESCRIPTOR
The standard AudioStreaming isochronous audio data Endpoint Descriptor is identical to the standard Endpoint
Descriptor defined in the USB Core Specifications. D7 of the bEndpointAddress field indicates whether the
Endpoint is an audio source (D7 = 0b1) or an audio sink (D7 = 0b0). The bmAttributes Field bits are set to reflect
the isochronous type of the Endpoint. The synchronization type is indicated by D3..2 and shall be set to
Asynchronous, Adaptive or Synchronous. For further details, refer to the USB Core Specifications.
Table 4-52: Standard AudioStreaming Isochronous Audio Data Endpoint Descriptor

Offset Field Size Value Description

0 bLength 1 Number Size of this Descriptor, in bytes: 7
1 bDescriptorType 1 Constant ENDPOINT Descriptor type
2 bEndpointAddress 1 Endpoint The address of the Endpoint on the USB
Device described by this Descriptor. The
address is encoded as follows:
D3..0: The Endpoint number,
determined by the designer.
D6..4: Reserved.
D7: Direction:
0 = OUT Endpoint.
1 = IN Endpoint.

Release 4.0 April 2023 129

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

3 bmAttributes 1 Bitmap D1..0: Transfer type:
01 = Isochronous.
D3..2: Synchronization Type:
01 = Asynchronous.
10 = Adaptive.
11 = Synchronous.
D5..4: Usage Type:
00 = Data Endpoint
or
10 = Implicit feedback Data Endpoint.
All other bits are reserved.
4 wMaxPacketSize 2 Number (Speed-dependent.) Indicates the maximum
packet size and potentially any bursting on
this Endpoint.
This is determined by the audio bandwidth
constraints of the Endpoint.
6 bInterval 1 Number Interval for polling Endpoint for data
transfers.
Note: For SuperSpeed and SuperSpeedPlus Endpoints, the SuperSpeed Endpoint Companion and
SuperSpeedPlus Endpoint Companion Descriptors would follow the standard Endpoint Descriptor.
See the USB 3.1 specification for details.

4.8.1.2 CLASS-SPECIFIC AUDIOSTREAMING ISOCHRONOUS AUDIO DATA ENDPOINT

DESCRIPTOR
There is no class-specific AudioStreaming isochronous Audio Data Endpoint Descriptor.

4.8.2 AUDIOSTREAMING ISOCHRONOUS FEEDBACK ENDPOINT DESCRIPTOR

This Descriptor is present only when one or more isochronous audio data Endpoints of the adaptive source type or
the asynchronous sink type are implemented.

4.8.2.1 STANDARD AUDIOSTREAMING ISOCHRONOUS FEEDBACK ENDPOINT DESCRIPTOR

The isochronous feedback Endpoint Descriptor is identical to the standard Endpoint Descriptor defined in the USB
Core Specifications. The bmAttributes field bits are set to reflect the isochronous type and synchronization type of
the Endpoint.
Table 4-53: Standard AudioStreaming Isochronous Feedback Endpoint Descriptor

Offset Field Size Value Description

0 bLength 1 Number Size of this Descriptor, in bytes: 7
1 bDescriptorType 1 Constant ENDPOINT Descriptor type.

Release 4.0 April 2023 130

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

2 bEndpointAddress 1 Endpoint The address of the Endpoint on the USB
Device described by this Descriptor. The
address is encoded as follows:
D3..0: The Endpoint number,
determined by the designer.
D6..4: Reserved.
D7: Direction:
0 = OUT Endpoint.
1 = IN Endpoint.
3 bmAttributes 1 Bitmap D1..0: Transfer type:
01 = Isochronous.
D3..2: Synchronization Type:
00 = No Synchronization.
D5..4: Usage Type:
01 = Feedback Endpoint.
All other bits are reserved.
4 wMaxPacketSize 2 Number Maximum packet size this Endpoint is
capable of sending or receiving when this
configuration is selected.
6 bInterval 1 Number Interval for polling Endpoint for data
transfers.

Note: For SuperSpeed and SuperSpeedPlus Endpoints, the SuperSpeed Endpoint Companion and
SuperSpeedPlus Endpoint Companion Descriptors would follow the standard Endpoint Descriptor.
See the USB 3.1 specification for details.

4.8.2.2 CLASS-SPECIFIC AUDIOSTREAMING ISOCHRONOUS FEEDBACK ENDPOINT

DESCRIPTOR
There is no class-specific AudioStreaming isochronous feedback Endpoint Descriptor.

4.9 CLASS-SPECIFIC STRING DESCRIPTORS

This specification defines a new type of class-specific String Descriptor. Class-specific String Descriptors are
retrieved through a class-specific Get String Command. See Section 5.3.5.1, “Class-specific String Descriptor ” for
details. All class-specific strings in the Audio Function shall use this new methodology. Strings that are part of the
standard Descriptor set shall use the standard string methodology.

Class-specific strings may be dynamic in nature, i.e., change during normal operation and inform the Host of such a
change by generating an interrupt with source type set to STRING.

The wLength field contains the length of the class-specific String Descriptor. Class-specific strings may be up to
65,525 bytes in length.

The wDescriptorType field shall be set to EXT_STRING.

The wDescriptorSubtype field indicates the Descriptor subtype for the String Descriptor. (Currently, only the value
STRING is defined.)

Release 4.0 April 2023 131

Universal Serial Bus Device Class Definition for Audio Devices

The wDescriptorID field contains a unique identifier for the class-specific String Descriptor in the range
[256..65,535].

The iLangID field contains a zero-based index into the LANGID code array as returned by the Device. A Device can
at most support 126 different languages since the LANGID code array is restricted to 254 bytes and each LANGID
code takes up 2 bytes. The range of the iLangID is therefore from 0 to 125 maximum.

The String field contains the actual Unicode encoded string as outlined in the USB Core Specifications.
Table 4-54: Class-specific String Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of the String Descriptor in bytes: 12+N.
2 wDescriptorType 2 Number EXT_STRING. Type of this Descriptor.
4 wDescriptorSubtype 2 Constant STRING Descriptor Subtype.
6 wDescriptorID 2 Number Unique ID for this class-specific String
Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.
10 iLangID 1 Number Zero-based index into the LANGID code
array as returned by the Device.
11 bReserved 1 0x00 Shall be set to zero.
12 String N Number Unicode UTF16LE encoded string. Follows
the definitions outlined in the USB Core
Specifications.

Release 4.0 April 2023 132

Universal Serial Bus Device Class Definition for Audio Devices

5 COMMANDS & REQUESTS

The following sections provides details about the Commands and Requests that are used to communicate and
control the Audio Function.

5.1 STANDARD REQUESTS

The Audio Device Class supports the standard Requests described in Section 9, “USB Device Framework,” of the
USB core specification. The Audio Device Class places no specific requirements on the values for the standard
Requests.

5.2 CLASS-SPECIFIC AUDIO FUNCTION MANAGEMENT COMMANDS

Class-specific Audio Function Management Commands are used to manipulate global Audio Function Capabilities.
These Commands are special in that they manipulate a global Audio Function Capability rather than a specific
AudioControl within the Audio Function. These Capabilities are addressed through the Audio Function’s single
AudioControl Interface. They do not use the normal Push/Pull Command paradigm (see further). Rather the
underlying Set/Get Request paradigm is used to interact with these Capabilities as follows.
Table 5-1: Set/Get Function Request Layout

bmRequest bRequest wValue wIndex wLength Data

Type
00100001B COMMIT Zero Zero Length of Parameter Block
SWITCH_FUNCTION and Parameter Block
Interface
10100001B SWITCH_FUNCTION

Bit D7 of the bmRequestType field shall be set to 0b0 for the Set Request and to 0b1 for the Get Request. It is a
class-specific Request (D6..5 = 0b01), directed at an AudioControl Interface of the Audio Function
(D4..0 = 0b00001).

The bRequest field shall contain the COMMIT or SWITCH_FUNCTION constant for the Set Request and the
SWITCH_FUNCTION constant for the Get request. If the field contains a value other than the allowed values for this
field, the Request shall return a Request Error.

The wValue field shall be set to zero. If the field contains a value other than zero, the Request shall return a
Request Error.
The value in the low byte of the wIndex field shall be appropriate to the recipient. Only appropriate AudioControl
Interface numbers may be used. If the Command specifies an unknown AudioControl Interface number, the
Request shall return a Request Error.

The high byte of the wIndex field shall be set to zero. If it contains a value other than zero, the Request shall return
a Request Error.

At this time, there are two global Capabilities defined as detailed below.

5.2.1 COMMIT
The Commit Capability is used to simultaneously update the CUR Attributes of all or a select group of
AudioControls within the Audio Function with their corresponding preloaded NEXT Attribute values in a
synchronized fashion. Only the Set Request is supported for this Capability. The Parameter Block shall contain

Release 4.0 April 2023 133

Universal Serial Bus Device Class Definition for Audio Devices

either zero or the wDescriptorID field value of a CommitGroup Descriptor. When zero is specified, all Armed
AudioControl CUR attributes within the entire Audio Function are updated. When a CommitGroup Descriptor ID is
specified, only the CUR Attributes of the Armed AudioControls that are Members of the indicated CommitGroup
are updated. If the Parameter Block contains a value other than zero or a valid CommitGroup Descriptor ID, the
Request shall return a Request Error.

If for some reason, one or more NEXT Attribute values have become invalid between the time the NEXT
Attribute(s) were preloaded (and checked for their validity at that time), then the Commit Command shall return a
Request Error and no updates to the CUR Attributes to any of the Armed AudioControls shall take place.

Note: This specification does not provide explicit means for the Host to find out which of the Armed
AudioControls caused the Commit Command to return a Request Error. The Host should implement
an appropriate strategy to recover gracefully from this situation.

The wLength field of the Set Request shall be set to two. If it contains a value other than two, the Request shall
return a Request Error.

The Commiting by CommitGroup functionality is optional. Whether the Audio Function supports this option is
indicated in the dOptControls field of the AC Self Descriptor (see Section 4.5.3.1, “AC Self Descriptor”).

Support for the Commit Capability is Conditionally Required. If none of the Audio Function’s AudioControls support
the NEXT Attribute, then the Commit Capability shall not be supported. However, if one or more AudioControls do
support the NEXT Attribute, then the Commit Capability shall be supported.

5.2.2 SWITCH FUNCTION

The Switch Function Capability is used to switch the entire Audio Function from its Base Revision Level operational
mode into a Higher Revision Level operational mode. The Capability is a Read-Only/Write-Once (ROW1).

When read, the current Audio Function Revision Level, as indicated by the bFunctionProtocol in the Interface
Association Descriptor, is returned as a single byte in the Parameter Block of the Get Request.

When written, the Parameter Block shall contain the single byte value as advertised in the bFunctionProtocol field
of the Interface Association Descriptor of one of the HRL Interface Descriptor Sets, which is described in the BOS
HRL Function Capability Descriptor in Section 8.2, “Discovery of Higher Version Level Support for an Audio
Function8.2.” The Device shall not complete the Status Stage of the Set Request until the Audio Function has fully
transitioned from BRL operation to HRL operation. Subsequent write operations to this AudioControl shall result in
a Request Error. The only way for the Host to choose a different HRL operation mode (if more than one is
supported by the Audio Function) is to reset the entire Device and restart operation in BRL mode.

The wLength field of the Set or Get Request shall be set to one. If it contains a value other than one, the Request
shall return a Request Error.

Support for the Switch Function Control is Conditionally Required. If the Audio Function does not support an HRL
operation mode, then this Request shall not be supported. However, If the Audio Function does support at least
one HRL operation mode, then the Switch Function Control shall be supported.

5.3 CLASS-SPECIFIC AUDIOCONTROL COMMANDS

Class-specific AudioControl Commands are used to manipulate audio-related AudioControls. These AudioControls
fall into two main groups: those that manipulate the Audio Function’s AudioControls, such as volume, tone,
selector position, etc. and those that describe or influence data transfer over an AudioStreaming interface, such as
the valid Alternate Settings Control.

Release 4.0 April 2023 134

Universal Serial Bus Device Class Definition for Audio Devices

• Control of an Audio Function is performed through the manipulation of the Attributes of individual
AudioControls that are embedded in the Entities of the Audio Function. (The AudioControl Interface itself is
considered to be an Entity in itself and uses Entity ID zero for access.)
The class-specific AudioControl Interface Descriptor contains a collection of Entity Descriptors, each indicating
which AudioControls are present in the Entity. Commands are always directed to the single AudioControl
Interface of the Audio Function. The Command contains enough information (Entity ID, Control Selector,
Control Attribute, etc.) for the Audio Function to decide to where a specific Command is to be routed.
• Control of the class-specific behavior of an AudioStreaming Interface is performed through manipulation of
Interface AudioControls. Commands are directed to the AudioStreaming Interface where the AudioControl
resides.

The Audio Device Class supports two additional class-specific Commands that do not manipulate AudioControls
inside the Audio Function:

• String Commands provide a class-specific method to retrieve String Descriptors from the Audio Function. The
String Command is introduced to overcome the USB core specification limitation that only provides for 255
Device-wide String Descriptors.
• Descriptor Commands provide a class-specific method to retrieve Descriptors from the Audio Function outside
the standard Descriptor retrieval during enumeration. The Descriptor Command is introduced to overcome
the USB core specification limitation that only provides for Descriptors that are a maximum of 255 bytes long.
It also enables dynamically changing Descriptors (after enumeration).

In general, all AudioControls and their associated Commands are optional, unless explicitly stated otherwise in the
AudioControl description (see Appendix A.25, “AudioControl Capabilities Overview” for an overview).

5.3.1 CLASS-SPECIFIC AUDIOCONTROL COMMAND STRUCTURE AND LAYOUT

To overcome the limitations imposed by the standard USB Requests, this specification introduces a new class-
specific approach to communicate with the internals of the Audio Function. Standard Requests limit the amount of
information that can be sent to the Device during the Setup transaction to four bytes (wValue and wIndex field of
the Setup packet).

The new Push/Pull Command structure allows for much more information to be exchanged between the Host and
the Device by moving the command related parameters from the Setup transaction to the Data transaction of the
transfer.

The Push Command is used to send information to the Device and the Pull Command is used to retrieve
information from the Device.

The Push Command consists of a single Set Request whereas the Pull Command is an atomic sequence of a Set
Request, followed by a Get request. All three Requests closely follow the standard USB Request layout as defined
in the USB Core Specification. The following table details their layout.
Table 5-2: Push and Pull Command Request Layout

bmRequest bRequest wValue wIndex wLength Data

Type
00100001B PUSH Zero Zero Length of Parameter Block
PULL and Parameter Block
Interface
10100001B PULL

Release 4.0 April 2023 135

Universal Serial Bus Device Class Definition for Audio Devices

Bit D7 of the bmRequestType field shall be set to 0b0 for the Set Request and to 0b1 for the Get Request. It is a
class-specific Request (D6..5 = 0b01), directed at an interface (AudioControl or AudioStreaming) of the Audio
Function (D4..0 = 0b00001).

The bRequest field shall contain the PUSH constant for the Push Set Request. It shall contain the PULL constant for
the Pull Set and Get Request. If the field contains a value other than these values respectively, the Request shall
return a Request Error.

The wValue field shall be set to zero. If the field contains a value other than zero, the Request shall return a
Request Error.

The value in the low byte of the wIndex field shall be appropriate to the recipient. Only appropriate Interface
numbers may be used. If the Command specifies an unknown Interface number, the Request shall return a
Request Error.

The high byte of the wIndex field shall be set to zero. If it contains a value other than zero, the Request shall return
a Request Error.

If an Audio Function does not support a certain Command, it shall indicate this by returning a Request Error when
that Command Set Request phase (see further) is issued to the Function. If a certain Push Command is supported,
the associated Pull Command shall also be supported. Pull Commands may be supported without the associated
Push Command being supported. If interrupts are supported, then all necessary Pull Commands shall be
implemented that are required to retrieve the appropriate information from the Audio Function in response to
these interrupts.

The following sections provide more details about the Push And Pull Commands and their associated Address
parameters and Data.

5.3.1.1 PUSH COMMAND

The Push Command is used to send information to the Device. It uses the Parameter Block of a single Set Request
to convey all the information needed by the Device to know how to execute the Push Command. This includes all
necessary addressing information (AddressPart) as well as the new value(s) (DataPart) for the addressed item, if
applicable. The Device shall not ACK the Set Request until it has changed all the internal state required by the Push
Command. A Device indicates that it cannot fulfill the Push Command by returning a Request Error.

The length of the Parameter Block is indicated in the wLength field of the Set Request. The layout of the Parameter
Block is described in the following section. If the parameter values are not supported, the Set Request shall return
a Request Error.

5.3.1.1.1 PUSH COMMAND PARAMETER BLOCK

The Push Command Parameter Block is divided into two separate Parts. The AddressPart of the Parameter Block is
always 12 bytes long and contains all addressing information for the Device to know which addressable item within
the Audio Function is targeted. The DataPart of the Parameter Block contains the data value(s) for the targeted
item. The layout of the DataPart is qualified by the parameters in the AddressPart. Refer to subsequent paragraphs
for a detailed description of the DataPart of the Parameter Block for all possible addressable items.
Table 5-3: Push Command Parameter Block Layout

Offset Field Size Value Description

AddressPart

Release 4.0 April 2023 136

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

0 wParam1 2 Number Command Parameter 1.
2 wParam2 2 Number Command Parameter 2.
4 wParam3 2 Number Command Parameter 3.
6 wParam4 2 Number Command Parameter 4.
8 wParam5 2 Number Command Parameter 5.
10 wParam6 2 Number Command Parameter 6.
DataPart
12 Data n Number The data value(s).

5.3.1.2 PULL COMMAND

The Pull Command is used to retrieve information from the Device. The Pull Command always consists of a Set
Request followed by its associated Get Request without an intervening Set Request to the same item. The
Parameter Block of the Set Request contains all addressing information (AddressPart) for the Device to know for
which item data is being requested by the Host. The Parameter Block of the associated Get Request contains the
actual requested Device data (DataPart). If the Device receives a Pull Command Get Request without having
received a Pull Command Set Request first, it shall indicate this by returning a Request Error on the Pull Command
Get Request.

5.3.1.2.1 PULL COMMAND SET REQUEST PARAMETER BLOCK

The Pull Command Set Request Parameter Block consists of the AddressPart only. It contains all addressing
information for the Device to know which addressable item within the Audio Function is targeted. The DataPart is
retrieved from the Device in the parameter Block of the associated atomic Pull Command Get Request.

The length of the Parameter Block is indicated in the wLength field of the Pull Command Set Request. This field
shall always be set to a value of 12. If the parameter values are not supported, the Set Request shall return a
Request Error.

The layout of the Pull Command Set Request Parameter Block is as follows.
Table 5-4: Pull Command Set Request Parameter Block Layout

Offset Field Size Value Description

0 wParam1 2 Number Command Parameter 1.

2 wParam2 2 Number Command Parameter 2.
4 wParam3 2 Number Command Parameter 3.
6 wParam4 2 Number Command Parameter 4.
8 wParam5 2 Number Command Parameter 5.
10 wParam6 2 Number Command Parameter 6.

Release 4.0 April 2023 137

Universal Serial Bus Device Class Definition for Audio Devices

5.3.1.2.2 PULL COMMAND GET REQUEST PARAMETER BLOCK

The data value(s) for the Pull Command are returned in the Parameter Block of the Pull Command Get Request. It
consists of the DataPart only and contains the data value(s) for the targeted item.

The length of the Parameter Block to return is indicated in the wLength field of the Get Request. If the Parameter
Block is longer than what is indicated in the wLength field, only the initial bytes of the Parameter Block are
returned. If the Parameter Block is shorter than what is indicated in the wLength field, the Device indicates the end
of the control transfer by sending a short packet when further data is requested. The layout of the Get Request
Parameter Block is qualified by the parameters in the Address Parameter Block of the associated Pull Command
Set Request. Refer to subsequent paragraphs for a detailed description of the DataPart for all possible addressable
items.

The layout of the Pull Command Get Request Parameter Block is as follows.
Table 5-5: Pull Command Get Request Parameter Block Layout

Offset Field Size Value Description

DataPart
0 Data n Number Data value(s) returned by the Device.

The remainder of this section describes the class-specific Commands and their characteristics used to manipulate
the incorporated AudioControls, class-specific Strings and Descriptors.

5.3.2 AUDIOCONTROL COMMANDS CHARACTERISTICS

When responding to an AudioControl Push Command, an Audio Function shall not complete the Status Stage of
the Control Transfer until the Command has been successfully interpreted and the Audio Function has successfully
completed the associated action. The Host always has the option to abort or abandon a pending Control Transfer if
it deems to take too long to complete.

The following sections describe the possible Commands that can be used to manipulate the AudioControls an
Audio Function exposes through its Interfaces, Endpoints, and Entities. The same layout of the Parameter Blocks is
used for both the Push and Pull Commands.

5.3.2.1 AUDIOCONTROL INTERRUPT CAPABILITIES

An AudioControl shall generate an interrupt when any of its Attributes change other than through Host
manipulation.

5.3.2.2 AUDIOCONTROL ATTRIBUTES

Each AudioControl within an Entity may have one or more Attributes associated with it. Currently defined
Attributes for an AudioControl are its:

• Current Attribute (CUR)

• Next Attribute (NEXT)
• Capabilities Attribute (CAP)
• Range Attribute (RANGE)

Attributes are manipulated by issuing the appropriate Push and Pull Commands to the targeted AudioControl as
detailed below.

Release 4.0 April 2023 138

Universal Serial Bus Device Class Definition for Audio Devices

5.3.2.2.1 CURRENT AND NEXT ATTRIBUTES

The Current (CUR) Attribute is used to manipulate the current setting of an AudioControl. Manipulating this
Attribute has immediate effect on the actual setting of the AudioControl.

In some cases, it is desirable to apply settings to multiple AudioControls simultaneously. For example, changing the
settings for a Parametric Equalizer Section (Center Frequency, Q Factor, and Gain) sequentially rather than
simultaneously may introduce totally undesired side effects and artifacts. To cover these cases, this specification
supports the NEXT Attribute and the Commit Command. Manipulating the NEXT Attribute Arms the AudioControl
with a next value without changing its CUR Attribute; this AudioControl is said to be Armed. The Commit Command
will only complete successfully after the NEXT Attributes of all the Armed AudioControls have been applied to their
respective CUR Attributes. This applies to either all AudioControls that have been updated since the last Commit
Command, or only to those AudioControls that are a Member of the CommitGroup that is indicated in the Commit
Command. This effectively changes all the affected AudioControls in the Audio Function simultaneously (within the
limits of the underlying firmware and hardware) and the NEXT and CUR Attributes of the affected AudioControls
will now have the same value. After the Commit Command completes, either successfully or resulting in a request
error, all affected Armed AudioControls return to the Unarmed state and subsequent Commit Commands have no
impact on these AudioControls unitl their NEXT Attribute is updated again.

Note that manipulating the CUR Attribute of an AudioControl that has a preloaded NEXT Attribute, does NOT
change the value of its NEXT Attribute. Whenever the value of the CUR Attribute is changed, the NEXT Attribute
retains its preloaded value. This means that a subsequent Commit Command will overwrite the value of the CUR
attribute that was issued between the write action to the NEXT Attribute and the exercution of the Commit
Command.

AudioControls that are intended to be manipulated simultaneously can declare their CUR Attribute as Read (r) and
solely rely on the update mechanism described above (staging all the NEXT Attributes involved, followed by a
Commit Command).

Whenever the USB Device containing the Audio Function receives a Set Configuration Request, all Armed
AudioControls shall no longer be Armed.

5.3.2.2.2 CAPABILITIES ATTRIBUTE

The mandatory Capabilities (CAP) Attribute is used to retrieve all necessary information from the AudioControl for
the Host software to determine how to successfully interact with the AudioControl. The data structure returned
via a Get(CAP) Command is as follows.
Bit D0 of the bmControlCaps field indicates the implemented support level for the CUR Attribute.

• For AudioControls that are allowed by specification to have the CUR Attribute implemented as Read-Only
(RO) or Read-optional-Write (RoW), bit D0 set to 0b0 indicates that the Audio Function has implemented
the CUR Attribute as Read (r). If D0 is set to 0b1, then the Audio Function has implemented the CUR
Attribute as Read-Write (rw).
• For AudioControls that are allowed by specification to have the CUR Attribute implemented as Write-Only
(WO) bit D0 shall be set to 0b0.
• For AudioControls that are required by specification to have the CUR Attribute implemented as
mandatory-Read-Write (mRW), bit D0 shall be set to 0b1.

Bit D1 of the bmControlCaps field indicates whether the NEXT Attribute is implemented (D1 = 0b1) or not (D1 =
0b0). If implemented, it shall always be implemented as Read-Write (rw).

Release 4.0 April 2023 139

Universal Serial Bus Device Class Definition for Audio Devices

Bit D2 of the bmControlCaps field indicates whether the RANGE Attribute uses the Triplet format (D2 = 0b0) or the
ValueList format (D2 = 0b1). Bit D2 shall be set to 0b0 when the RANGE Attribute is Prohibited for the
AudioControl.

Bit D3 of the bmControlCaps field indicates whether the AudioControl is calibrated for a specific use (D3 = 0b1) or
not (D3 = 0b0). When set, bit D3 indicates that there is a preferred value of the AudioControl that is calibrated
such that the resulting effect on the audio stream is deemed the preferred level of impact for that specific use by
the manufacturer.

Currently the use of Bit D3 is only supported for Gain Controls that are in a microphone signal path. See Section
5.3.3.5.3, “Gain Control” for details. The use of this bit for other AudioControls is reserved for future use.

Bit D4 of the bmControlCaps field indicates whether the AudioControl is intended by the manufacturer to be
hidden from the user (D4 = 0b1) or not (D4 = 0b0).

Note: The CUR and CAP Attributes shall always be implemented and therefore, there is no need to express
this in the bmControlCaps field.
Table 5-6: Capabilities Attribute DataPart

wLength 1

Offset Field Size Value Description

0 bmControlCaps 1 Bitmap D0: CUR Attribute implementation
support level.
0 = Read (r) for RoW Controls or Write
(w) for WO Controls.
1 = Read-Write (rw).
D1: NEXT Attribute Implementation.
0 = Not Implemented.
1 = Implemented.
D2: RANGE Attribute Format.
0 = Array of Triplets.
1 = ValueList.
D3: Calibrated AudioControl.
0 = No.
1 = Yes.
D4: User-Hidden.
0 = No.
1 = Yes.
D7..5: Reserved.

5.3.2.2.3 RANGE ATTRIBUTE

The RANGE Attribute provides information about the limitations the AudioControl imposes on the allowed settings
of the CUR and NEXT Attributes. The RANGE Attribute either consists of an array of sub-ranges or an enumeration
of possible values. The format a particular AudioControl uses for its RANGE Attribute is indicated by the RANGE
Attribute Format bit in the AudioControl’s Capabilities Attribute.

If the RANGE Attribute Format bit indicates “Array of Triplets”, then sub-ranges are described via the Minimum
(MIN), Maximum (MAX), and Resolution (RES) fields. They are always grouped in triplets of the form [MIN, MAX,

Release 4.0 April 2023 140

Universal Serial Bus Device Class Definition for Audio Devices

RES]. The RANGE Attribute supports an array of these triplets so that discontinuous multiple subranges can be
accurately reported. The first element in the Parameter Block contains the number of subranges the AudioControl
supports. Subsequent triplet elements in the Parameter Block correspond to each of the subranges. The subranges
shall be ordered in ascending order (from lower values to higher values). Individual subranges shall not overlap
(i.e., the MAX value of the previous subrange shall be less than the MIN value of the next subrange). If a subrange
consists of only a single value, the corresponding triplet shall contain that value for both its MIN and MAX sub-
Attribute and the RES sub-Attribute shall be set to zero.

If the RANGE Attribute Format bit indicates “ValueList”, then the RANGE Attribute exposes an enumeration of
possible values for the CUR and NEXT Attributes. The first value in the list indicates the number of elements in the
enumeration.

In all cases, the values returned by the RANGE Attribute shall use the same format as the CUR and NEXT Attributes
as defined by this specification.

As an example, consider a (hypothetical) Control that takes the following values for its CUR Attribute:

• - dB
• -70 dB to -40 dB in steps of 3 dB
• -38 dB to -20 dB in steps of 2 dB
• -19 dB to 0 dB in steps of 1 dB

One possible layout of the RANGE Attribute is then:

RANGE Attribute Format = 0

wNumSubRanges = 3
RANGE(1) = [-70, -40, 3]
RANGE(2) = [-38, -20, 2]
RANGE(3) = [-19, 0, 1]

Another way of representing the same Control is:

wNumSubRanges = 3
RANGE(1) = [-70, -43, 3]
RANGE(2) = [-40, -22, 2]
RANGE(3) = [-20, 0, 1]

It is left to the designer to choose a suitable representation.

As a second example, consider a Mixer Control that is implemented as Read (r) and that has a fixed value of -6 dB.
The RANGE Attribute for that Mixer Control is then:

Range Attribute Format = 1

wNumValues = 1
Value(1) = -6

In the Sections that describe the AudioControls in detail, the support level for the AudioControl Attributes is
explicitly called out.

5.3.2.3 AUDIOCONTROL READ/WRITE PRIVILEGES

Note: It is important to make a distinction between the Read/Write privilege allowed by the specification
and the actual Read/Write privilege an Audio Function chooses to implement for an AudioControl’s

Release 4.0 April 2023 141

Universal Serial Bus Device Class Definition for Audio Devices

Attributes. The specification privileges are expressed as Read-Only (RO) or Read-optional-Write

(RoW) for AudioControl Attributes that may be implemented as either Read (r) or Read-Write (rw)
respectively, and mandatory-Read-Write (mRW) for AudioControl Attributes that shall be
implemented as Read-Write (rw). The specification privileges are further expressed as Write-Only
(WO) for AudioControl Attributes that shall be implemented as Write (w).

The actual Read/Write privilege an Audio Function chooses to implement for an AudioControl’s
Attributes (indicated by the abbreviations (r) for Read, (w) for Write, and (rw) for Read-Write) can be
retrieved via the mandatory CAP Attribute for each AudioControl. For details, see Section 5.3.2.2.2,
“Capabilities Attribute.”

This specification defines the Read/Write privileges of an AudioControl based on the Read/Write privilege of its
CUR Attribute as follows:

• Some AudioControls are defined as Read-Only (RO). For this type of AudioControls, the CUR Attribute shall be
Read (r) and the NEXT Attribute shall not be supported. It is recommended that the RANGE Attribute be
supported, if only to assist the Host with meaningful information for UI display purposes.
• Most AudioControls are defined as Read-optional-Write (RoW). This means that a particular implementation
can decide whether to implement the AudioControl’s CUR Attribute either as Read-Write (rw) or as Read (r).
Most AudioControls that implement their CUR Attribute as Read (r) usually do not support the NEXT Attribute.
However, some AudioControls may implement their CUR Attribute as Read (r) and provide a NEXT Attribute
(always implemented as Read-Write (rw)) and rely on the presence and use of the Commit Control to change
the current value. In other words, these AudioControls cannot have the value of their CUR Attribute changed
by the driver directly but only through an update via the NEXT Attribute followed by a Commit. For
AudioControls that are implemented as Read-Write (rw), the NEXT Attribute may be supported and shall
always be implemented as Read-Write (rw).
• The Commit Control is the only AudioControl defined as Write-Only (WO) by the current specification. The
Commit Control’s CUR Attribute shall be implemented as Write (w) and the NEXT Attribute shall not be
supported. The RANGE Attribute shall not be supported for this AudioControl.
• All AudioControls shall support the Capabilities Attribute and it shall be implemented as Read (r).
• All AudioControls shall have the RANGE Attribute, either explicitly implemented by advertising MIN, MAX, and
RES or VALUELIST values. However, some AudioControls have a range that is fixed by this specification or
inherent by their type. These AudioControls shall not advertise a RANGE Attribute. All Boolean type
AudioControls fall into this category, for example. Explicitly advertised RANGE attributes shall always be
implemented as Read (r). Note, however, that an implementation is allowed to update the value(s) of the
RANGE attribute as a result of an external event.

In summary, the specification may allow the Read/Write privilege of an Attribute to be Read-Only (RO), Read-
optional-Write (RoW), mandatory-Read-Write (mRW), or Write-Only (WO). An implementation shall always
express the Read/Write privilege of an Attribute as either Read (r), Read-Write (rw), or Write (w).

5.3.2.4 AUDIOCONTROL COMMAND ADDRESSPART LAYOUT

The AudioControl Push and Pull Commands are used to manipulate an Attribute of an AudioControl inside the
Audio Function. The following table details the AddressPart layout.

Release 4.0 April 2023 142

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-7: AudioControl Command AddressPart Layout

Offset Field Size Value Description

0 wEntityID 2 Number Entity ID of the addressed Item.
2 wCS 2 Constant Control Selector for the targeted Item.
4 wAttribute 2 Constant Attribute of the targeted Item.
6 wOCN 2 Number Output Channel Number on which the
targeted Item resides.
8 wICN 2 Number Input Channel Number on which the
targeted Item resides.
10 wIPN 2 Number Input Pin Number on which the targeted
Item resides.

The wEntityID field shall contain the unique ID of the Entity (Clock Entity ID, Unit ID, Terminal ID, Power Domain
ID, or Connector ID) within which the targeted AudioControl resides. When addressing AudioControls residing
within an Interface, the wEntityID field shall be set to zero. The values in the wEntityID field shall be appropriate
to the recipient. Only existing Entities in the Audio Function or in the AudioStreaming Interfaces may be used. If
the Command specifies an unknown or non-Entity ID, the Command Set Request shall return a Request Error.

The wCS field specifies the Control Selector (CS). The Control Selector indicates which type of AudioControl this
Command is manipulating. If the Command specifies an unknown or unsupported CS for the targeted Entity, the
Command Set Request shall return a Request Error.

The wAttribute field contains a constant, identifying which Attribute of the targeted AudioControl is to be
manipulated. Possible Attributes for an AudioControl are its:

• Current Attribute (CUR)

• Next Attribute (NEXT)
• Capabilities Attribute (CAP)
• Range Attribute (RANGE)

If the targeted AudioControl does not support modification of a certain Attribute, the Command Set Request shall
return a Request Error when an attempt is made to modify that Attribute. In most cases, only the CUR Attribute
will be supported for the Push Command.

This specification does not prevent a designer from having the Audio Function adjust a RANGE Attribute due to an
external event and alert the Host of this change through an interrupt.

For the list of Request constants, refer to Appendix A.22, “Class-specific Attribute Codes.”

The wOCN field specifies the Output Channel Number (OCN). The OCN is equal to the number of the Output Pin
Channel on which the AudioControl resides. If an AudioControl is output channel independent, then the OCN shall
be set to zero (OCN = 0). If the Command specifies an unknown or unsupported OCN for the targeted Entity, the
Command Set Request shall return a Request Error.

The wICN field specifies the Input Channel Number (ICN). The ICN is equal to the number of the Input Pin Channel
on which the AudioControl resides. If an AudioControl is input channel independent, then the ICN shall be set to
zero (ICN = 0). If the Command specifies an unknown or unsupported ICN for the targeted Entity, the Command Set
Request shall return a Request Error.

Release 4.0 April 2023 143

Universal Serial Bus Device Class Definition for Audio Devices

The wIPN field specifies the Input Pin Number (IPN). The IPN is equal to the Input Pin Number on which the
AudioControl resides. If an AudioControl is Input Pin independent, then the IPN shall be set to zero (IPN = 0). If the
Command specifies an unknown or unsupported IPN for the targeted Entity, the Command Set Request shall
return a Request Error.

The OCN:ICN:IPN triplet shall always reflect the correct values, i.e., for a Unit that only has a single Input Pin, the
IPN shall be set to one, and the ICN shall be set equal to the OCN since the input channel and the output channel
are exactly the same.

5.3.2.5 AUDIOCONTROL COMMAND DATAPART LAYOUT

With a few exceptions, almost all AudioControl Commands manipulate a single AudioControl Attribute parameter
during a Push or Pull Command. For those Commands, the possible DataPart layouts for the CUR, NEXT, and
RANGE Attributes can be divided into three layout categories, depending on the byte size of the AudioControl’s
CUR Attributes. See further for the Layout 1 (Section 5.3.2.6.1, “Layout 1 DataPart”), Layout 2 (Section 5.3.2.6.2,
“Layout 2 DataPart”), and Layout 3 (Section 5.3.2.6.3, “Layout 3 DataPart”) definitions. The layout for the CAP
Attribute is defined in Section 5.3.2.2.2, “Capabilities Attribute”. The following paragraphs specify the layout of the
CUR, NEXT, and RANGE DataParts for the three categories.

For those Commands that use a custom DataPart layout, the actual layout is explicitly defined in the relevant
sections.

5.3.2.6 USE OF WILDCARDS IN THE OCN:ICN:IPN TRIPLET

This specification allows for a wildcard value to be used in any or all of the values in the OCN:ICN:IPN triplet. The
wildcard value shall be 0xFFFF. When a value of the OCN:ICN:IPN triplet is set to the wildcard value, this means
that the Attribute of all AudioControls in the range of that wildcard triplet value are manipulated at the same time.
The DataPart shall contain an Attribute value for all the AudioControls in the range of that wildcard triplet. Note
that only Attribute values for AudioControls that are actually implemented in the Entity, as indicated by the
d(a)BmOptControls field(s) in the Entity Descriptor, shall be present in the DataPart.

The ordering of the values in the DataPart is from lowest channel number to highest channel number, OCN first,
followed by ICN, followed by IPN. As a hypothetical example, consider an Entity that has 2 Input Pins, 2 input
channels on Input Pin 1, 1 input channel on Input Pin 2, and 2 channels on its Output Pin. Assume further that
AudioControls are present as indicated by the dots in the following figure.
Figure 5-1: OCN:ICN:IPN Example

Release 4.0 April 2023 144

Universal Serial Bus Device Class Definition for Audio Devices

For this Entity, the DataPart for the CUR Attribute of the AudioControl would look like this for the following uses of
the wildcard:

OCN:ICN:IPN = 0xFFFF:0xFFFF:0xFFFF: CUR1:1:1; CUR2:2:1; CUR1:1:2; CUR2:1:2

OCN:ICN:IPN = 0xFFFF:0xFFFF:0x0001: CUR1:1:1; CUR2:2:1

OCN:ICN:IPN = 0xFFFF:0x0001:0xFFFF: CUR1:1:1; CUR1:1:2; CUR2:1:2

OCN:ICN:IPN = 0x0001:0xFFFF:0xFFFF: CUR1:1:1; CUR1:1:2

OCN:ICN:IPN = 0xFFFF:0x0002:0x0001: CUR2:2:1

Other Attributes can be retrieved in the same fashion. It is the responsibility of the Host to check whether the
expected volume of information will fit within the limitations of the DataPart maximum size (65535 – 12 = 65523
bytes). Note that for the RANGE Attribute, the actual size of the information returned for each AudioControl may
not be known beforehand. Therefore, the use of wildcards while manipulating the RANGE Attribute needs to be
approached with caution.

5.3.2.6.1 LAYOUT 1 DATAPART

The DataPart for a 1-byte sized CUR or NEXT Attribute of an AudioControl is as follows:
Table 5-8: 1-byte AudioControl CUR or NEXT DataPart

wLength 1

Offset Field Size Value Description

0 bCUR 1 Number The setting for the CUR or NEXT Attribute of
bNEXT the targeted AudioControl.

The associated DataPart for the RANGE Attribute of that AudioControl when expressed as an Array of Triplets is as
follows:
Table 5-9: 1-byte AudioControl RANGE DataPart (Array of Triplets)

wLength 2+3*n

Offset Field Size Value Description

0 wNumSubRanges 2 Number The number of subranges of the targeted
AudioControl: n.
2 bMIN(1) 1 Number The setting for the MIN Attribute of the first
subrange of the targeted AudioControl.
3 bMAX(1) 1 Number The setting for the MAX Attribute of the first
subrange of the targeted AudioControl.
4 bRES(1) 1 Number The setting for the RES Attribute of the first
subrange of the targeted AudioControl.
… … … … …
2+3*(n-1) bMIN(n) 1 Number The setting for the MIN Attribute of the last
subrange of the targeted AudioControl.
3+3*(n-1) bMAX(n) 1 Number The setting for the MAX Attribute of the last
subrange of the targeted AudioControl.

Release 4.0 April 2023 145

Universal Serial Bus Device Class Definition for Audio Devices

wLength 2+3*n
4+3*(n-1) bRES(n) 1 Number The setting for the RES Attribute of the last
subrange of the targeted AudioControl.

The associated DataPart for the RANGE Attribute of that AudioControl when expressed as a ValueList is as follows:
Table 5-10: 1-byte AudioControl RANGE DataPart (ValueList)

wLength 2+n

Offset Field Size Value Description

0 wNumValues 2 Number The number of values in the ValueList: n.
2 bValue(1) 1 Number The first value in the ValueList for the targeted
AudioControl.
… … … … …
2+(n-1) bValue(n) 1 Number The last value in the ValueList for the targeted
AudioControl.

5.3.2.6.2 LAYOUT 2 DATAPART

The DataPart for a 2-byte sized CUR or NEXT Attribute of an AudioControl is as follows:
Table 5-11: 2-byte AudioControl CUR or NEXT DataPart

wLength 2

Offset Field Size Value Description

0 wCUR 2 Number The setting for the CUR or NEXT Attribute of
wNEXT the targeted AudioControl.

The associated DataPart for the RANGE Attribute of that AudioControl when expressed as an Array of Triplets is as
follows:
Table 5-12: 2-byte AudioControl RANGE DataPart (Array of Triplets)

wLength 2+6*n

Offset Field Size Value Description

0 wNumSubRanges 2 Number The number of subranges of the targeted
AudioControl: n.
2 wMIN(1) 2 Number The setting for the MIN Attribute of the first
subrange of the targeted AudioControl.
4 wMAX(1) 2 Number The setting for the MAX Attribute of the first
subrange of the targeted AudioControl.
6 wRES(1) 2 Number The setting for the RES Attribute of the first
subrange of the targeted AudioControl.
… … … … …
2+6*(n-1) wMIN(n) 2 Number The setting for the MIN Attribute of the last
subrange of the targeted AudioControl.
4+6*(n-1) wMAX(n) 2 Number The setting for the MAX Attribute of the last
subrange of the targeted AudioControl.

Release 4.0 April 2023 146

Universal Serial Bus Device Class Definition for Audio Devices

wLength 2+6*n
6+6*(n-1) wRES(n) 2 Number The setting for the RES Attribute of the last
subrange of the targeted AudioControl.

The associated DataPart for the RANGE Attribute of that AudioControl when expressed as a ValueList is as follows:
Table 5-13: 2-byte AudioControl RANGE DataPart (ValueList)

wLength 2+2*n

Offset Field Size Value Description

0 wNumValues 2 Number The number of values in the ValueList: n.
2 wValue(1) 1 Number The first value in the ValueList for the targeted
AudioControl.
… … … … …
2+2*(n-1) wValue(n) 1 Number The last value in the ValueList for the targeted
AudioControl..

5.3.2.6.3 LAYOUT 3 DATAPART

The DataPart for a 4-byte sized CUR or NEXT Attribute of an AudioControl is as follows:
Table 5-14: 4-byte AudioControl CUR or NEXT DataPart

wLength 4

Offset Field Size Value Description

0 dCUR 4 Number The setting for the CUR or NEXT Attribute of
dNEXT the targeted AudioControl.

The associated DataPart for the RANGE Attribute of that AudioControl when expressed as an Array of Triplets is as
follows:
Table 5-15: 4-byte AudioControl RANGE DataPart (Array of Triplets)

wLength 2+12*n

Offset Field Size Value Description

0 wNumSubRanges 2 Number The number of subranges of the targeted
AudioControl: n
2 dMIN(1) 4 Number The setting for the MIN Attribute of the first
subrange of the targeted AudioControl.
6 dMAX(1) 4 Number The setting for the MAX Attribute of the first
subrange of the targeted AudioControl.
10 dRES(1) 4 Number The setting for the RES Attribute of the first
subrange of the targeted AudioControl.
… … … … …
2+12*(n-1) dMIN(n) 4 Number The setting for the MIN Attribute of the last
subrange of the targeted AudioControl.
6+12*(n-1) dMAX(n) 4 Number The setting for the MAX Attribute of the last
subrange of the targeted AudioControl.

Release 4.0 April 2023 147

Universal Serial Bus Device Class Definition for Audio Devices

wLength 2+12*n
10+12*(n-1) dRES(n) 4 Number The setting for the RES Attribute of the last
subrange of the targeted AudioControl.

The associated DataPart for the RANGE Attribute of that AudioControl when expressed as a ValueList is as follows:
Table 5-16: 4-byte AudioControl RANGE DataPart (ValueList)

wLength 2+4*n

Offset Field Size Value Description

0 wNumValues 2 Number The number of values in the ValueList: n.
2 dValue(1) 1 Number The first value in the ValueList for the targeted
AudioControl..
… … … … …
2+4*(n-1) dValue(n) 1 Number The last value in the ValueList for the targeted
AudioControl

5.3.2.7 AUDIOCONTROL SECTIONS LAYOUT

The layout of the following sections that describe the possible AudioControls in more detail is as follows:

First, there is a paragraph briefly describing the functionality of the AudioControl. Most AudioControls are optional
for implementations to support. However, some AudioControls shall be implemented by all class-compliant Audio
Function implementations. Some AudioControls may be required under certain conditions. All AudioControls are
clearly marked as Mandatory, Optional, or Conditionally Required in the header of the AudioControl Table, using
the abbreviations MAN, OPT, or CR respectively (see below). If an AudioControl is marked as Conditionally
Required, then the conditions are explained in the descriptive paragraph preceding the AudioControl Table.

Note: The mandatory or optional character of the AudioControl is qualified by the fact whether the Audio
Function implements a certain Entity that contains the AudioControl. For example, the Power Domain
Control is marked as Mandatory, but obviously it is only mandatory if the Audio Function implements
a Power Domain.

Then a standardized AudioControl Table follows that summarizes several different aspects of the AudioControl and
applicable values for various fields in the AudioControl Command that is used to manipulate the AudioControl.
The first rows of the table enumerate the possible Attributes (CUR, NEXT). For each Attribute, the required support
level for the Attribute is indicated in the SL (Support Level) column. Mandatory (M) indicates that the AudioControl
shall support this Attribute. Optional (O) indicates that the AudioControl may choose to support the Attribute,
while Prohibited (P) indicates that the AudioControl is not allowed to support the Attribute. Implicit (I) indicates
that the Attribute support is fixed by the specification or by the type of the AudioControl.

The RW column indicates the level of freedom (allowed by this specification) an implementation has regarding the
Read-Write privilege of the Attribute: whether the Attribute can be implemented as Read-Only (RO), mandatory-
Read-Write (mRW), Read with optional Write (RoW), or Write-Only (WO).

The RANGE rows together specify the range of values that are applicable for the CUR and NEXT Attribute. The
VALUE LIST row specifies a list of values that is applicable for the CUR and NEXT Attribute.

The acronym N/A is used to indicate “Not Applicable”.

Release 4.0 April 2023 148

Universal Serial Bus Device Class Definition for Audio Devices

The CS row contains the Control Selector value that shall be used for that AudioControl.

The OCN:ICN:IPN row contains information about the values that can be used to target a particular AudioControl
within the Entity.

Finally, if the AudioControl has a custom DataPart layout (other than the default DataPart layouts), that layout is
explicitly listed in subsequent rows, as applicable.

The following syntax is used to indicate values and value ranges:

[value] ::= single value

[value1 to value2] ::= a single value in the range between and including value1 and value2

[value1 to value2]+ ::= list of values, containing one or more elements in the range between and including value1
and value2

Release 4.0 April 2023 149

Universal Serial Bus Device Class Definition for Audio Devices

5.3.2.7.1 EXAMPLE 1: AUDIOCONTROL TABLE WITH DEFAULT DATAPART LAYOUT

Table 5-17: AudioControl Table with Default DataPart Layout

Name SL RW MAN/OPT/CR

CUR M RO/RoW/mRW/WO

NEXT M/O/P mRW

MIN Applicable range

MAX Applicable range

RANGE M/I RO
RES Applicable range

VLIST Applicable Value List

Field Value

CS Control Selector Value

OCN:ICN:IPN As applicable

Parameter Block Layout 1, 2, or 3

5.3.2.7.2 EXAMPLE 2: AUDIOCONTROL TABLE WITH CUSTOM DATAPART LAYOUT

Table 5-18: AudioControl Table with Custom DataPart

Name SL RW MAN/OPT/CR

CUR M RO/RoW/mRW/WO

NEXT M/O/P mRW

MIN Applicable range

MAX Applicable range

RANGE M/I RO
RES Applicable range

VLIST Applicable Value List

Field Value

CS Control Selector Value

OCN:ICN:IPN As applicable

Parameter Block Length Length in bytes of the CUR or NEXT custom DataPart

Offset Field Size Value Description

0 Field Name x As Applicable …

… … … … …

5.3.3 AUDIOCONTROL COMMANDS

The following sections describe the Commands an Audio Function may support for its AudioControl Interface. The
same layout of the DataParts is used for both the Push and Pull Commands.

AudioControl Commands are directed to the Entity that contains the targeted AudioControl via the single
AudioControl Interface of the Audio Function.

Release 4.0 April 2023 150

Universal Serial Bus Device Class Definition for Audio Devices

5.3.3.1 COMMON AUDIOCONTROLS

The following sections describe AudioControls that may appear in several Entity types. They are described here
only once and a reference to these AudioControl descriptions is provided for all those Entities that may
incorporate any of these AudioControls.

5.3.3.1.1 BYPASS CONTROL

The Bypass Control is implemented either on each Pin channel intersection individually or on the Unit as whole. In
the former case, it is used to include the functionality offered by an Entity on a particular OCN:ICN:IPN Node or to
bypass that functionality on that Node. In the latter case, it used to either include the functionality of an Entity in
the signal path or bypass the Entity entirely (OCN=ICN=IPN = 0). The input Cluster on Input Pin one is then routed
unaltered to the output of the Entity.

For most Entities, the Bypass Control is optional, for some it is required to be implemented.
Table 5-19: Bypass Control Characteristics

Bypass SL RW OPT or MAN

CUR M RoW [0 (0x00) to 1 (0x01)]; Active or Bypassed respectively.

NEXT O mRW

MIN N/A

MAX N/A
RANGE I N/A
RES N/A

VLIST N/A

Field Value

CS XX_BYPASS_CONTROL (where XX shall be replaced by the appropriate two-letter abbreviation for the particular Entity)

OCN:ICN:IPN As applicable

Parameter Block Layout 1 (See Section 5.3.2.6.1, “Layout 1 DataPart”)

5.3.3.1.2 CLUSTER CONTROL

The Cluster Control is used to select among different logical Cluster Configurations the Entity may support on its
Output Pin. A value of zero indicates that the Cluster Configuration is inherited from the Entity’s Input Pin 1. For an
Input Terminal (that has no Input Pin 1), a value of zero indicates that its associated AudioStreaming Interface is
currently set to Alternate Setting zero and consequently, there is no Cluster Configuration available. This also
means that the Active Cluster Control of the Input Terminal shall report that the Cluster is Inactive.

Whenever this AudioControl is present, the Cluster Active Control shall also be present. It is important to note that
this does not mean that the Cluster content or the state of the Cluster Active Control is directly inherited from
Input Pin 1. The range for this AudioControl is from 0 to the number of Clusters, supported on the Output Pin of
the Entity, as reflected in the Entity’s wNrClusters Descriptor field.

Release 4.0 April 2023 151

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-20: Cluster Control Characteristics

Mode Select SL RW OPT

CUR M RoW [0 to bNrClusters]

NEXT O mRW

MIN N/A

MAX N/A
RANGE I N/A
RES N/A

VLIST N/A

Field Value

CS XX_CLUSTER_CONTROL (where XX shall be replaced by the appropriate two-letter abbreviation for the particular Entity)

OCN:ICN:IPN 0:0:0

Parameter Block Layout 1 (See Section 5.3.2.6.1, “Layout 1 DataPart”)

5.3.3.1.3 CLUSTER ACTIVE CONTROL

The Read-Only Cluster Active Control indicates whether the Cluster present on the Output Pin of the Entity is
currently Active or Inactive. The Entity that contains this AudioControl shall decide by itself which value to report,
based on a variety of criteria, including the state of the Clusters being received on its Input Pin(s).
Table 5-21: Cluster Active Control Characteristics

Cluster Active SL RW MAN

CUR M RO [0 (0x00) to 1 (0x01)]; Active or Inactive respectively.

NEXT P N/A

MIN N/A

MAX N/A
RANGE I N/A
RES N/A

VLIST N/A

Field Value

CS XX_CLUSTER_ACTIVE_CONTROL (where XX shall be replaced by the appropriate two-letter abbreviation for the particular Entity)

OCN:ICN:IPN 0:0:0

Parameter Block Layout 1 (See Section 5.3.2.6.1, “Layout 1 DataPart”)

5.3.3.1.4 LATENCY CONTROL

An Audio Function shall either not support this AudioControl (D0 = 0b0 in the bmOptControls field of the class-
specific AudioControl Interface Descriptor) or support this Read-Only AudioControl for every Terminal and Unit
within the Audio Function (D0 = 0b1 in the bmOptControls field of the class-specific AudioControl Interface
Descriptor). Terminal latencies shall include all latencies incurred by A/D or D/A converters, encoders, decoders,
etc.

Note: The presence of the Latency Controls is advertised in the class-specific AudioControl Interface
Descriptor and not repeated in every Terminal and Unit Descriptor. Its functionality is described here,
although the AudioControl Interface by itself does not contain a Latency Control.

Release 4.0 April 2023 152

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-22: Latency Control Characteristics

Latency SL RW OPT

CUR M RO [0 ns (0x00000000) to 4,294,967,295 ns (0xFFFFFFFF)]

NEXT P N/A

MIN N/A

MAX N/A
RANGE P N/A
RES N/A

VLIST N/A

Field Value

CS XX_LATENCY_CONTROL (where XX shall be replaced by the appropriate two-letter abbreviation for the particular Entity)

OCN:ICN:IPN 0:0:0

Parameter Block Layout 3 (See Section 5.3.2.6.3, “Layout 3 DataPart”)

5.3.3.2 TERMINAL COMMAND

This Command is used to manipulate the AudioControls inside a Terminal of the Audio Function. The exact layout
of the Command is defined in Section 5.3.2.5, “AudioControl Command DataPart Layout.”

The following paragraphs present a detailed description of all possible AudioControls a Terminal may incorporate.
For each AudioControl, the supported Attributes and their value ranges are specified. Also, the appropriate
AudioControl Selector value and the layout type of the Parameter Blocks are listed. The Control Selector codes are
defined in Appendix A.23.1, “Terminal Control Selectors.”

5.3.3.2.1 CLUSTER CONTROL

This AudioControl shall only be present in an Input Terminal that does not represent an AudioStreaming OUT
Interface. All Output Terminals and any Input Terminal that represents an AudioStreaming OUT Interface shall not
include this AudioControl. It shall always be implemented as Read (r) and shall never return a value of zero. See
Section 5.3.3.1.2, “Cluster Control” for a detailed description. XX shall be replaced by TE.

5.3.3.2.2 CLUSTER ACTIVE CONTROL

This AudioControl shall always be present in an Input Terminal. An Output Terminal shall not include this
AudioControl. See Section 5.3.3.1.3, “Cluster Active Control” for a detailed description. XX shall be replaced by TE.

5.3.3.2.3 VOLTAGE CONTROL

This AudioControl shall only be present in an Input Terminal that represents a microphone that requires either a
phantom power supply or a bias voltage to properly operate. The Voltage Control is used to set the voltage level of
the power supply.

Release 4.0 April 2023 153

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-23: Voltage Control Characteristics

Voltage SL RW OPT

CUR M RoW

NEXT O mRW

MIN [0 V (0x0000) to +255.9961 V (0xFFFF)]

MAX [0 V (0x0000) to +255.9961 V (0xFFFF)]

RANGE M RO
RES [0 V (0x0000) to +255.9961 V (0xFFFF)]

VLIST [0 V (0x0000) to +255.9961 V (0xFFFF)]+

Field Value

CS TE_VOLTAGE_CONTROL

OCN:ICN:IPN The Pin channel number on which the AudioControl resides (including the Primary channel 0):0:0

Parameter Block Layout 2 (See Section 5.3.2.6.2, “Layout 2 DataPart”)

5.3.3.2.4 MOMENTARY EXPOSURE LEVEL CONTROL

The Momentary Exposure Level Control is used to report the Momentary Exposure Level (MEL) for headphones, or
a headphone jack as defined by EN 50332-3:2017, a CENELEC standard that defines a methodology for measuring
the maximum Sound Pressure Level (SPL) experienced by a person listening to headphones. Consequently, only
Output Terminals, representing headphones shall be permitted to have this MEL Control. If implemented, this
AudioControl shall be Read-Only and shall have only the CUR attribute. Responding to the Pull Command returns
the CUR attribute. The value of the MEL Control CUR attribute may range from 0.00 dB(A) to +255.9961 dB(A)
(0xFFFF) in steps of 1/256 dB(A) or 0.00390625 dB(A) (0x0001).

The Control Selector field shall be set to TE_MEL_CONTROL and the Channel Number field shall be set to zero
(Cluster-wide AudioControl). The parameter block for this AudioControl Command uses Layout 2 (See Section
5.2.1.3.2, “Layout 2 Parameter Block.”)

If the Device implements this AudioControl, it shall expose an interrupt endpoint in its AudioControl Interface as
this AudioControl relies on interrupt endpoint communication to deliver a new MEL value every second. Old values
will be overwritten if not read within the one second period. Polling the CUR attribute of this AudioControl is not a
viable alternative and will lead to unreliable results since the availability of a new value solely relies on a 1 s timer
maintained by the Device which could potentially be asynchronous to any clocks to which the Host may have
access.

Release 4.0 April 2023 154

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-24: Momentary Exposure Level Control Characteristics

Voltage SL RW OPT

CUR M RO [0 dB (0x0000) to +255.9961 dB (0xFFFF)]

NEXT O N/A

MIN N/A

MAX N/A
RANGE P N/A
RES N/A

VLIST N/A

Field Value

CS TE_MOMENTARY_EXPOSURE_LEVEL_CONTROL

OCN:ICN:IPN 0:0:0

Parameter Block Layout 2 (See Section 5.3.2.6.2, “Layout 2 DataPart”)

5.3.3.2.5 OVERLOAD CONTROL

The Overload Control is used to indicate the existence of a thermal overload condition within a Terminal. (E.g.,
thermal overload in powered speakers; an amplifier, etc.)
Table 5-25: Overload Control Characteristics

Overload SL RW OPT

CUR M RO [0 (0x00) to 1 (0x01)]; Normal or Overload Condition respectively.

NEXT P N/A

MIN N/A

MAX N/A
RANGE I N/A
RES N/A

VLIST N/A

Field Value

CS TE_ OVERLOAD_CONTROL

OCN:ICN:IPN 0:0:0

Parameter Block Layout 1 (See Section 5.3.2.6.1, “Layout 1 DataPart”)

5.3.3.2.6 SIGNAL CLIPPING CONTROL

The Signal Clipping Control is used to indicate the existence of signal clipping condition within a Terminal, i.e., a
signal level exists outside of the (analog or digital) range that can be handled by the signal path.

Release 4.0 April 2023 155

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-26: Clipping Control Characteristics

Signal Clipping SL RW OPT

CUR M RO [0 (0x00) to 1 (0x01)]; Normal or Clipping Condition respectively.

NEXT P N/A

MIN N/A

MAX N/A
RANGE I N/A
RES N/A

VLIST N/A

Field Value

CS TE_CLIPPING_CONTROL

OCN:ICN:IPN 0:0:0

Parameter Block Layout 1 (See Section 5.3.2.6.1, “Layout 1 DataPart”)

5.3.3.2.7 LATENCY CONTROL

See Section 5.3.3.1.4, “Latency Control” for a detailed description. XX shall be replaced by TE.

5.3.3.3 MIXER UNIT COMMAND

This Command is used to manipulate the AudioControls inside the Mixer Unit. The exact layout of the Command is
defined in Section 5.3.2.5, “AudioControl Command DataPart Layout.”

The following paragraphs present a detailed description of all possible AudioControls the Mixer Unit may
incorporate. For each AudioControl, the supported Attributes and their value ranges are specified. Also, the
appropriate Control Selector value and the layout type of the Parameter Blocks are listed. The Control Selector
codes are defined in Appendix A.23.2, “Mixer Unit Control Selectors.”

5.3.3.3.1 MIXER CONTROL

The Mixer Control is used to manipulate the level at which an input channel of the Mixer Unit is mixed into one of
its output channels.

Silence (i.e., - dB) is represented by code 0x8000. If the AudioControl supports this setting, it shall explicitly
indicate this by either advertising the subrange triplet [MIN: 0x8000; MAX: 0x8000; RES: 0x0000] or as the value
0x8000 being part of the VLIST enumeration.

Release 4.0 April 2023 156

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-27: Mixer Control Characteristics

Mixer SL RW MAN

MCUR M RoW

MNEXT O mRW

MIN [- dB (0x8000), -127.9961 dB (0x8001) to +127.9961 dB (0x7FFF)]

MAX [- dB (0x8000), -127.9961 dB (0x8001) to +127.9961 dB (0x7FFF)]

MRANGE M RO
RES [0 dB (0x0000) to +255.9961 dB (0xFFFF)]

VLIST [- dB (0x8000), -127.9961 dB (0x8001) to +127.9961 dB (0x7FFF)]+

Field Value

CS MU_MIXER_CONTROL

OCN:ICN:IPN As applicable

Parameter Block Layout 2 (See Section 5.3.2.6.2, “Layout 2 DataPart”)

5.3.3.3.2 CLUSTER CONTROL

See Section 5.3.3.1.2, “Cluster Control” for a detailed description. XX shall be replaced by MU.

5.3.3.3.3 CLUSTER ACTIVE CONTROL

This AudioControl shall always be present. See Section 5.3.3.1.3, “Cluster Active Control” for a detailed description.
XX shall be replaced by TE.

5.3.3.3.4 LATENCY CONTROL

See Section 5.3.3.1.4, “Latency Control” for a detailed description. XX shall be replaced by MU.

5.3.3.4 SELECTOR UNIT COMMAND

This Command is used to manipulate the AudioControls inside a Selector Unit. The exact layout of the Command is
defined in Section 5.3.2.5, “AudioControl Command DataPart Layout.”

The following paragraphs present a detailed description of all possible AudioControls the Selector Unit may
incorporate. For each AudioControl, the supported Attributes and their value ranges are specified. Also, the
appropriate AudioControl Selector value and the layout type of the Parameter Blocks are listed. The Control
Selector codes are defined in Appendix A.23.3, “Selector Unit Control Selectors.”

5.3.3.4.1 SELECTOR CONTROL

The Selector Control provides the capability to select a particular Cluster from several incoming Clusters on the
Selector Unit’s Input Pins. The value zero, if available, is used to disconnect the output from all the inputs. In this
case, the output Cluster is Inactive. The range for this AudioControl is from zero (disconnect option supported) or
one (disconnect option not supported) to the number of Input Pins the Selector Unit supports, as reflected in the
Entity’s bNrInputPins Descriptor field.

Release 4.0 April 2023 157

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-28: Selector Control Characteristics

Selector SL RW MAN

CUR M RoW [0 to bNrInputPins]

NEXT O mRW

MIN 0 or 1

MAX bNrInputPins
RANGE M RoW
RES 1

VLIST N/A

Field Value

CS SU_SELECTOR_CONTROL

OCN:ICN:IPN 0:0:0

Parameter Block Layout 1 (See Section 5.3.2.6.1, “Layout 1 DataPart”)

5.3.3.4.2 LATENCY CONTROL

See Section 5.3.3.1.4, “Latency Control” for a detailed description. XX shall be replaced by SU.

5.3.3.5 FEATURE UNIT COMMAND

This Command is used to manipulate the AudioControls inside the Feature Unit. The exact layout of the Command
is defined in Section 5.3.2.5, “AudioControl Command DataPart Layout.”

The following paragraphs present a detailed description of all possible AudioControls the Feature Unit may
incorporate. For each AudioControl, the supported Attributes and their value ranges are specified. Also, the
appropriate AudioControl Selector value and the layout type of the Parameter Blocks are listed. The Control
Selector codes are defined in Appendix A.23.4, “Feature Unit Control Selectors.”

5.3.3.5.1.1 BYPASS CONTROL

See Section 5.3.3.1.1, “Bypass Control” for a detailed description. XX shall be replaced by FU.

5.3.3.5.2 MUTE CONTROL

The Mute Control is one of the building blocks of the Feature Unit. It is used to temporarily suppress all audio on
an outgoing channel.

Release 4.0 April 2023 158

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-29: Mute Control Characteristics

Mute SL RW OPT

CUR M RoW [0 (0x00) to 1 (0x01)]; Unmuted or Muted respectively

NEXT O mRW

MIN N/A

MAX N/A
RANGE I N/A
RES N/A

VLIST N/A

Field Value

CS FU_MUTE_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 1 (See Section 5.3.2.6.1, “Layout 1 DataPart”)

5.3.3.5.3 GAIN CONTROL

The Gain Control is one of the building blocks of the Feature Unit. It is used to scale (amplify or attenuate) the
audio level on a channel in the outgoing Cluster of the Feature Unit.

A Gain Control can be used to control volume for output signal paths and level for input signal paths.

Silence (i.e., - dB) is represented by code 0x8000. If the AudioControl supports this setting, it shall explicitly
indicate this by either advertising the subrange triplet [MIN: 0x8000; MAX: 0x8000; RES: 0x0000] or as the value
0x8000 being part of the VLIST enumeration.
Table 5-30: Gain Control Characteristics

Gain SL RW OPT

CUR M RoW

NEXT O mRW

MIN [- dB (0x8000), -127.9961 dB (0x8001) to +127.9961 dB (0x7FFF)]

MAX [- dB (0x8000), -127.9961 dB (0x8001) to +127.9961 dB (0x7FFF)]

RANGE M RO
RES [0 dB (0x0000) to +255.9961 dB (0xFFFF)]

VLIST [- dB (0x8000), -127.9961 dB (0x8001) to +127.9961 dB (0x7FFF)]+

Field Value

CS FU_GAIN_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 2 (See Section 5.3.2.6.2, “Layout 2 DataPart”)

When the Calibrated bit D3 in the CAP Attribute is set on a microphone signal path’s Gain Control, it indicates that
the implementation has been calibrated so that, when the Gain Control is set to 0 dB, an eminently useful signal
level will be produced at the AudioStreaming Interface when an average user positions the USB Audio Device in
the manufacturer’s intended manner and speaks at a conversational level. It is highly recommended that the
power-up value for the Gain Control be set to 0 dB when the Calibrated bit is set.
In specific engineering terms, the rms level at the AudioStreaming Interface will measure -26 dBFS when the USB
Audio Device is positioned in the manufacturer’s intended manner with respect to a Head and Torso Simulator

Release 4.0 April 2023 159

Universal Serial Bus Device Class Definition for Audio Devices

(HATS) with its mouth simulator producing a superwideband (50 - 14,000 Hz) pink noise stimulus signal at a
nominal level (89 dBSPL), as defined in IEEE Std 269-2019, "IEEE Standard for Measuring Electroacoustic
Performance of Communication Devices”.

Note that there are two common definitions of dBFS and they differ by 3 dB. This document uses the definition
found within IEEE Std 269-2019.

• dBFS: Decibels are relative to full-scale digital saturation. 0 dBFS is the level of a square-wave signal whose
amplitude is at the maximum digital value.

If the microphone signal path happens to have more than one Gain Control in series, each of the Gain Controls
must have the Calibrated bit in its CAP Attribute set for the calibration to be meaningful.

If the USB Audio Device has any other audio signal processing on the microphone signal path that can be adjusted
by the Host driver, the implementation's mic gain calibration shall be performed with those signal processing
parameters set at nominal values, which are very likely the values the manufacturer decided to have those
parameters take on after initial power-up.

5.3.3.5.4 BASS CONTROL

The Bass Control is one of the building blocks of the Feature Unit. It influences the general lower frequency
behavior on an outgoing channel. Other parameters that also influence the lower frequency behavior, such as cut-
off frequency, cannot be altered through this Command.
Table 5-31: Bass Control Characteristics

Bass SL RW OPT

CUR M RoW

NEXT O mRW

MIN [-128.0000 dB (0x8000) to +127.9961 dB (0x7FFF)]

MAX [-128.0000 dB (0x8000) to +127.9961 dB (0x7FFF)]

RANGE M RO
RES [0 dB (0x0000) to +255.9961 dB (0xFFFF)]

VLIST [-128.0000 dB (0x8001) to +127.9961 dB (0x7FFF)]+

Field Value

CS FU_BASS_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 2 (See Section 5.3.2.6.2, “Layout 2 DataPart”)

5.3.3.5.5 MID CONTROL

The Mid Control is one of the building blocks of the Feature Unit. It influences the general medium frequency
behavior on an outgoing channel. Other parameters that also influence the medium frequency behavior, such as
cut-off frequencies, cannot be altered through this Command.

Release 4.0 April 2023 160

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-32: Mid Control Characteristics

Mid SL RW OPT

CUR M RoW

NEXT O mRW

MIN [-128.0000 dB (0x8000) to +127.9961 dB (0x7FFF)]

MAX [-128.0000 dB (0x8000) to +127.9961 dB (0x7FFF)]

RANGE M RO
RES [0 dB (0x0000) to +255.9961 dB (0xFFFF)]

VLIST [-128.0000 dB (0x8001) to +127.9961 dB (0x7FFF)]+

Field Value

CS FU_MID_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 2 (See Section 5.3.2.6.2, “Layout 2 DataPart”)

5.3.3.5.6 TREBLE CONTROL

The Treble Control is one of the building blocks of the Feature Unit. It influences the general higher frequency
behavior on an outgoing channel. Other parameters that also influence the higher frequency behavior, such as cut-
off frequency, cannot be altered through this Command.
Table 5-33: Treble Control Characteristics

Treble SL RW OPT

CUR M RoW

NEXT O mRW

MIN [-128.0000 dB (0x8000) to +127.9961 dB (0x7FFF)]

MAX [-128.0000 dB (0x8000) to +127.9961 dB (0x7FFF)]

RANGE M RO
RES [0 dB (0x0000) to +255.9961 dB (0xFFFF)]

VLIST [-128.0000 dB (0x8001) to +127.9961 dB (0x7FFF)]+

Field Value

CS FU_TREBLE_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 2 (See Section 5.3.2.6.2, “Layout 2 DataPart”)

5.3.3.5.7 GRAPHIC EQUALIZER CONTROL

The Graphic Equalizer Control is one of the optional building blocks of the Feature Unit. The Audio Device Class
definition provides for standard support of a third octave graphic equalizer. The bands are defined according to the
ANSI S1.11-2004 standard. Bands are numbered from 14 (midband frequency of 25 Hz) up to 43 (midband
frequency of 20,000 Hz), making a total of 30 possible bands. The following table lists the band numbers and their
midband frequencies
Table 5-34: Band Numbers and Nominal Midband Frequencies (ANSI S1.11-2004 Standard)

Band Nr. Midband Freq. Band Nr. Midband Freq. Band Nr. Midband Freq.
14 25 Hz 24* 250 Hz 34 2500 Hz

Release 4.0 April 2023 161

Universal Serial Bus Device Class Definition for Audio Devices

Band Nr. Midband Freq. Band Nr. Midband Freq. Band Nr. Midband Freq.
15* 31.5 Hz 25 315 Hz 35 3150 Hz
16 40 Hz 26 400 Hz 36* 4000 Hz
17 50 Hz 27* 500 Hz 37 5000 Hz
18* 63 Hz 28 630 Hz 38 6300 Hz
19 80 Hz 29 800 Hz 39* 8000 Hz
20 100 Hz 30* 1000 Hz 40 10000 Hz
21* 125 Hz 31 1250 Hz 41 12500 Hz
22 160 Hz 32 1600 Hz 42* 16000 Hz
23 200 Hz 33* 2000 Hz 43 20000 Hz
Note: Bands marked with an asterisk (*) are those present in an octave equalizer.

A Feature Unit that supports the Graphic Equalizer Control is not required to implement the full set of filters. A
subset (for example, octave bands) may be implemented. During a Pull Command, the bmBandsPresent field in
the Parameter Block is a bitmap indicating which bands are effectively implemented and thus reported back in the
returned Parameter Block. Consequently, the number of bits set in this field determines the total length of the
returned Parameter Block. During a Push Command, a bit set in the bmBandsPresent field indicates there is a new
setting for that band in the Parameter Block that follows. The new values shall be in ascending order. If the
number of bits set in the bmBandsPresent field does not match the number of parameters specified in the
following block, the Command Set Request shall return a Request Error.

The Parameter Block for the CUR and NEXT Attributes of the Graphic Equalizer Control is as follows:

Release 4.0 April 2023 162

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-35: Graphic Equalizer CUR or NEXT Control Characteristics

Graphic
SL RW OPT
Equalizer

CUR M RoW

NEXT O mRW

VALUE LIST N/A

Field Value

CS FU_GRAPHIC_EQUALIZER_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Length 4 + (2 * number of bits set in bmBandsPresent: 2 * NrBits)

Offset Field Size Value Description

0 bmBandsPresent 4 Bitmap A bit set indicates the band is present:

D0: Band 14 is present.
D1: Band 15 is present.
…
D29: Band 43 is present.
D30: Reserved.
D31: Reserved.

4 wCUR(Lowest) 2 Number The setting for the CUR Attribute of the lowest band
present:
[-127.9961 dB (0x8001) to +127.9961 dB (0x7FFF)]

… … … … …

4+2*(NrBits-1) wCUR(Highest) 2 Number The setting for the CUR Attribute of the highest band
present.

The Parameter Block for the RANGE Attribute of the Graphic Equalizer Control is as follows:

Release 4.0 April 2023 163

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-36: Graphic Equalizer RANGE Control Characteristics

Graphic
SL RW OPT
Equalizer

RANGE M RO

VALUE LIST N/A

Field Value

CS FU_GRAPHIC_EQUALIZER_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Length 2+6*n

Offset Field Size Value Description

0 wNumSubRanges 2 Number The number of subranges of the Graphic Equalizer

Control: n

2 wMIN(1) 2 Number The setting for the MIN Attribute of the first
subrange of the Graphic Equalizer Control

4 wMAX(1) 2 Number The setting for the MAX Attribute of the first
subrange of the Graphic Equalizer Control

6 wRES(1) 2 Number The setting for the RES Attribute of the first
subrange of the Graphic Equalizer Control

… … … … …

2+6*(n-1) wMIN(n) 1 Number The setting for the MIN Attribute of the last
subrange of the Graphic Equalizer Control

4+6*(n-1) wMAX(n) 1 Number The setting for the MAX Attribute of the last
subrange of the Graphic Equalizer Control

6+6*(n-1) wRES(n) 1 Number The setting for the RES Attribute of the last subrange
of the Graphic Equalizer Control

5.3.3.5.8 AUTOMATIC GAIN CONTROL

The Automatic Gain Control (AGC) is one of the building blocks of the Feature Unit. It enables or disables the
circuitry that tries to optimize the gain setting automatically and continuously on an outgoing channel. The
algorithm that achieves this optimization is not defined by this specification and is implementation dependent.

Release 4.0 April 2023 164

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-37: Automatic Gain Control Characteristics

Automatic Gain SL RW OPT

CUR M RoW [0 (0x00) to 1 (0x01)]; Automatic Gain Disabled or Enabled respectively.

NEXT O mRW

MIN N/A

MAX N/A
RANGE I N/A
RES N/A

VLIST N/A

Field Value

CS FU_AUTOMATIC_GAIN_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 1 (See Section 5.3.2.6.1, “Layout 1 DataPart”)

5.3.3.5.9 DELAY CONTROL

The Delay Control is one of the building blocks of the Feature Unit. It impacts the amount of delay that is incurred
on an outgoing channel.
Table 5-38: Delay Control Characteristics

Delay SL RW OPT

CUR M RoW

NEXT O mRW

MIN [0 s (0x00000000) to 1023.999999761581 s (0xFFFFFFFF)]

MAX [0 s (0x00000000) to 1023.999999761581 s (0xFFFFFFFF)]

RANGE M RO
RES [0 s (0x00000000) to 1023.999999761581 s (0xFFFFFFFF)]

VLIST [0 s (0x00000000) to 1023.999999761581 s (0xFFFFFFFF)]+

Field Value

CS FU_DELAY_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 3 (See Section 5.3.2.6.3, “Layout 3 DataPart”)

5.3.3.5.10 BASS BOOST CONTROL

The Bass Boost Control is one of the building blocks of the Feature Unit. It enables or disables the circuitry that
boosts the lower frequency content on an outgoing channel. The algorithm that achieves this boost is not defined
by this specification and is implementation dependent.

Release 4.0 April 2023 165

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-39: Bass Boost Control Characteristics

Bass Boost SL RW OPT

CUR M RoW [0 (0x00) to 1 (0x01)]; Bass Boost Disabled or Enabled respectively.

NEXT O mRW

MIN N/A

MAX N/A
RANGE I N/A
RES N/A

VLIST N/A

Field Value

CS FU_BASS_BOOST_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 1 (See Section 5.3.2.6.1, “Layout 1 DataPart”)

5.3.3.5.11 LOUDNESS CONTROL

The Loudness Control is one of the building blocks of the Feature Unit. It enables or disables the circuitry that
boosts the bass content at lower volumes on an outgoing channel. The algorithm that achieves this boost is not
defined by this specification and is implementation dependent.
Table 5-40: Loudness Control Characteristics

Loudness SL RW OPT

CUR M RoW [0 (0x00) to 1 (0x01)]; Loudness Disabled or Enabled respectively.

NEXT O mRW

MIN N/A

MAX N/A
RANGE I N/A
RES N/A

VLIST N/A

Field Value

CS FU_LOUDNESS_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 1 (See Section 5.3.2.6.1, “Layout 1 DataPart”)

5.3.3.5.12 INPUT GAIN PAD CONTROL

The Input Gain Pad Control is one of the building blocks of the Feature Unit. It is used to scale (amplify or
attenuate) the audio level on a channel in the outgoing Cluster of the Feature Unit, typically in a few coarse steps.

Silence (i.e., - dB) is represented by code 0x8000. If the AudioControl supports this setting, it shall explicitly
indicate this by either advertising the subrange triplet [MIN: 0x8000; MAX: 0x8000; RES: 0x0000] or as the value
0x8000 being part of the VLIST enumeration.

Release 4.0 April 2023 166

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-41: Input Gain Pad Control Characteristics

Input Gain Pad SL RW OPT

CUR M RoW

NEXT O mRW

MIN [- dB (0x8000), -127.9961 dB (0x8001) to +127.9961 dB (0x7FFF)]

MAX [- dB (0x8000), -127.9961 dB (0x8001) to +127.9961 dB (0x7FFF)]

RANGE M RO
RES [0 dB (0x0000) to +255.9961 dB (0xFFFF)]

VLIST [- dB (0x8000), -127.9961 dB (0x8001) to +127.9961 dB (0x7FFF)]+

Field Value

CS FU_INPUT_GAIN_PAD_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 2 (See Section 5.3.2.6.2, “Layout 2 DataPart”)

5.3.3.5.13 PHASE INVERTER CONTROL

The Phase Inverter Control is one of the building blocks of the Feature Unit. It inverts (180°phase shift) the signal
on an outgoing channel.
Table 5-42: Phase Inverter Control Characteristics

Phase Inverter SL RW OPT

CUR M RoW [0 (0x00) to 1 (0x01)]; Phase Inverter Disabled or Enabled respectively.

NEXT O mRW

MIN N/A

MAX N/A
RANGE I N/A
RES N/A

VLIST N/A

Field Value

CS FU_PHASE_INVERTER_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 1 (See Section 5.3.2.6.1, “Layout 1 DataPart”)

5.3.3.5.14 LATENCY CONTROL

See Section 5.3.3.1.4, “Latency Control” for a detailed description. XX shall be replaced by FU.

5.3.3.6 SAMPLE RATE CONVERTER UNIT COMMAND

This Command is used to manipulate the AudioControls inside the SRC Unit. The exact layout of the Command is
defined in Section 5.3.2.5, “AudioControl Command DataPart Layout.”

The following paragraphs present a detailed description of all possible AudioControls the SRC Unit may
incorporate. For each AudioControl, the supported Attributes and their value ranges are specified. Also, the
appropriate AudioControl Selector value and the layout type of the Parameter Blocks are listed. The Control
Selector codes are defined in Appendix A.23.5, “SRC Unit Control Selectors.”

Release 4.0 April 2023 167

Universal Serial Bus Device Class Definition for Audio Devices

5.3.3.6.1 LATENCY CONTROL

See Section 5.3.3.1.4, “Latency Control” for a detailed description. XX shall be replaced by RU.

5.3.3.7 EFFECT UNIT COMMAND

This Command is used to manipulate the AudioControls inside Effect Unit. The exact layout of the Command is
defined in Section 5.3.2.5, “AudioControl Command DataPart Layout.”

The following paragraphs present a detailed description of all possible AudioControls Effect Unit may incorporate.
For each AudioControl, the supported Attributes and their value ranges are specified. Also, the appropriate
AudioControl Selector value and the layout type of the Parameter Blocks are listed. The Control Selector codes are
defined in Appendix A.23.6, “Effect Unit Control Selectors.”

5.3.3.7.1 PARAMETRIC EQUALIZER SECTION EFFECT UNIT

At least one of the optional AudioControls shall be implemented.

5.3.3.7.1.1 BYPASS CONTROL

See Section 5.3.3.1.1, “Bypass Control” for a detailed description. XX shall be replaced by PE.

5.3.3.7.1.2 CENTER FREQUENCY CONTROL

The Center Frequency Control is used to manipulate the center frequency of the filter implemented on an outgoing
channel.
Table 5-43: Center Frequency Control Characteristics

Center
SL RW OPT
Frequency

CUR M RoW

NEXT O mRW

MIN [0 Hz (0x00000000) to 65,535.9999847 Hz (0xFFFFFFFF)]

MAX [0 Hz (0x00000000) to 65,535.9999847 Hz (0xFFFFFFFF)]

RANGE M RO
RES [0 Hz (0x00000000) to 65,535.9999847 Hz (0xFFFFFFFF)]

VLIST [0 Hz (0x00000000) to 65,535.9999847 Hz (0xFFFFFFFF)]+

Field Value

CS PE_CENTER_FREQUENCY_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 3 (See Section 5.3.2.6.3, “Layout 3 DataPart”)

5.3.3.7.1.3 QFACTOR CONTROL

The Qfactor Control is used to manipulate the range of frequencies affected around the center frequency of the
filter implemented on an outgoing channel.

Note: The Q-factor of a filter is defined as the ratio of the center frequency to the bandwidth measured at the
-3 dB point. The result of a Q setting of 10 for a filter set to 1000 Hz is a bandwidth of 100 Hz.
Likewise, a center frequency of 5325 Hz and a Q setting of 7.25 results in a bandwidth of
734.48275862 Hz.

Release 4.0 April 2023 168

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-44: Qfactor Control Characteristics

Qfactor SL RW OPT

CUR M RoW

NEXT O mRW

MIN [0 (0x00000000) to 65,535.9999847 (0xFFFFFFFF)]

MAX [0 (0x00000000) to 65,535.9999847 (0xFFFFFFFF)]

RANGE M RO
RES [0 (0x00000000) to 65,535.9999847 (0xFFFFFFFF)]

VLIST [0 (0x00000000) to 65,535.9999847 (0xFFFFFFFF)]+

Field Value

CS PE_QFACTOR_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 3 (See Section 5.3.2.6.3, “Layout 3 DataPart”)

5.3.3.7.1.4 GAIN CONTROL

The Gain Control is used to manipulate the amount of gain or attenuation at the center frequency of the filter
implemented on an outgoing channel.
Table 5-45: Gain Control Characteristics

Gain SL RW OPT

CUR M RoW

NEXT O mRW

MIN [-128.0000 dB (0x8000) to +127.9961 dB (0x7FFF)]

MAX [-128.0000 dB (0x8000) to +127.9961 dB (0x7FFF)]

RANGE M RO
RES [0 dB (0x0000) to +255.9961 dB (0xFFFF)]

VLIST [-128.0000 dB (0x8000) to +127.9961 dB (0x7FFF)]+

Field Value

CS PE_GAIN_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 2 (See Section 5.3.2.6.2, “Layout 2 DataPart”)

5.3.3.7.1.5 LATENCY CONTROL

See Section 5.3.3.1.4, “Latency Control” for a detailed description. XX shall be replaced by PE.

5.3.3.7.2 REVERBERATION EFFECT UNIT

At least one of the optional AudioControls shall be implemented.

5.3.3.7.2.1 BYPASS CONTROL

See Section 5.3.3.1.1, “Bypass Control” for a detailed description. XX shall be replaced by RV.

Release 4.0 April 2023 169

Universal Serial Bus Device Class Definition for Audio Devices

5.3.3.7.2.2 TYPE CONTROL

The Type Control is a macro AudioControl that selects among potentially different algorithms and internal
parameter settings to obtain the desired reverberation effect on an outgoing channel. The CUR and NEXT range for
this AudioControl is from one to the number of supported Reverberation Types, as reflected in the Entity’s
bNrTypes Descriptor field.
Table 5-46: Type Control Characteristics

Type SL RW OPT

CUR M RoW [1 to bNrTypes]

NEXT O mRW [1 to bNrTypes]

MIN N/A

MAX N/A
RANGE P N/A
RES N/A

VLIST N/A

Field Value

CS RV_TYPE_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 1 (See Section 5.3.2.6.1, “Layout 1 DataPart”)

5.3.3.7.2.3 LEVEL CONTROL

The Level Control is used to set the amount of reverberant sound introduced on an outgoing channel, expressed as
a percentage, compared to the level of the original signal.
Table 5-47: Level Control Characteristics

Level SL RW OPT

CUR M RoW

NEXT O mRW

MIN [0 % (0x00) to 255 % (0xFF)]

MAX [0 % (0x00) to 255 % (0xFF)]

RANGE M RO
RES [0 % (0x00) to 255 % (0xFF)]

VLIST [0 % (0x00) to 255 % (0xFF)]+

Field Value

CS RV_LEVEL_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 1 (See Section 5.3.2.6.1, “Layout 1 DataPart”)

5.3.3.7.2.4 TIME CONTROL

The Time Control is used to set the time over which the reverberation, introduced on an outgoing channel, will
continue

Release 4.0 April 2023 170

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-48: Time Control Characteristics

Time SL RW OPT

CUR M RoW

NEXT O mRW

MIN [0 s (0x0000) to 255.9961 s (0xFFFF)]

MAX [0 s (0x0000) to 255.9961 s (0xFFFF)]

RANGE M RO
RES [0 s (0x0000) to 255.9961 s (0xFFFF)]

VLIST [0 s (0x0000) to 255.9961 s (0xFFFF)]+

Field Value

CS RV_TIME_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 2 (See Section 5.3.2.6.2, “Layout 2 DataPart”)

5.3.3.7.2.5 DELAY FEEDBACK CONTROL

The Delay Feedback Control is used when the reverb type is set to Type 6 (Delay) or Type 7 (Panning Delay). It is
ignored for all other Reverb Types. It sets the way in which delay repeats on an outgoing channel.
Table 5-49: Feedback Control Characteristics

Feedback SL RW OPT

CUR M RoW

NEXT O mRW

MIN [0 % (0x00) to 255 % (0xFF)]

MAX [0 % (0x00) to 255 % (0xFF)]

RANGE M RO
RES [0 % (0x00) to 255 % (0xFF)]

VLIST [0 % (0x00) to 255 % (0xFF)]+

Field Value

CS RV_FEEDBACK_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 1 (See Section 5.3.2.6.1, “Layout 1 DataPart”)

5.3.3.7.2.6 PRE-DELAY CONTROL

The Pre-Delay Control is used to set the delay time between the original source and the initial reflection in the
reverberation, introduced on an outgoing channel. The Pre-Delay Control shall support the CUR and RANGE(MIN,
MAX, RES) Attributes. The settings for the CUR, MIN, MAX, and RES Attributes may range from 0 ms (0x0000) to
65535 ms (0xFFFF) in steps of 1 ms (0x0001).

The Control Selector field shall be set to RV_PREDELAY_CONTROL and the Channel Number field indicates the
desired Channel.

Release 4.0 April 2023 171

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-50: Pre-Delay Control Characteristics

Pre-Delay SL RW OPT

CUR M RoW

NEXT O mRW

MIN [0 ms (0x0000) to 65,535 ms (0xFFFF)]

MAX [0 ms (0x0000) to 65,535 ms (0xFFFF)]

RANGE M RO
RES [0 ms (0x0000) to 65,535 ms (0xFFFF)]

VLIST [0 ms (0x0000) to 65,535 ms (0xFFFF)]+

Field Value

CS RV_PREDELAY_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 2 (See Section 5.3.2.6.2, “Layout 2 DataPart”)

5.3.3.7.2.7 DENSITY CONTROL

The Density Control is used to set the density of reflections in the reverberant sound introduced on an outgoing
channel.
Table 5-51: Density Control Characteristics

Density SL RW OPT

CUR M RoW

NEXT O mRW

MIN [0 (0x00) to 100 % (0x64)]

MAX [0 (0x00) to 100 % (0x64)]

RANGE M RO
RES [0 (0x00) to 100 % (0x64)]

VLIST [0 (0x00) to 100 % (0x64)]+

Field Value

CS RV_DENSITY_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 1 (See Section 5.3.2.6.1, “Layout 1 DataPart”)

5.3.3.7.2.8 HI-FREQ ROLL-OFF CONTROL

The Hi-Freq Roll-Off Control is used to set the frequency of a low pass filter on the reverberant sound introduced
on an outgoing channel.

Release 4.0 April 2023 172

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-52: Hi-Freq Roll-Off Control Characteristics

Hi-Freq Roll-Off SL RW OPT

CUR M RoW

NEXT O mRW

MIN [0 Hz (0x00000000) to 65,535.9999847 Hz (0xFFFFFFFF)]

MAX [0 Hz (0x00000000) to 65,535.9999847 Hz (0xFFFFFFFF)]

RANGE M RO
RES [0 Hz (0x00000000) to 65,535.9999847 Hz (0xFFFFFFFF)]

VLIST [0 Hz (0x00000000) to 65,535.9999847 Hz (0xFFFFFFFF)]+

Field Value

CS PE_HIFREQ_ROLLOFF_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 3 (See Section 5.3.2.6.3, “Layout 3 DataPart”)

5.3.3.7.2.9 LATENCY CONTROL

See Section 5.3.3.1.4, “Latency Control” for a detailed description. XX shall be replaced by RV.

5.3.3.7.3 MODULATION DELAY EFFECT UNIT

5.3.3.7.3.1 BYPASS CONTROL

See Section 5.3.3.1.1, “Bypass Control” for a detailed description. XX shall be replaced by MD.

5.3.3.7.3.2 BALANCE CONTROL

The Balance Control is used to set the amount of effect sound introduced on an outgoing channel. A setting of -
100 % results in 100 % of original signal and 0 % of effected signal. A setting of 0 % results in equal amounts of
original signal and effected signal. A setting of +100 % results in 0 % of original signal and 100 % of effected signal.

Release 4.0 April 2023 173

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-53: Balance Control Characteristics

Balance SL RW OPT

CUR M RoW

NEXT O mRW

MIN [-100 % (0x9C) to 100 % (0x64)]

MAX [-100 % (0x9C) to 100 % (0x64)]

RANGE M RO
RES [-100 % (0x9C) to 100 % (0x64)]

VLIST [-100 % (0x9C) to 100 % (0x64)]+

Field Value

CS MD_BALANCE_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 1 (See Section 5.3.2.6.1, “Layout 1 DataPart”)

5.3.3.7.3.3 RATE CONTROL

The Rate Control is used to set the speed (frequency) of the modulator of the delay time introduced on an
outgoing channel. Higher values result in faster modulation.
Table 5-54: Rate Control Characteristics

Rate SL RW OPT

CUR M RoW

NEXT O mRW

MIN [0 Hz (0x0000) to 255.9961 Hz (0xFFFF)]

MAX [0 Hz (0x0000) to 255.9961 Hz (0xFFFF)]

RANGE M RO
RES [0 Hz (0x0000) to 255.9961 Hz (0xFFFF)]

VLIST [0 Hz (0x0000) to 255.9961 Hz (0xFFFF)]+

Field Value

CS MD_RATE_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 2 (See Section 5.3.2.6.2, “Layout 2 DataPart”)

5.3.3.7.3.4 DEPTH CONTROL

The Depth Control is used to set the depth at which the effect sound introduced on an outgoing channel is
modulated. Higher values result in deeper modulation.

Release 4.0 April 2023 174

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-55: Depth Control Characteristics

Depth SL RW OPT

CUR M RoW

NEXT O mRW

MIN [0 ms (0x0000) to 255.9961 ms (0xFFFF)]

MAX [0 ms (0x0000) to 255.9961 ms (0xFFFF)]

RANGE M RO
RES [0 ms (0x0000) to 255.9961 ms (0xFFFF)]

VLIST [0 ms (0x0000) to 255.9961 ms (0xFFFF)]

Field Value

CS MD_DEPTH_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 2 (See Section 5.3.2.6.2, “Layout 2 DataPart”)

5.3.3.7.3.5 TIME CONTROL

The Time Control is used to set the length of the delay time introduced on an outgoing channel. Higher values
result in longer times.
Table 5-56: Time Control Characteristics

Time SL RW OPT

CUR M RoW

NEXT O mRW

MIN [0 ms (0x0000) to 255.9961 ms (0xFFFF)]

MAX [0 ms (0x0000) to 255.9961 ms (0xFFFF)]

RANGE M RO
RES [0 ms (0x0000) to 255.9961 ms (0xFFFF)]

VLIST [0 ms (0x0000) to 255.9961 ms (0xFFFF)]+

Field Value

CS MD_TIME_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 2 (See Section 5.3.2.6.2, “Layout 2 DataPart”)

5.3.3.7.3.6 FEEDBACK LEVEL CONTROL

The Feedback Level Control is used to set the level at which the effected (delayed) sound introduced on an
outgoing channel is mixed back into its own input. Higher values result in higher level of feedback, thereby
resulting in more audible delay repeats.

Release 4.0 April 2023 175

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-57: Feedback Control Characteristics

Feedback SL RW OPT

CUR M RoW

NEXT O mRW

MIN [0 % (0x00) to 255 % (0xFF)]

MAX [0 % (0x00) to 255 % (0xFF)]

RANGE M RO
RES [0 % (0x00) to 255 % (0xFF)]

VLIST [0 % (0x00) to 255 % (0xFF)]+

Field Value

CS MD_FEEDBACK_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 1 (See Section 5.3.2.6.1, “Layout 1 DataPart”)

5.3.3.7.3.7 LATENCY CONTROL

See Section 5.3.3.1.4, “Latency Control” for a detailed description. XX shall be replaced by MD.

5.3.3.7.4 DYNAMIC RANGE COMPRESSOR/EXPANDER EFFECT UNIT

5.3.3.7.4.1 BYPASS CONTROL

See Section 5.3.3.1.1, “Bypass Control” for a detailed description. XX shall be replaced by DR.

5.3.3.7.4.2 RATIO CONTROL

The Ratio Control is used to influence the slope of the active part of the static input-to-output transfer
characteristic of the DRC on an outgoing channel.

Release 4.0 April 2023 176

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-58: Compression Ratio Control Characteristics

Compression
SL RW OPT
Ratio

CUR M RoW

NEXT O mRW

MIN [0 (0x0000) to 255.9961 (0xFFFF)]

MAX [0 (0x0000) to 255.9961 (0xFFFF)]

RANGE M RO
RES [0 (0x0000) to 255.9961 (0xFFFF)]

VLIST [0 (0x0000) to 255.9961 (0xFFFF)]+

Field Value

CS DR _RATIO_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 2 (See Section 5.3.2.6.2, “Layout 2 DataPart”)

5.3.3.7.4.3 MAXAMPL CONTROL

The MaxAmpl Control is used to set the upper boundary of the active input range of the DRC on an outgoing
channel.
Table 5-59: MaxAmpl Control Characteristics

MaxAmpl SL RW OPT

CUR M RoW

NEXT O mRW

MIN [-128.0000 dB (0x8000) to +127.9961 dB (0x7FFF)]

MAX [-128.0000 dB (0x8000) to +127.9961 dB (0x7FFF)]

RANGE M RO
RES [0 dB (0x0000) to +255.9961 dB (0xFFFF)]

VLIST [-128.0000 dB (0x8000) to +127.9961 dB (0x7FFF)]+

Field Value

CS DR_MAXAMPL_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 2 (See Section 5.3.2.6.2, “Layout 2 DataPart”)

5.3.3.7.4.4 THRESHOLD CONTROL

The Threshold Control is used to set the lower boundary of the active input range of the effect on an outgoing
channel.

Release 4.0 April 2023 177

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-60: Threshold Control Characteristics

Threshold SL RW OPT

CUR M RoW

NEXT O mRW

MIN [-128.0000 dB (0x8000) to +127.9961 dB (0x7FFF)]

MAX [-128.0000 dB (0x8000) to +127.9961 dB (0x7FFF)]

RANGE M RO
RES [0 dB (0x0000) to +255.9961 dB (0xFFFF)]

VLIST [-128.0000 dB (0x8000) to +127.9961 dB (0x7FFF)]+

Field Value

CS DR_THRESHOLD_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 2 (See Section 5.3.2.6.2, “Layout 2 DataPart”)

5.3.3.7.4.5 ATTACK TIME CONTROL

The Attack Time Control is used to determine the response of the DRC on an outgoing channel to a step increase in
the signal level on the corresponding input channel.
Table 5-61: Attack Time Control Characteristics

Attack Time SL RW OPT

CUR M RoW

NEXT O mRW

MIN [0 ms (0x0000) to 255.9961 ms (0xFFFF)]

MAX [0 ms (0x0000) to 255.9961 ms (0xFFFF)]

RANGE M RO
RES [0 ms (0x0000) to 255.9961 ms (0xFFFF)]

VLIST [0 ms (0x0000) to 255.9961 ms (0xFFFF)]+

Field Value

CS DR_ATTACK_TIME_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 2 (See Section 5.3.2.6.2, “Layout 2 DataPart”)

5.3.3.7.4.6 RELEASE TIME CONTROL

The Release Time Control is used to determine the recovery response of the DRC on an outgoing channel to a step
decrease in the signal level on the corresponding input channel.

Release 4.0 April 2023 178

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-62: Release Time Control Characteristics

Release Time SL RW OPT

CUR M RoW

NEXT O mRW

MIN [0 ms (0x0000) to 255.9961 ms (0xFFFF)]

MAX [0 ms (0x0000) to 255.9961 ms (0xFFFF)]

RANGE M RO
RES [0 ms (0x0000) to 255.9961 ms (0xFFFF)]

VLIST [0 ms (0x0000) to 255.9961 ms (0xFFFF)]+

Field Value

CS DR_RELEASE_TIME_CONTROL

OCN:ICN:IPN OCN = ICN = The Pin channel number on which the AudioControl resides (including the Primary channel 0); IPN = 1

Parameter Block Layout 2 (See Section 5.3.2.6.2, “Layout 2 DataPart”)

5.3.3.7.4.7 LATENCY CONTROL

See Section 5.3.3.1.4, “Latency Control” for a detailed description. XX shall be replaced by DR.

5.3.3.8 PROCESSING UNIT COMMAND

This Command is used to manipulate the AudioControls inside the Processing Unit. The exact layout of the
Command is defined in Section 5.3.2.5, “AudioControl Command DataPart Layout.”

The following paragraphs present a detailed description of all possible AudioControls the Processing Unit may
incorporate. For each AudioControl, the supported Attributes and their value ranges are specified. Also, the
appropriate AudioControl Selector value and the layout type of the Parameter Blocks are listed. The Control
Selector codes are defined in Appendix A.23.7, “Processing Unit Control Selectors.”

5.3.3.8.1 UP/DOWN-MIX PROCESSING UNIT

5.3.3.8.1.1 BYPASS CONTROL

See Section 5.3.3.1.1, “Bypass Control” for a detailed description. XX shall be replaced by UD.

5.3.3.8.1.2 CLUSTER CONTROL

The Cluster Control is used to indirectly change the behavior of the Up/Down-mix Processing Unit. The operational
mode of the Up/Down-mix Processing Unit is changed so that the outgoing Cluster matches the configuration as
indicated by the Cluster Control. See Section 5.3.3.1.2, “Cluster Control” for a detailed description. XX shall be
replaced by UD.

5.3.3.8.1.3 CLUSTER ACTIVE CONTROL

This AudioControl shall always be present. See Section 5.3.3.1.3, “Cluster Active Control” for a detailed description.
XX shall be replaced by UD.

5.3.3.8.1.4 LATENCY CONTROL

See Section 5.3.3.1.4, “Latency Control” for a detailed description. XX shall be replaced by UD.

Release 4.0 April 2023 179

Universal Serial Bus Device Class Definition for Audio Devices

5.3.3.8.2 CHANNEL REMAP PROCESSING UNIT

5.3.3.8.2.1 BYPASS CONTROL

See Section 5.3.3.1.1, “Bypass Control” for a detailed description. XX shall be replaced by CR.

5.3.3.8.2.2 CLUSTER CONTROL

The Cluster Control is used to indirectly change the behavior of the Channel Remap Processing Unit. The
remapping mode of the Channel Remap Processing Unit is changed so that the outgoing Cluster matches the
configuration as indicated by the Cluster Control. See Section 5.3.3.1.2, “Cluster Control” for a detailed description.
XX shall be replaced by CR.

5.3.3.8.2.3 CLUSTER ACTIVE CONTROL

This AudioControl shall always be present. See Section 5.3.3.1.3, “Cluster Active Control” for a detailed description.
XX shall be replaced by CR.

5.3.3.8.2.4 LATENCY CONTROL

See Section 5.3.3.1.4, “Latency Control” for a detailed description. XX shall be replaced by CR.

5.3.3.8.3 STEREO EXTENDER PROCESSING UNIT

5.3.3.8.3.1 BYPASS CONTROL

See Section 5.3.3.1.1, “Bypass Control” for a detailed description. XX shall be replaced by ST.

5.3.3.8.3.2 WIDTH CONTROL

The Width Control is used to change the spatial appearance of the stereo image, produced by the Stereo Extender
Processing Unit.
Table 5-63: Width Control Characteristics

Width SL RW MAN

CUR M RoW

NEXT O mRW

MIN [0 (0x00) to 255 (0xFF)]

MAX [0 (0x00) to 255 (0xFF)]

RANGE M RO
RES [0 (0x00) to 255 (0xFF)]

VLIST [0 (0x00) to 255 (0xFF)]+

Field Value

CS ST_WIDTH_CONTROL

OCN:ICN:IPN 0:0:0

Parameter Block Layout 1 (See Section 5.3.2.6.1, “Layout 1 DataPart”)

5.3.3.8.3.3 LATENCY CONTROL

See Section 5.3.3.1.4, “Latency Control” for a detailed description. XX shall be replaced by ST.

Release 4.0 April 2023 180

Universal Serial Bus Device Class Definition for Audio Devices

5.3.3.8.4 MULTI-FUNCTION PROCESSING UNIT

5.3.3.8.4.1 BYPASS CONTROL

The Bypass Control shall be implemented for a Multi-Function Processing Unit. See Section 5.3.3.1.1, “Bypass
Control” for a detailed description. XX shall be replaced by MF.

5.3.3.8.4.2 CLUSTER CONTROL

See Section 5.3.3.1.2, “Cluster Control” for a detailed description. XX shall be replaced by MF.

5.3.3.8.4.3 CLUSTER ACTIVE CONTROL

This AudioControl shall always be present. See Section 5.3.3.1.3, “Cluster Active Control” for a detailed description.
XX shall be replaced by MF.

5.3.3.8.4.4 ALGO PRESENT CONTROL

The Algo Present Control is a mandatory Read-Only AudioControl that indicates which of the predefined algorithms
are currently supported. The CUR Attribute returns a 32-bit AlgorithmsPresent bitmap, indicating which of the
predefined algorithms are currently available. Multiple bits may be set simultaneously. The CUR Attribute value
may change dynamically and generate an interrupt. The bit allocation in the AlgorithmsPresent bitmap is as
follows:

• D0: Beam Forming.

• D1: Acoustic Echo Cancellation.
• D2: Active Noise Cancellation.
• D3: Blind Source Separation.
• D4: Noise Suppression/Reduction.
• D23..5: Reserved.
• D31..24: Vendor-defined Algorithms.

Release 4.0 April 2023 181

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-64: Algo Present Control Characteristics

Algo Present SL RW MAN

CUR M RO

NEXT P N/A

MIN N/A

MAX N/A
RANGE I N/A
RES N/A

VLIST N/A

Field Value

CS MF_ALGO_PRESENT_CONTROL

OCN:ICN:IPN 0:0:0

Parameter Block Layout 3 (See Section 5.3.2.6.3, “Layout 3 DataPart”)

5.3.3.8.4.5 ALGO ENABLE CONTROL

The Algo Enable Control provides a means to selectively enable or disable the implemented algorithms. The CUR
Attribute is a 32-bit EnableAlgorithms bitmap, where each bit set to one enables the corresponding algorithm. The
bit allocations are identical to those of the AlgorthmsPresent bitmap as defined in Section 5.3.3.8.4.4, “Algo
Present Control” above.
Table 5-65: Enable Algo Control Characteristics

Enable Algo SL RW OPT

CUR M mRW 32-bit EnableAlgorithms bitmap

NEXT O mRW

MIN N/A

MAX N/A
RANGE I N/A
RES N/A

VLIST N/A

Field Value

CS MF_ALGO_ENABLE_CONTROL

OCN:ICN:IPN 0:0:0

Parameter Block Layout 3 (See Section 5.3.2.6.3, “Layout 3 DataPart”)

5.3.3.8.4.6 LATENCY CONTROL

See Section 5.3.3.1.4, “Latency Control” for a detailed description. XX shall be replaced by MF.

5.3.3.9 EXTENSION UNIT COMMAND

Because this specification has no knowledge about the inner workings of the Extension Unit, it is impossible to
define a Command that can manipulate specific Extension Unit AudioControls. However, this specification defines
several AudioControls the Extension Unit shall or may support.

This Command is used to manipulate the AudioControls inside the Extension Unit. The exact layout of the
Command is defined in Section 5.3.2.5, “AudioControl Command DataPart Layout.”

Release 4.0 April 2023 182

Universal Serial Bus Device Class Definition for Audio Devices

The following paragraphs present a detailed description of all possible AudioControls the Extension Unit may
incorporate. For each AudioControl, the supported Attributes and their value ranges are specified. Also, the
appropriate Control Selector value and the layout type of the Parameter Blocks are listed. Issuing non-supported
Control Selectors to the Extension Unit leads to a Request Error on the Set request part of the Command. The
Control Selector codes are defined in Appendix A.23.8, “Extension Unit Control Selectors.”

5.3.3.9.1 BYPASS CONTROL

The Bypass Control shall be implemented for an Extension Unit. See Section 5.3.3.1.1, “Bypass Control” for a
detailed description. XX shall be replaced by XU.

5.3.3.9.2 CLUSTER CONTROL

If implemented, the Extension Unit shall always implement the Cluster Control as Read (r). Vendor defined events
within the Extension Unit may cause the Cluster Control to change its value and generate an interrupt, informing
the Audio Class driver of the change. See Section 5.3.3.1.2, “Cluster Control” for a detailed description. XX shall be
replaced by XU.

5.3.3.9.3 CLUSTER ACTIVE CONTROL

This Control shall always be present. See Section 5.3.3.1.3, “Cluster Active Control” for a detailed description. XX
shall be replaced by XU.

5.3.3.9.4 LATENCY CONTROL

See Section 5.3.3.1.4, “Latency Control” for a detailed description. XX shall be replaced by XU.

5.3.3.10 CLOCK SOURCE COMMAND

This Command is used to manipulate the AudioControls inside a Clock Source Entity of the Audio Function. The
exact layout of the Command is defined in Section 5.3.2.5, “AudioControl Command DataPart Layout.”

The following paragraphs present a detailed description of all possible AudioControls a Clock Source Entity may
incorporate. For each AudioControl, the supported Attributes and their value ranges are specified. Also, the
appropriate Control Selector value and the layout type of the Parameter Blocks are listed. The Control Selector
codes are defined in Appendix A.23.9, “Clock Source Control Selectors.”

5.3.3.10.1 SAMPLING FREQUENCY CONTROL

The Sampling Frequency Control is used to manipulate the actual sampling frequency of the clock signal that is
generated by the Clock Source Entity.

For predictable results, this AudioControl should not be written while a related audiostream is active.

Release 4.0 April 2023 183

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-66: Sampling Frequency Control Characteristics

Sampling
SL RW MAN
Frequency

CUR M RoW*

NEXT O/P* mRW

MIN [0 Hz (0x00000000) to 4,294,967,295 Hz (0xFFFFFFFF)]

MAX [0 Hz (0x00000000) to 4,294,967,295 Hz (0xFFFFFFFF)]

RANGE M RO
RES [0 Hz (0x00000000) to 4,294,967,295 Hz (0xFFFFFFFF)]

VLIST [0 Hz (0x00000000) to 4,294,967,295 Hz (0xFFFFFFFF)]+

Field Value

CS CS_SAM_FREQ_CONTROL

OCN:ICN:IPN 0:0:0

Parameter Block Layout 3 (See Section 5.3.2.6.3, “Layout 3 DataPart”)

* In many cases, the Clock Source Entity represents a crystal oscillator-based generator with a single fixed frequency. In some other cases, the Clock
Source entity represents an external clock which cannot be controlled by the Audio Function hardware. In all those cases, the CUR Attribute shall be
implemented as Read and the NEXT Attribute shall not be supported. However, the RANGE Attribute shall always be supported to indicate the
possible values the CUR Attribute may assume.

5.3.3.10.2 CLOCK VALID CONTROL

The Clock Valid Control is used to indicate if the clock signal that is generated by the Clock Source Entity is valid
(stable and reliable).
Table 5-67: Status Control Characteristics

Clock Valid SL RW MAN

CUR M RO [0 (0x00) to 1 (0x01)]; Invalid Clock or Valid Clock respectively.

NEXT P N/A

MIN N/A

MAX N/A
RANGE I N/A
RES N/A

VLIST N/A

Field Value

CS CS_CLOCK_VALID_CONTROL

OCN:ICN:IPN 0:0:0

Parameter Block Layout 1 (See Section 5.3.2.6.1, “Layout 1 DataPart”)

5.3.3.11 CLOCK SELECTOR COMMAND

This Command is used to manipulate the AudioControls inside a Clock Selector Entity of the Audio Function. The
exact layout of the Command is defined in Section 5.3.2.5, “AudioControl Command DataPart Layout.”

The following paragraphs present a detailed description of all possible AudioControls a Clock Selector Entity may
incorporate. For each AudioControl, the supported Attributes and their value ranges are specified. Also, the
appropriate Control Selector value and the layout type of the Parameter Blocks are listed. The Control Selector
codes are defined in Appendix A.23.10,”Clock Selector Control Selectors.”

Release 4.0 April 2023 184

Universal Serial Bus Device Class Definition for Audio Devices

5.3.3.11.1 SELECTOR CONTROL

The Selector Control allows selecting among several clock signals.

The Selector Control provides the capability to select a clock signal from several incoming clocks on the Clock
Selector Unit’s Input Pins. The range for this AudioControl is from 0 to the number of Input Pins the Clock Selector
Entity supports, as reflected in the Entity’s bNrInputPins Descriptor field.
Table 5-68: Clock Selector Control Characteristics

Clock Selector SL RW MAN

CUR M RoW [1 to bNrInputPins]

NEXT O mRW

MIN N/A

MAX N/A
RANGE N/A N/A
RES N/A

VLIST N/A

Field Value

CS CX_CLOCK_SELECTOR_CONTROL

OCN:ICN:IPN 0:0:0

Parameter Block Layout 1 (See Section 5.3.2.6.1, “Layout 1 DataPart”)

5.3.3.12 CONNECTOR ENTITY COMMAND

This Command is used to manipulate the AudioControls inside a Connector Entity of the Audio Function. The exact
layout of the Command is defined in Section 5.3.2.5, “AudioControl Command DataPart Layout.”

The following paragraphs present a detailed description of all possible AudioControls a Connector Entity may
incorporate. For each AudioControl, the supported Attributes and their value ranges are specified. Also, the
appropriate Control Selector value and the layout type of the Parameter Blocks are listed. The Control Selector
codes are defined in Appendix A.23.11, “Connector Control Selectors.”

5.3.3.12.1 INSERTION DETECT CONTROL

If implemented, the Insertion Detect Control shall be supported as Read-Only. It is used to report whether a
(compatible) plug has been inserted into the receptacle of the Connector. A value of zero indicates that nothing
has been inserted. A value of one indicates that a plug has been inserted.

Release 4.0 April 2023 185

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-69: Insertion Detect Control Characteristics

Insertion Detect SL RW OPT

CUR M RO [0 (0x00) to 1 (0x01)]; Nothing Inserted or Plug Inserted respectively

NEXT P N/A

MIN N/A

MAX N/A
RANGE I N/A
RES N/A

VLIST N/A

Field Value

CS CO_INSERTION_CONTROL

OCN:ICN:IPN 0:0:0

Parameter Block Layout 1 (See Section 5.3.2.6.1, “Layout 1 DataPart”)

5.3.3.13 POWER DOMAIN COMMAND

This Command is used to manipulate the AudioControls inside a Power Domain of the Audio Function. The exact
layout of the Command is defined in Section 5.3.2.5, “AudioControl Command DataPart Layout.”

The following paragraphs present a detailed description of all possible AudioControls a Power Domain may
incorporate. For each AudioControl, the supported Attributes and their value ranges are specified. Also, the
appropriate Control Selector value and the layout type of the Parameter Blocks are listed. The Control Selector
codes are defined in Appendix A.23.12, “Power Domain Control Selectors.”

5.3.3.13.1 POWER STATE CONTROL

This AudioControl shall be supported when the Audio Function advertises the existence of a Power Domain
through the presence of a Power Domain Entity Descriptor. (It is therefore not necessary to advertise the existence
of the AudioControl via the regular bmOptControls mechanism.)

The Power State Control is used to selectively bring parts of the Audio Function (a Power Domain) into different
Power States and to report their current Power State.

Release 4.0 April 2023 186

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-70: Power State Control Characteristics

Power State SL RW MAN

CUR M mRW [0 (0x00) to 4 (0x04)]

NEXT O mRW

MIN N/A

MAX N/A
RANGE I N/A
RES N/A

VLIST N/A

Field Value

CS PD _POWER_STATE_CONTROL

OCN:ICN:IPN 0:0:0

Parameter Block Layout 1 (See Section 5.3.2.6.1, “Layout 1 DataPart”)

5.3.4 AUDIOSTREAMING COMMANDS

The following sections describe the Commands an Audio Function may support for its AudioStreaming Interfaces.
The same layout of the Parameter Blocks is used for both the Set and Pull Commands.

AudioStreaming Commands may be directed either to the AudioStreaming Interface or to the associated
isochronous data Endpoint, depending on the location of the AudioControl to be manipulated.

5.3.4.1 INTERFACE AUDIOCONTROL COMMAND

This Command is used to manipulate the AudioControls inside an AudioStreaming Interface of the Audio Function.
The exact layout of the Command is defined in Section 5.3.2.5, “AudioControl Command DataPart Layout.”

The following paragraphs present a detailed description of all possible AudioControls an AudioStreaming Interface
may incorporate. For each AudioControl, the supported Attributes and their value ranges are specified. Also, the
appropriate Control Selector value and the layout type of the Parameter Blocks are listed. The Control Selector
codes are defined in Appendix A.23.13, “AudioStreaming Interface Control Selectors.”

5.3.4.1.1 ACTIVE ALTERNATE SETTING CONTROL

This AudioControl is used to inform Host software which Alternate Setting of an AudioStreaming Interface is
currently active. The main purpose of this AudioControl is to notify the Host (through an interrupt) that the last
selected Alternate Setting is no longer valid. There may be a variety of reasons why this could happen. For
instance, increasing the sampling frequency of a Clock Source Entity may render the current Alternate Setting of an
interface connected to that Clock Source invalid because more bandwidth is now needed than is available in the
current Alternate Setting.

This specification does not allow an interface to change from one active Alternate Setting to another without Host
intervention. Whenever an Alternate Setting becomes invalid, the interface is required to switch to (idle) Alternate
Setting zero. If this situation may occur in the Audio Function, this AudioControl (and the Valid Alternate Settings
Control) shall be present. It always provides the currently active Alternate Setting for the interface. The Host
software needs to then take appropriate action to reactivate the interface by switching to a valid Alternate Setting.
The value of an Active Alternate Setting Control CUR Attribute shall only be either the last set Alternate Setting or
zero.

Release 4.0 April 2023 187

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-71: Active Alternate Setting Control Characteristics

Active Alternate
SL RW OPT
Setting

CUR M RO Current Active Alternate Setting: 0 or lastly set Alternate Setting.

NEXT P N/A

MIN N/A

MAX N/A
RANGE I N/A
RES N/A

VLIST N/A

Field Value

CS AS_ACTIVE_ALT_SETTING_CONTROL

OCN:ICN:IPN 0:0:0

Parameter Block Layout 1 (See Section 5.3.2.6.1, “Layout 1 DataPart”)

5.3.4.1.2 VALID ALTERNATE SETTINGS CONTROL

This AudioControl is used to inform Host software what the currently possible valid Alternate Settings are for an
AudioStreaming Interface. If the Active Alternate Setting Control is present, then this AudioControl shall also be
present. It always provides a list of currently valid Alternate Settings for the interface. The value of a Valid
Alternate Setting Control CUR Attribute is a bitmap, returned in the bmValidAltSettings field, that contains a bit
for each possible active Alternate Setting. The bSize field indicates the length in bytes of the bmValidAltSettings
field.

A bit set means that this Alternate Setting is currently valid. A bit cleared means that this Alternate Setting is
currently not valid. Bit D0 corresponds to Alternate Setting 0 and shall always be set since it is always a possible
valid setting. Bit D1 corresponds to Alternate Setting 1. Bit Dm corresponds to Alternate Setting m. All bits that do
not correspond to an existing Alternate Setting shall be set to 0. An attempt to set the interface to an invalid
Alternate Setting (through the standard Set Interface Command) shall result in a Request Error.

Release 4.0 April 2023 188

Universal Serial Bus Device Class Definition for Audio Devices

Table 5-72: Valid Alternate Settings Control Characteristics

Valid Alternate
SL RW OPT
Settings

CUR M RO

NEXT P N/A

MIN N/A

MAX N/A
RANGE I N/A
RES N/A

VLIST N/A

Field Value

CS AS_VALID_ALT_SETTINGS_CONTROL

OCN:ICN:IPN 0:0:0

Parameter Block Length 1+n

Offset Field Size Value Description

0 bSize 1 Number Indicates the number of bytes in the

bmValidAltSettings bitmap: n

1 bmValidAltSettings n Bitmap Each set bit represents a currently valid Alternate

Setting for the interface.

5.3.4.2 ENDPOINT COMMAND

There is no endpoint-specific Command defined at this time.

5.3.5 ADDITIONAL COMMANDS

5.3.5.1 CLASS-SPECIFIC STRING DESCRIPTOR COMMAND

To overcome the limitations of the standard Get String Descriptor Request (limited to 255 Device-wide strings),
this specification provides a class-specific extension to retrieve a larger set of potentially larger String Descriptors
from the Audio Function.

The Pull CS String Descriptor Command is used to retrieve class-specific String Descriptors and shall be supported if
the Audio Function contains at least one class-specific String Descriptor. These String Descriptors are uniquely
identified within the Audio Function through their wStrDescriptorID and iLangID values.

5.3.5.1.1 CS STRING DESCRIPTOR COMMAND ADDRESSPART LAYOUT

Table 5-73: CS String Descriptor Command AddressPart Layout

Offset Field Size Value Description

0 wStrDescriptorID 2 Number String Descriptor ID of the targeted CS
String Descriptor.
2 iLang 2 Number iLang ID for the String Descriptor.
4 wAttribute 2 Constant STRING Attribute.
6 Reserved 2 Number Not used. Shall be set to zero.
8 Reserved 2 Number Not used. Shall be set to zero.

Release 4.0 April 2023 189

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

10 Reserved 2 Number Not used. Shall be set to zero.

The wStrDescriptorID field specifies the String Descriptor’s Descriptor ID. For maximum interoperability between
this class-specific method to retrieve String Descriptors from the Audio Function and the standard Get String
Descriptor Request, all Audio Function related standard String Descriptors, including the LANGID code array at
index 0, that can be retrieved using the standard Get String Descriptor Request shall also be retrievable using this
class-specific String Descriptor Command by specifying zero in the high byte and the standard String Descriptor
index in the low byte of the wStrDescriptorID value. Whether other than Audio Function related standard String
Descriptors present in the Device can be retrieved using this new method is implementation dependent. Any newly
defined class-specific String Descriptor that uses the wStrDescriptorID value as an index, shall use
wStrDescriptorID values starting from 256 onwards.

The iLangID field specifies the iLang ID for the String Descriptor. It contains a zero-based index into the LANGID
code array as returned by the Device. A Device shall at most support 126 different languages since the LANGID
code array is restricted to 254 bytes and each LANGID code takes up 2 bytes. The range of the iLangID is therefore
from 0 to 125 maximum.

The values specified in the wStrDescriptorID and iLang fields shall be appropriate to the recipient. Only existing
String Descriptor ID values in the Audio Function may be indexed and only appropriate iLang IDs may be used. If
the Command specifies an unknown wStrDescriptorID or iLangID value, the Pull Command Set Request shall
return a Request Error.

The wAttribute field shall contain the STRING constant. If the Audio Function does not support class-specific String
Descriptors, the Pull Command Set Request shall return a Request Error.

All Reserved fields shall be set to zero. If any of the Reserved fields is not zero, the Pull Command Set Request shall
return a Request Error.

5.3.5.1.2 CS STRING DESCRIPTOR COMMAND DATAPART LAYOUT

The length of the Descriptor to return is indicated in the wLength field of the Pull CS String Descriptor Command
Get Request. If the Descriptor is longer than what is indicated in the wLength field, only the initial bytes of the
Descriptor are returned. If the Descriptor is shorter than what is indicated in the wLength field, the Device
indicates the end of the control transfer by sending a short packet when further data is requested.
The layout of the Parameter Block follows the class-specific String Descriptor definition as outlined in Section 4.9,
“Class-specific String Descriptors.”

5.3.5.2 EXTENDED DESCRIPTOR COMMAND

To overcome the limitations of the standard Get Descriptor Command (limited to maximum 255 bytes long), this
specification provides a class-specific method to retrieve larger Descriptors from the Audio Function. Also, since
Extended Descriptors can report changes dynamically, they may be used whenever there is a need for the
Descriptor to indicate that some of its values have changed (even when its length is less than 256 bytes).

The Pull Extended Descriptor Command is used to retrieve Extended Descriptors and shall be supported whenever
the Audio Function uses at least one Extended Descriptor. An Extended Descriptor is uniquely identified by its
wDescriptorID value and is therefore global to the Audio Function and is always retrieved through the
AudioControl Interface.

Release 4.0 April 2023 190

Universal Serial Bus Device Class Definition for Audio Devices

5.3.5.2.1 EXTENDED DESCRIPTOR COMMAND ADDRESSPART LAYOUT

Table 5-74: Extended Descriptor Command AddressPart Layout

Offset Field Size Value Description

0 wDescriptorID 2 Number Extended Descriptor ID of the targeted
Extended Descriptor.
2 Reserved 2 Number Not used. Shall be set to zero.
4 wAttribute 2 Constant EXTENDED_DESCRIPTOR Attribute.
6 Reserved 2 Number Not used. Shall be set to zero.
8 Reserved 2 Number Not used. Shall be set to zero.
10 Reserved 2 Number Not used. Shall be set to zero.

The wDescriptorID field specifies the Extended Descriptor ID. The value specified in the wDescriptorID field shall
be appropriate to the recipient. Only existing Extended Descriptor ID values in the Audio Function may be used. If
the Command specifies an unknown wDescriptorID, the Pull Command Set Request shall return a Request Error.

The wAttribute field shall contain the EXTENDED_DESCRIPTOR constant. If the Audio Function does not support
Extended Descriptors, the Pull Command Set Request shall return a Request Error.

All Reserved fields shall be set to zero. If any of the Reserved fields is not zero, the Pull Command Set Request shall
return a Request Error.

5.3.5.2.2 EXTENDED DESCRIPTOR COMMAND DATAPART LAYOUT

The length of the Descriptor to return is indicated in the wLength field of the Pull Extended Descriptor Command
Get Request. If the Descriptor is longer than what is indicated in the wLength field, only the initial bytes of the
Descriptor are returned. If the Descriptor is shorter than what is indicated in the wLength field, the Device
indicates the end of the control transfer by sending a short packet when further data is requested.

The layout of the Parameter Block follows the Extended Descriptor definitions as outlined in Section 4.2, “Class-
specific Descriptors.”

5.3.5.3 PAGED EXTENDED DESCRIPTOR COMMAND

To accommodate limited resource Host implementations, this specification supports paged access to the Extended
Descriptors. This Command allows retrieval of an Extended Descriptor in 256-byte pages at a time. Each page start
at a 256-byte boundary and the page numbering is zero-based. A maximum of one page can be retrieved at a time.
Therefore, the maximum allowed value in the wLength field is 256.

The Pull Paged Extended Descriptor Command shall be supported whenever the Audio Function uses at least one
Extended Descriptor.

An Extended Descriptor is uniquely identified by the wDescriptorID value and is therefore global to the Audio
Function and is always retrieved through the AudioControl Interface.

Release 4.0 April 2023 191

Universal Serial Bus Device Class Definition for Audio Devices

5.3.5.3.1 PAGED EXTENDED DESCRIPTOR COMMAND ADDRESSPART LAYOUT

Table 5-75: Paged Extended Descriptor Command AddressPart Layout

Offset Field Size Value Description

0 wDescriptorID 2 Number Extended Descriptor ID of the targeted
Extended Descriptor.
2 wPageNumber 2 Number Not used. Shall be set to zero.
4 wAttribute 2 Constant PAGED_EXTENDED_DESCRIPTOR Attribute.
6 Reserved 2 Number Not used. Shall be set to zero.
8 Reserved 2 Number Not used. Shall be set to zero.
10 Reserved 2 Number Not used. Shall be set to zero.

The wDescriptorID field specifies the Extended Descriptor ID.

The wPageNumber field specifies the zero-based page number of the page to be retrieved.

The values specified in the wDescriptorID and wPageNumber fields shall be appropriate to the recipient. Only
existing Extended Descriptor ID values and Page numbers in the Audio Function may be used. If the Command
specifies an unknown wDescriptorID or a non-existent Page Number, the Pull Command Set Request shall return a
Request Error.

The wAttribute field shall contain the PAGE_EXTENDED_DESCRIPTOR constant. If the Audio Function does not
support Extended Descriptors, the Pull Command Set Request shall return a Request Error.

All Reserved fields shall be set to zero. If any of the Reserved fields is not zero, the Pull Command Set Request shall
return a Request Error.

5.3.5.3.2 PAGED EXTENDED DESCRIPTOR COMMAND DATAPART LAYOUT

The length of the Descriptor to return is indicated in the wLength field of the Pull Paged Extended Descriptor
Command Get Request and shall therefore be equal to or less than 256. If the Descriptor is longer than what is
indicated in the wLength field, only the initial bytes of the Descriptor are returned. If the Descriptor is shorter than
what is indicated in the wLength field, the Device indicates the end of the control transfer by sending a short
packet when further data is requested.

The layout of the Parameter Block follows the Extended Descriptor definitions as outlined in Section 4.2, “Class-
specific Descriptors.”

Release 4.0 April 2023 192

Universal Serial Bus Device Class Definition for Audio Devices

6 INTERRUPTS

Interrupts are used to inform the Host that a change has occurred in the current state of the Audio Function. This
specification currently defines four different Interrupt Source Types:

• AudioControl CUR Attribute Change: An AudioControl inside the Audio Function has changed its CUR
Attribute value (any AudioControl inside an Entity, AudioControl or
AudioStreaming Interface, or any AudioControl related to an audio
Endpoint).
• Extended Descriptor Change: An Extended Descriptor has changed one or more of its fields.
• Class-specific String Descriptor Change: A class-specific String Descriptor has changed.

A change of state in the Audio Function is most often the result of an event, either user-initiated or Device-
initiated. Insertion or removal of a connector is a typical example of a user-initiated event. As a result, the Host
could switch selectors or mixers so as to play audio out of the just inserted device (e.g. a headphone) and stop
playing audio out of the current device (e.g. a speaker set).
An example of a Device-initiated event is an external device (e.g. an A/V receiver) automatically switching from
PCM to AC-3 encoded data on its optical digital output, depending on the material that is currently being played. If
this device is connected to the optical digital input of an Audio Function that has auto-detect capabilities, the
interface on that Audio Function may need to be reconfigured (e.g., to start the AC-3 decoding process), causing
other interfaces to change some aspect of their format, or even become unusable. The Device could issue an
interrupt, letting the Host know that the Audio Function needs reconfiguration.

6.1 INTERRUPT MESSAGE

This specification defines four different Interrupt Message Types, one for each defined Interrupt Source Type:

• AudioControl
• Extended Descriptor
• Class-specific String Descriptor

The Interrupt Message consists of a common Interrupt Message Header, followed by an Interrupt Message Body.
The layout of the Interrupt Message Body depends on the wAttribute field as detailed below.

6.1.1 INTERRUPT MESSAGE HEADER

The interrupt Message Header is always six bytes in length.
The wLength field contains the total length of the Interrupt Message, in bytes.

The wAttribute field shall be set to the Attribute value that caused the interrupt.

The bSourceNumber field shall contain the AudioControl or AudioStreaming interface number of the interface that
contains the source of the interrupt.

The Body field the layout is defined below for the different Interrupt Source Types.

Note: The MaxPacketSize of the interrupt endpoint shall be set such that all Interrupt Messages that the
Device is able to generate fit within the endpoint buffer.

Release 4.0 April 2023 193

Universal Serial Bus Device Class Definition for Audio Devices

Table 6-1: Interrupt Message Header Format

Offset Field Size Value Description

0 wLength 2 Number Size of this Interrupt Message, in bytes:
6+n.
2 wAttribute 2 Number The Attribute that caused the interrupt.
4 bSourceNumber 1 Number The Interface number of the interface from
which the interrupt originates.
5 Reserved 1 Number Not used. Shall be set to zero.
6 Body n N/A Source-specific Interrupt Message Body.

6.1.2 AUDIOCONTROL INTERRUPT MESSAGE BODY

The wAttribute field shall be set to CUR when the interrupt is caused by a change in the CUR Attribute of an
AudioControl. It shall be set to RANGE when the interrupt is caused by a change in the RANGE Attribute of an
AudioControl.

The wEntityID field contains the EntityID of the Entity that contains the AudioControl from which the interrupt
originates. This is only applicable when the bSourceNumber field indicates the AudioControl interface number. The
wEntityID field shall be set to zero otherwise (not used).

The wControlSelector field contains the Control Selector value (CS) of the AudioControl from which the interrupt
originates.

The wOCN field contains the Output Channel Number of the AudioControl from which the interrupt originates.

The wICN field contains the Input Channel Number of the AudioControl from which the interrupt originates.

The wIPN field contains the Input Pin Number of the AudioControl from which the interrupt originates.

The ParamBlock field is only meaningful if the wAttribute field specifies CUR and the AudioControl uses one of the
predefined DataPart Layouts 1, 2, or 3 as defined in Section 5.3.2.5, “AudioControl Command DataPart Layout.” In
this case, the ParamBlock field contains the value of the CUR Attribute of the AudioControl (one, two, or four
bytes).

If the AudioControl uses a custom DataPart Layout, then the ParamBlock field is not present. The Host should
query the AudioControl directly to obtain the AudioControl’s most recent Attribute value(s) and to re-enable
interrupt generation.

See also Section 6.2, “Interrupt Behavior” for the timing details for the value(s) returned in the ParamBlock field.
Table 6-2: AudioControl Interrupt Message Format

Offset Field Size Value Description

0 wLength 2 Number Size of this Interrupt Message, in bytes:
16+n (n = 0|1|2|4).
2 wAttribute 2 Number The AudioControl Attribute from which the
interrupt originates.
4 bSourceNumber 1 Number The interface number of the interface from
which the interrupt originates.
5 Reserved 1 Number Not used. Shall be set to zero.

Release 4.0 April 2023 194

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

6 wEntityID 2 Number ID of the Entity that contains the
AudioControl from which the interrupt
originates.
8 wControlSelector 2 Number The Control Selector value of the
AudioControl from which the interrupt
originates.
10 wOCN 2 Number The OCN of the AudioControl from which
the interrupt originates.
12 wICN 2 Number The ICN of the AudioControl from which the
interrupt originates.
14 wIPN 2 Number The IPN of the AudioControl from which the
interrupt originates.
16 DataPart n Number The DataPart for the AudioControl, if
present.

6.1.3 EXTENDED DESCRIPTOR INTERRUPT MESSAGE BODY

The wAttribute field shall be set to EXTENDED_DESCRIPTOR.

The wDescriptorID field contains the Descriptor ID of the Extended Descriptor from which the interrupt originates.

The Host should issue a Pull Extended Descriptor Command to obtain the most recent version of the Descriptor
and to re-enable interrupt generation.
Table 6-3: Extended Descriptor Interrupt Message Format

Offset Field Size Value Description

0 wLength 2 Number Size of this Interrupt Message, in bytes: 8.
2 wAttribute 2 Constant EXTENDED_DESCRIPTOR.
4 bSourceNumber 1 Number The interface number of the interface from
which the interrupt originates.
5 Reserved 1 Number Not used. Shall be set to zero.
6 wDescriptorID 2 Number ID of the Extended Descriptor from which
the interrupt originates.

6.1.4 CLASS-SPECIFIC STRING DESCRIPTOR INTERRUPT MESSAGE BODY

The wAttribute field shall be set to CLASS_SPECIFIC_STRING_DESCRIPTOR.

The wStrDescrID field contains the Descriptor ID of the Class-specific String Descriptor from which the interrupt
originates.

The Host should issue a Get Class-specific String Descriptor Command to obtain the most recent version of the
String and to re-enable interrupt generation.
Table 6-4: Class-specific String Descriptor Interrupt Message Format

Offset Field Size Value Description

0 wLength 2 Number Size of this Interrupt Message, in bytes: 8.

Release 4.0 April 2023 195

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

2 wAttribute 2 Constant CLASS_SPECIFIC_STRING_DESCRIPTOR.
4 bSourceNumber 1 Number The interface number of the AudioControl
interface from which the interrupt
originates.
5 Reserved 1 Number Not used. Shall be set to zero.
6 wStrDescrID 2 Number ID of the Class-specific String Descriptor
from which the interrupt originates.

6.2 INTERRUPT BEHAVIOR

If the Host directly changes the CUR Attribute on any AudioControl, that AudioControl shall not generate an
interrupt. However, any AudioControl CUR or RANGE Attribute within the Audio Function that changes value as a
side effect of the Host performing the direct Attribute change shall generate an interrupt.

Extended Descriptor and Class-specific String changes shall never create any side effects within the Audio Function.

An Interrupt Message is generated when an event (external or as a side effect of another Host-initiated change)
occurs that changes the value of an Attribute.

For AudioControls that use a predefined DataPart, once the Audio Function has detected a change and therefore
started the Interrupt Message generation process, the Audio Function ignores any additional changes to that
interrupt source, i.e., it does not create an internal interrupt queue to keep track of these changes. The Audio
Function then captures the most current value of the interrupt source (Control CUR DataPart) and simultaneously
re-enables the interrupt generation capability of the interrupt source. The Audio Function then creates the
AudioControl Interrupt Message with that captured CUR DataPart and posts that Message to the interrupt
Endpoint. This strategy guarantees that the value delivered in the Interrupt Message is reflecting all changes to it
that may have occurred between the moment the first change was detected (and the interrupt process was
initiated) and the moment the Interrupt Message was effectively scheduled to be delivered to the Host.

For AudioControls that use a custom DataPart and for Extended and Class-specific String Descriptor interrupt
sources, once the Audio Function has detected a change and therefore started the Interrupt Message generation
process, it ignores any additional changes to that interrupt source, i.e., it does not create an internal interrupt
queue to keep track of these changes. It then creates the corresponding Interrupt Message and posts that
Message to the interrupt Endpoint. The interrupt generation for that interrupt source shall remain disabled until
the Host retrieves the interrupt source value from the interrupt source through the appropriate Pull Command.

Release 4.0 April 2023 196

Universal Serial Bus Device Class Definition for Audio Devices

7 AUDIO DATA FORMATS

Audio Data formats can be divided in two main groups:

• Simple Audio Data Formats

• Extended Audio Data Formats

Simple Audio Data Formats can then be subdivided into two groups according to Type.

Note: Type II and Type IV Audio Data Formats are no longer supported in this version of the specification.

The first group, Type I, deals with audio data streams that are transmitted over USB and are constructed on a
sample-by-sample basis. Each audio sample is represented by a single independent symbol, contained in an audio
subslot. Different mapping schemes may be used to transform the audio samples into symbols.

Note: Mapping is considered to take place on a per-audio-sample base. Each audio sample generates one
symbol (e.g., A-law companding where a 16-bit audio sample is mapped into an 8-bit symbol).

If multiple physical audio channels are formatted into a single audio channel cluster, then samples at time x of
subsequent channels are first contained in audio subslots. These audio subslots are then interleaved, according to
the cluster channel ordering as described in this specification, and then grouped into an audio slot. The audio
samples, taken at time x+1, are interleaved in the same fashion to generate the next audio slot and so on. The
notion of physical channels is explicitly preserved during transmission. A typical example of Type I formats is the
standard PCM audio data. The following figure illustrates the concept.
Figure 7-1: Type I Audio Stream

The second group, Type III, contains Audio Data Formats that use encapsulation as described in the ISO/IEC 61937
standard before being sent over USB. One or more non-PCM encoded audio data streams are packed into
“pseudo-stereo samples” and transmitted as if they were real stereo PCM audio samples. The sampling frequency
of these pseudo samples (transport sampling frequency, as reported by the Clock Frequency Control of the

Release 4.0 April 2023 197

Universal Serial Bus Device Class Definition for Audio Devices

associated Clock Source Entity) either matches the sampling frequency of the original non-encoded PCM audio
data streams (native sampling frequency) or there is an integer ratio relationship between them. Therefore, clock
recovery at the receiving end is relatively easy.

In addition to the Simple Audio Data Formats described above, Extended Audio Data Formats are defined. These
are based on the Simple Audio Data Formats Type I and III definitions, but they provide an optional packet header
and for the Extended Audio Data Format Type I, an optional synchronous (i.e., sample accurate) control channel.

The following sections explain the different Audio Data Formats and Format Types in more detail.

7.1 SERVICE INTERVAL AND SERVICE INTERVAL PACKET DEFINITIONS

In the following paragraphs, the packetizing process for audio is described in terms of Service Interval and Service
Interval Packets. This provides a consistent model of ‘one Service Interval Packet (SIP) per Service Interval (SI)’,
irrespective of the actual transactions on the USB and the version of USB used.

A Service Interval is defined as:

𝑆𝑒𝑟𝑣𝑖𝑐𝑒 𝐼𝑛𝑡𝑒𝑟𝑣𝑎𝑙 = 𝐵𝑢𝑠 𝐼𝑛𝑡𝑒𝑟𝑣𝑎𝑙 ∗ 2(𝑏𝐼𝑛𝑡𝑒𝑟𝑣𝑎𝑙−1)

where Bus Interval has a value of 1 ms for full-speed isochronous Endpoints and 125 s for high-speed,
SuperSpeed, and Enhanced SuperSpeed isochronous Endpoints and where bInterval is the value specified in the
bInterval field of the standard Endpoint descriptor.

A Service Interval Packet is defined as the amount of isochronous data that is transported during a Service Interval.
For high-speed high-bandwidth Endpoints, the Service Interval Packet is the concatenation of the one to three
physical packets that are transferred over the bus in a Service Interval.

Note: For high-speed high-bandwidth Endpoints, the Service Interval is equal to the Bus Interval since the
bInterval field is required to be set to one.

For SuperSpeed and SuperSpeed+ Endpoints, the Service Interval Packet includes all data transferred in the Service
Interval, including bursts and multipliers.

7.2 SIMPLE AUDIO DATA FORMATS

7.2.1 TYPE I FORMATS

The following sections describe the Audio Data Formats that belong to Type I. Several terms and their definition
are presented.

7.2.1.1 USB PACKETS

Audio data streams that are inherently continuous shall be packetized when sent over the USB. The quality of the
packetizing algorithm directly influences the amount of effort needed to reconstruct a reliable sample clock at the
receiving side.

Furthermore, the chosen size of the Service Interval has a direct impact on the amount of buffer memory needed
on both sides of the pipe and on the incurred latency over the pipe. Shorter Service Intervals minimize buffer
requirements and therefore also latency at the potential expense of higher power consumption. Indeed, longer
Service Intervals potentially allow the bus (and parts of the sender and receiver’s hardware) to enter lower power
states for longer periods of time, thus conserving more power.

Release 4.0 April 2023 198

Universal Serial Bus Device Class Definition for Audio Devices

7.2.1.2 CONTINUOUS AND BURST MODE

This specification defines two possible modes to transport audio data streams over the USB:

• Continuous Mode: occurs when the Service Interval is smaller or equal to one USB Frame time of 1 ms. This
mode minimizes buffer requirements and latency at the potential expense of higher power consumption. Note
however that choosing a Service Interval that is smaller than one USB Frame time may result in excessive
system level interrupts.
• Burst Mode: occurs when the Service Interval is larger than one USB Frame time. This mode provides for
opportunities to save more power by allowing various system components to enter low power states for
extended periods of time at the expense of larger buffer sizes and increased latency.

Devices may choose to expose multiple Alternate Settings of their AudioStreaming interface(s) with different
Service Interval settings for each Alternate Setting, thus allowing the Host to choose a setting that best fits the
desired use case. However, all devices shall expose at least one Alternate Setting (besides the zero bandwidth
Alternate setting 0) that supports Continuous Mode (Service Interval <= 1 ms).

In cases where an Explicit Feedback is required to operate the isochronous data endpoint, the Service Interval of
the Feedback endpoint shall be equal to or larger than the Service Interval of the corresponding isochronous data
endpoint.

7.2.1.2.1 SERVICE INTERVAL PACKET SIZE CALCULATION

The goal shall be to keep the instantaneous number of audio slots per SI, ni as close as possible to the average
number of audio slots per SI, nav. The average nav shall be calculated as follows:
𝑇𝑆𝐼
𝑛𝑎𝑣 =
∆𝑡
where 𝑇𝑆𝐼 is the duration of an SI and ∆𝑡 is the sample time (1/𝐹𝑠 ). In some cases, 𝑛𝑎𝑣 will be a number with a
fractional part.

If the sampling rate is a constant, the allowable variation on ni is limited to two audio slots, that is, ni = 2. This
implies that 𝑛𝑖 may vary between 𝐼𝑁𝑇(𝑛𝑎𝑣 ) − 1 (small SIP), 𝐼𝑁𝑇(𝑛𝑎𝑣 ) (medium SIP) and 𝐼𝑁𝑇(𝑛𝑎𝑣 ) + 1 (large SIP).
For all 𝑖:

𝑛𝑖 = 𝐼𝑁𝑇(𝑛𝑎𝑣 ) − 1 | 𝐼𝑁𝑇(𝑛𝑎𝑣 ) | 𝐼𝑁𝑇(𝑛𝑎𝑣 ) + 1

Furthermore, a large SIP shall be generated as soon as it becomes available. Typically, a source will generate small
SIPs as long as the accumulated fractional part of 𝑛𝑎𝑣 remains < 1. Once the accumulated fractional part of 𝑛𝑎𝑣
becomes ≥ 1, the source shall send a large SIP and decrement the accumulator by 1.

Note that in some cases (for example, asynchronous Endpoint operating at low frequency), the above formula will
result in 𝑛𝑖 = 0 occasionally, i.e., no payload is available for the Service Interval. In this case, a zero-length packet
shall be sent in that Service Interval.

Due to possible different notions of time in the source and the sink (they may each have their own independent
sampling clock), the (small SIP)/(large SIP) pattern generated by the source may be different from what the sink
expects. Therefore, the sink shall be capable to always accept a large SIP.

Example:

Release 4.0 April 2023 199

Universal Serial Bus Device Class Definition for Audio Devices

Assume 𝐹𝑆 = 44,100 Hz and 𝑇𝑉𝐹 = 1ms. Then 𝑛𝑎𝑣 = 44.1 audio slots. Since the source can only send an integer
number of audio slots per SI, it will send small SIPs of 44 audio slots. Each SI, it therefore sends ‘0.1 slot’ too few
and it will accumulate this fractional part in an accumulator. After having sent 9 small SIPs of 44 audio slots, at the
tenth SI it will have exactly one audio slot in excess and therefore can send a large SIP containing 45 audio slots.
Decrementing the accumulator by 1 brings it back to 0 and the process can start all over again. The source will thus
produce a repetitive pattern of 9 small SIPs of 44 audio slots followed by 1 large SIP of 45 audio slots. The following
table illustrates the process:
Table 7-1: Packetization

#SI 𝒏𝒂𝒗 𝒏𝒊 Fraction Accumulator

n 44.1 44 0.1 0.1
n+1 44.1 44 0.1 0.2
n+2 44.1 44 0.1 0.3
n+3 44.1 44 0.1 0.4
n+4 44.1 44 0.1 0.5
n+5 44.1 44 0.1 0.6
n+6 44.1 44 0.1 0.7
n+7 44.1 44 0.1 0.8
n+8 44.1 44 0.1 0.9
n+9 44.1 45 0.1 1.0 -> 0
n+10 44.1 44 0.1 0.1
n+11 44.1 44 0.1 0.2
… … … … …

7.2.1.3 PITCH CONTROL

If the sampling rate can be varied (to implement pitch control), the allowable variation on ni is limited to one audio
slot per SI. For all 𝑖:
𝑛𝑖+1 = 𝑛𝑖 |𝑛𝑖 ± 1
Pitch control is restricted to adaptive Endpoints only. AudioStreaming interfaces that support pitch control on their
isochronous Endpoint are required to report this in the class-specific Endpoint Descriptor. In addition, a Set/Get
Pitch Control Command is required to enable or disable the pitch control functionality.

7.2.1.4 AUDIO SUBSLOT

The basic structure used to represent audio data is the audio subslot. An audio subslot holds a single audio sample.
An audio subslot always contains an integer number of bytes.

This specification limits the possible audio subslot sizes (wSubslotSize) to 1, 2, 3, 4, or 8 bytes per audio subslot. An
audio sample is represented using a number of bits (wBitResolution) less than or equal to the total number of bits
available in the audio subslot, i.e., wBitResolution  wSubslotSize*8.

AudioStreaming Endpoints shall be constructed in such a way that a valid transfer may take place as long as the
reported audio subslot size (wSubslotSize) is respected during transmission. If the reported bits per sample
(wBitResolution) do not correspond with the number of significant bits used during transfer, the device will either

Release 4.0 April 2023 200

Universal Serial Bus Device Class Definition for Audio Devices

discard trailing significant bits ([actual_bits_per_sample] > wBitResolution) or interpret trailing zeroes as
significant bits ([actual_bits_per_sample] < wBitResolution).

7.2.1.5 AUDIO SLOT

An audio slot consists of a collection of audio subslots, each containing an audio sample of a different physical
audio channel, taken at the same moment in time. The number of audio subslots in an audio slot equals the
number of logical audio channels in the audio channel cluster. The ordering of the audio subslots in the audio slot
obeys the rules set forth in this specification. All audio subslots shall have the same audio subslot size.

7.2.1.6 AUDIO STREAMS

An audio stream is a concatenation of a potentially very large number of audio slots, ordered according to
ascending time. Streams are packetized when transported over USB whereby SIPs can only contain an integer
number of audio slots. Each packet always starts with the same channel, and the channel order is respected
throughout the entire transmission. If, for any reason, there are no audio slots available to construct a SIP, a Zero-
Length Packet shall be sent instead.

7.2.1.7 TYPE I SUPPORTED FORMATS

The following paragraphs list all currently supported Type I Audio Data Formats. The allowed values for the
wFormat field of the class-specific AudioStreaming Self descriptor for the different Type I Audio Data Formats can
be found in Appendix B.1, “Audio Data Formats .”

7.2.1.7.1 PCM FORMAT

The PCM (Pulse Coded Modulation) format is the most used audio format to represent audio data streams. The
audio data is not compressed and uses a signed two’s-complement fixed point format. It is left-justified (the sign
bit is the MSb) and data is padded with trailing zeroes to fill the remaining unused bits of the subslot. The binary
point is located to the right of the sign bit so that all values lie within the range [-1, +1).

7.2.1.7.2 PCM8 FORMAT

The PCM8 format is introduced to be compatible with the legacy 8-bit wave format. Audio data is uncompressed
and uses 8 bits per sample (wBitResolution = 8). In this case, data is unsigned fixed-point, left-justified in the audio
subslot, MSb first. The range is [0..255].

7.2.1.7.3 IEEE_FLOAT FORMAT

The IEEE_FLOAT format is based on the [ANSI/IEEE-754] floating-point standard. Audio data is represented using
the basic single-precision format. The basic single-precision number is 32 bits wide and has an 8-bit exponent and a
24-bit mantissa. Both mantissa and exponent are signed numbers, but neither is represented in two’s-complement
format. The mantissa is stored in sign magnitude format and the exponent in biased form (also called excess-n
form). In biased form, there is a positive integer (called the bias) which is subtracted from the stored number to
get the actual number. For example, in an eight-bit exponent, the bias is 127. To represent 0, the number 127 is
stored. To represent -100, 27 is stored. An exponent of all zeroes and an exponent of all ones are both reserved for
special cases, so in an eight-bit field, exponents of -126 to +127 are possible. In the basic floating-point format, the
mantissa is assumed to be normalized so that the most significant bit is always one, and therefore is not stored.
Only the fractional part is stored. Denormalized (exponent = 0) values are interpreted as zero.

Release 4.0 April 2023 201

Universal Serial Bus Device Class Definition for Audio Devices

The 32-bit IEEE-754 floating-point word is broken into three fields. The most significant bit stores the sign of the
mantissa, the next group of 8 bits stores the exponent in biased form, and the remaining 23 bits store the
magnitude of the fractional portion of the mantissa. For further information, refer to the ANSI/IEEE-754 standard.

The data is conveyed over USB using 32 bits per sample (wBitResolution = 32; wSubslotSize = 4).

7.2.1.7.4 ALAW FORMAT AND LAW FORMAT

Starting from 12- or 16-bits linear PCM samples, simple mapping down to 8-bits per sample (one byte per sample)
can be achieved by using logarithmic companding. The companded audio data uses 8 bits per sample
(wBitsPerSample = 8). Data is signed fixed point, left-justified in the subslot, MSb first. The companded range is
[-128..128]. The difference between Alaw and Law compression lies in the formulae used to achieve it. Refer to
the ITU G.711 standard for further details.

7.2.1.7.5 DSD FORMAT

The Direct-Stream Digital (DSD) format uses pulse-density modulation encoding—a technology primarily used to
store audio signals on SACD (Super Audio CD) digital storage media.

Audio data is stored as single-bit delta-sigma modulated digital audio; i.e. a sequence of single-bit values at a
sampling rate of 2.8224 MHz (64 times the CD audio sampling rate of 44.1 kHz) for basic sampling rate DSD64,
5.6448 MHz for DSD128 (2X-rate DSD), 11.2896 MHz for DSD256 (4X-rate DSD), 22.5792 MHz for DSD 512 (8X-rate
DSD), and 45.1584 MHz for DSD1024 (16X-rate DSD). 48 kHz based DSD streams are also in existence. In that case,
the bitstream sampling rates are 3.072 MHz, 6.144 MHz, 12.288 MHz, 24.576 MHz, and 49.152 MHz respectively.

No matter what sampling rate the DSD stream uses, the audio subslot size is fixed to 64 bits (wBitResolution = 64;
wSubslotSize = 8) so that, at the transport layer, the DSD stream always looks like 64-bit PCM-like data. Therefore,
the transport sampling rate of the DSD stream, packetized as 64-bit PCM samples, is always 1/64 of the DSD
sampling rate and the Clock Source Entity, connected to the Terminal, representing the DSD AudioStreaming
interface shall always advertise this transport sampling rate:

• 44.1 kHz, 88.2 kHz, 176.4 kHz, 352.8 kHz, or 705.6 kHz for 44.1 kHz based DSD streams
• 48 kHz, 96 kHz, 192 kHz, 384 kHz, or 768 kHz for 48 kHz based DSD streams

The subslot data is a 64-bit value in little-endian notation where bit D0 (LSb) is the most recent and bit D63 (MSb)
is the least recent. This will result in the most recent bit D0 to appear on the USB wire first.

7.2.2 TYPE III FORMATS

These formats are based upon the ISO/IEC 61937 standard. The ISO/IEC 61937 standard describes a method to
transfer non-PCM encoded audio bit streams over an ISO/IEC 60958 digital audio interface, together with the
transfer of the accompanying “Channel Status” and “User Data.”

The ISO/IEC 60958 standard specifies a widely used method of interconnecting digital audio equipment with two-
channel linear PCM audio. The ISO/IEC 61937 standard describes a way in which the ISO/IEC 60958 interface shall
be used to convey non-PCM encoded audio bit streams for consumer applications.

The same basic techniques used in ISO/IEC 61937 are reused here to convey non-PCM encoded audio bit streams
over a Type III formatted audio stream. From a USB transfer standpoint, the data streaming over the interface
looks exactly like two-channel 16-bit PCM audio data.

Release 4.0 April 2023 202

Universal Serial Bus Device Class Definition for Audio Devices

7.2.2.1 TYPE III SUPPORTED FORMATS

Refer to the ISO/IEC 60958 and ISO/IEC 61937 (several parts) specifications for detailed format information. The
allowed values for the wFormat field of the class-specific AudioStreaming Self descriptor for the different Type III
Audio Data Formats can be found in Appendix B.1, “Audio Data Formats .”

The following is a list of formats that is covered by the above specifications.

• PCM_IEC60958
• AC-3
• MPEG-1_Layer1
• MPEG-1_Layer2/3 or MPEG-2_NOEXT
• MPEG-2_EXT
• MPEG-2_AAC_ADTS
• MPEG-2_Layer1_LS
• MPEG-2_Layer2/3_LS
• DTS-I
• DTS-II
• DTS-III
• ATRAC
• ATRAC2/3
• WMA
• E-AC-3
• MAT
• DTS-IV
• MPEG-4_HE_AAC
• MPEG-4_HE_AAC_V2
• MPEG-4_AAC_LC
• DRA
• MPEG-4_HE_AAC_SURROUND
• MPEG-4_AAC_LC_SURROUND
• MPEG-H_3D_AUDIO
• AC4
• MPEG-4_AAC_ELD

7.3 EXTENDED AUDIO DATA FORMATS

In addition to the Simple Audio Data Formats described above, Extended Audio Data Formats Type I and III are
defined. These are based on the Simple Audio Data Formats Type I and III definitions, but they provide an optional
Header and for the Extended Audio Data Format Type I, an optional synchronous (i.e. sample accurate) Control
Stream.

Each SIP shall start with a SIPDescriptor, followed by the following optional components:

• Header
• Audio data, formatted according to the Simple Audio Data Formats Type I or III
• Synchronous vendor-specific Control Stream (only for Extended Audio Format Type I)
These three components may be optionally present on a per-SIP basis.

Release 4.0 April 2023 203

Universal Serial Bus Device Class Definition for Audio Devices

The SIPDescriptor is exactly 4 bytes long and has the following layout:

The wFlags field indicates which of the components are present in the SIP as follows:

• Bit D0 indicates whether a Header is present in the SIP (D0=0b1) or not (D0=0b0). When the Header is not
present, the wHeaderLength field shall be set to zero (0x0000).
• Bit D1 indicates whether an AudioSlot is present in all the Extended AudioSlots of the SIP (D1=0b1) or not
(D1=0b0).
• Bit D2 indicates whether a Control Stream is present. In other words, it indicates whether a Control Word is
present in all the Extended AudioSlots of the Extended Type I SIP (D2=0b1) or not (D2=0b0). This bit shall be
set to zero (D2=0b0) for Extended Type III formats.
• Bits D15..3 are reserved.

The wHeaderLength field indicates the total length of the Header in bytes. This includes all the SubHeaders that
together make up the total Header.

The presence of the SIPDescriptor allows for maximum flexibility. For example, it is possible to create an Audio
Stream consisting of a Control Stream only, without Header or audio data. It is also possible to create an Audio
Stream where Headers are only present occasionally.

If present, the Header immediately follows the SIPDescriptor. There is only one Header allowed per SIP. The
Header may consist of one or more SubHeaders, each having a different purpose. Each SubHeader shall start with
a wSubHeaderID field, uniquely identifying the purpose of the SubHeader. The length of each SubHeader and the
semantics of the remaining fields in the SubHeader are defined by this specification. The structure and
composition of the Header and the order in which the SubHeaders appear in the Header may change from SIP to
SIP.
Table 7-2: SIPDescriptor Layout

Offset Field Size Value Description

0 wFlags 2 Bitmap D0: Header present.
D1: AudioSlot present.
D2: Control Stream present.
D15..3: Reserved.
2 wHeaderLength 2 Number Total length of the Header, in bytes.

7.3.1 EXTENDED TYPE I FORMATS

The following figure illustrates the possible layout of a SIP. The greyed parts are optional.

Note: To have a meaningful stream, at least one of the optional components shall be present.

Release 4.0 April 2023 204

Universal Serial Bus Device Class Definition for Audio Devices

Figure 7-2: Extended Type I Format

AudioSlot

AudioSlot

AudioSlot

AudioSlot

AudioSlot

AudioSlot
Audio Data, formatted according to
SIPDescriptor

Simple Audio Format Type I

Header

Control

Control

Control

Control

Control

Control
One Control Word per AudioSlot

Extended AudioSlot (x-1) Extended AudioSlot (x) Extended AudioSlot (x+1)

Control (x+1)
Control (x-1)

Control (x)
SubSlot 1

SubSlot 2

SubSlot n

SubSlot 1

SubSlot 2

SubSlot n

SubSlot 1

SubSlot 2

SubSlot n
Time
x-1 x x+1 x+2

An Extended AudioSlot is the concatenation of a Control Word, followed by the Type I AudioSlot. The Control
Stream therefore consists of a sequence of Control Words, where each Control Word is synchronous to its
associated AudioSlot. There are as many Control Words per SIP as there are AudioSlots in the SIP. The byte size of
the Control Words is independent of the AudioSubSlot size and is the same for each AudioSlot.

7.3.2 EXTENDED TYPE III FORMATS

The following figure illustrates the possible layout of a SIP. The greyed parts are optional.

Note: In order to have a meaningful stream, at least one of the optional components shall be present.
Figure 7-3: Extended Type III Format
SIPDescriptor

Header

Audio Data, formatted according to Simple Audio Format

Type III

The Header is optionally followed by the actual encoded audio frame data.

Release 4.0 April 2023 205

Universal Serial Bus Device Class Definition for Audio Devices

7.4 CLASS-SPECIFIC AUDIOSTREAMING SELF DESCRIPTOR

The wStartDelayUnits field indicates the units that are used for the wStartDelay field, immediately following. The
wStartDelay field indicates how long it takes this Interface to reliably produce valid outgoing data or effectively
consume incoming data. This time is measured using as a time reference the first IN or OUT PID that occurs at the
start of an audio stream. See Section 3.14.2, “AudioStreaming Interface.”

The wFormat field indicates the Audio Data Format that is supported by this Alternate Setting of the
AudioStreaming Interface.

A complete list of currently supported Audio Data Formats, their wFormat field value, and their usage can be
found in Appendix B, “Audio Data Format Codes.”

For Simple and Extended Type I formats, the wSubslotSize field can be set to 1, 2, 3, 4, or 8, and the
wBitResolution field indicates how many bits of the total number of available bits in the audio subslot are
effectively used to convey audio information.

For Simple and Extended Type III formats, the wSubslotSize field shall be set to 2 and the wBitResolution field
shall be set to 16.

For Simple Type I and Type III formats, the wAuxProtocols and wControlSize fields are not used and shall be set to
zero.

For Extended Type I and Extended Type III formats, this specification defines several Auxiliary Protocols (See
Section 7.5, “Auxiliary Protocols.”). The wAuxProtocols field contains a bitmap identifying which Auxiliary
Protocols this AudioStreaming interface’s Alternate Setting requires. For each Auxiliary Protocol, the
AudioStreaming interface may offer up to two Alternate Settings, one in which the Auxiliary Protocol is required
and the other in which it is not. For example, an AudioStreaming Interface may offer two Alternate Settings, one
indicating the required use of HDCP and the other indicating that it operates without HDCP.

For Extended Type I formats, the wControlSize field indicates the size in bytes of each Control Channel Word in the
stream. It shall be set to zero for all other formats.
Table 7-3: Class-specific AudioStreaming Self Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this descriptor in bytes: 28.

2 wDescriptorType 2 Constant EXT_INTERFACE descriptor type.

4 wDescriptorSubtype 2 Constant AS_SELF descriptor subtype.
6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.
10 dOptControls 4 Bitmap D0: Active Alternate Setting Control.
D1: Valid Alternate Settings Control.
D31..2: Reserved.
14 wStartDelayUnits 2 Number Indicates the units used for the wStartDelay
field:
0: Undefined.
1: Milliseconds.
2: Decoded PCM samples.
3..255: Reserved.

Release 4.0 April 2023 206

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

16 wStartDelay 2 Number Indicates the time it takes this Interface to
reliably produce or consume streamed data.
Units used depend on the value of the
bStartDelayUnits field.
18 wFormat 2 Number The Audio Data Format used to
communicate with this Alternate Setting of
the AudioStreaming Interface. See Section
7, “Audio Data Formats” for further details.
20 wSubslotSize 2 Number The number of bytes occupied by one audio
subslot.
22 wBitResolution 2 Number The number of effectively used bits from
the available bits in an audio subslot.
24 wAuxProtocols 2 Bitmap Bitmap, indicating which Auxiliary Protocols
are required.
26 wControlSize 2 Number Size of the Control Channel Words, in bytes.

7.5 AUXILIARY PROTOCOLS

7.5.1 HDCP PROTOCOL

The HDCP protocol uses the HDCP SubHeader to convey the periodic Link Synchronization information, required by
HDCP 2.3. HDCP 2.3 requires that a 32-bit streamCtr value and a 64-bit inputCtr value be sent periodically from the
content transmitter to the content receiver. The Header construct in the Extended Format Types shall be used to
convey that information periodically as follows.

The HDCP SubHeader has the following layout:

Table 7-4: HDCP SubHeader Layout

Offset Field Size Value Description

0 wLength 2 Number Size of this SubHeader, in bytes. 18.
2 wSubHeaderID 2 Number HDCP_ENCRYPTION
4 wOffset 2 Number Offset to first full 16-byte encrypted data
block in the SIP.
6 dStreamCtr 4 Number The streamCtr value as assigned by the
HDCP transmitter.
10 qInputCtr 8 Number The inputCtr value associated with the first
full 16-byte encrypted data block in the SIP.
The wLength field contains the size of this SubHeader, expressed in bytes.

The wSubHeaderID field contains the HDCP_ENCRYPTION constant to uniquely identify this SubHeader as an HDCP
SubHeader.

The wOffset field contains an offset value ranging from 0 to 15, indicating at which byte position, measured from
the start of the audio data in the SIP, the first full 16-byte encrypted data block (HDCP Cipherblock) starts. The
value in the qInputCtr field shall pertain to that same 16-byte encrypted data block.

Note: Since an HDCP Cipher Block is always 16 bytes or 128 bits long, the start of a full Cipher Block is
guaranteed to occur within the first 16 bytes of the actual audio data payload. Also, when a Control

Release 4.0 April 2023 207

Universal Serial Bus Device Class Definition for Audio Devices

Stream is present, the wOffset field value only pertains to the actual audio data bytes (the Control
Stream is not encrypted) and the Control Stream should be separated from the actual audio data
bytes before the offset is applied.

The dStreamCtr field contains the streamCtr value associated with this stream as assigned by the HDCP
transmitter.

The qInputCtr field contains the inputCtr value associated with the first full 16-byte encrypted audio data block in
the SIP.

The HDCP SubHeader shall be present at least every HDCP_PACKET_HEADER_TIME (see Appendix B.3, “Audio
Format General Constants”) but may occur more frequently. The presence of the HDCP SubHeader indicates to the
HDCP receiver that the content is HDCP-encrypted.

7.6 ADDING NEW AUDIO DATA FORMATS

Adding new Audio Data Formats to this specification is achieved by proposing a fully documented Audio Data
Format to the Audio Device Class Working Group. Upon acceptance, they will register the new Audio Data Format
(attribute a unique value for the wFormat field of the class-specific AudioStreaming Self descriptor) and update
this document accordingly. This process will also guarantee that new releases of generic USB audio drivers will
support the newly registered Audio Data Formats.

It Is always possible to use vendor-specific definitions if the above procedure is considered unsatisfactory.

7.7 ADDING NEW AUXILIARY PROTOCOLS

Adding new Auxiliary Protocols to this specification is achieved by proposing a fully documented Auxiliary Protocol
to the Audio Device Class Working Group. Upon acceptance, they will register the new Auxiliary Protocol (attribute
a unique Auxiliary Protocol constant) and update this document accordingly. This process will also guarantee that
new releases of generic USB audio drivers will be able to support the newly registered Auxiliary Protocols.

Release 4.0 April 2023 208

Universal Serial Bus Device Class Definition for Audio Devices

8 BACKWARDS COMPATIBILITY CONSIDERATIONS

ADC 4.0 compliant Devices are not required to include configurations that are compliant with previous versions of
the ADC Specification. In other words, backwards compatibility is optional.

To ensure backwards compatibility with a certain revision level of the Audio Device Class-specification, called the
Base Revision Level (BRL), a Device that is also capable of operating its Audio Function at Higher Revision Levels
(HRL) than the BRL, is called a multi-mode Device and shall do all of the following.

• Provide the BRL Audio Function Descriptor Set during enumeration that is fully compliant with the BRL Audio
Device Class-specification.
• Be able to operate its Audio Function in compliance with and as indicated by the BRL Audio Function
Descriptor Set.
• Make available to the Host one or more HRL Audio Function Descriptor Sets that each are fully compliant with
their respective HRL Audio Device Class-specifications. How this is accomplished is detailed further below.
• Be able to operate its Audio Function in compliance with those HRL Audio Device Class-specifications as
indicated by their respective HRL Audio Function Descriptor Sets, once instructed to do so. How this is
accomplished is detailed further below.

For this 4.0 version of the Specification, Revision Levels 2.0 and 4.0 shall be the only Base Revision Levels
supported and Revision Level 4.0 shall be the only Higher Revision Level supported if the Device supports Base
Revision Level 2.0.

Note: It is anticipated that when future version of this Specification (4.1+) become available, Devices may
choose to advertise multiple Higher Revision Level modes of operation to ensure the broadest
possible level of compatibility. For example, a Device may advertise a 2.0 BRL mode of operation,
together with both 4.0 and 5.0 HRL modes of operation.

Both the BRL Audio Function Descriptor Set and the HRL Audio Function Descriptor Sets follow the familiar ADC
layout. All Descriptors in both Sets follow the traditional Descriptor layout as described in the USB Specifications,
i.e., they start with a bLength field, followed by a bDescriptorType field. Also, see Section 4.3, “Audio Function
Descriptor Set” for more details.

To ensure backwards compatibility, the HRL Audio Function Descriptor Sets shall use the identical footprint, layout,
and contents for their standard Interface and Endpoint Descriptors (shown in white in the figures in Section 4.3.1,
“Audio Function Descriptor Set Layouts”) as is used for the BRL standard Interface Descriptors. The only difference
is that the standard Interface Descriptors now indicate the Higher Revision Level in their bFunctionProtocol and
bInterfaceProtocol fields. Indeed, for the Audio Function to be able to switch from BRL operation to a different
HRL operating mode, the resources that are allocated once by the USB stack during BRL enumeration need to
match the resource requirements of all the HRL operating modes the Audio Function supports. Therefore, the
standard Descriptor Set exposed by the BRL Audio Function Descriptor Set and all the exposed HRL Audio Function
Descriptor Sets shall be identical, except for the values in the bFunctionProtocol and bInterfaceProtocol fields.

The class-specific Descriptors (shown in grey in the figures in Section 4.3.1, “Audio Function Descriptor Set
Layouts”) can be substantially different between the BRL and HRL Audio Function Descriptor Sets. Nevertheless, it
is imperative that the class-specific Descriptors still accurately reflect the actual BRL or HRL operational
functionality respectively.

As already mentioned before, the BRL Audio Function Descriptor Set is provided to the Host during enumeration of
the Device, while the HRL Interface Descriptor Sets can be retrieved using a method described further below.

Release 4.0 April 2023 209

Universal Serial Bus Device Class Definition for Audio Devices

Note: The Interface Association Descriptor of the BRL Audio Function Descriptor Set shall not contain a
reference to any MIDI Streaming Interfaces. For more information, see the Universal Serial Bus Device
Class Definition for MIDI Devices.

8.1 SIMPLE MULTI-MODE DEVICE EXAMPLE

The following is a hypothetical example to illustrate the procedure outlined above.

Suppose a vendor wants to build a Device that can operate in two modes. A (legacy) mode that is compliant with
the ADC 2.0 specification and another mode that is compliant with this ADC 4.0 specification.

For maximum backwards compatibility, the vendor first creates the set of standard Interface and Endpoint
Descriptors as illustrated in Figure 8-1.
Figure 8-1: Standard Interface Descriptor Set Layout

Interface Association

Interface #0 Interface #1 Interface #2

AUDIO_CONTROL AUDIO_STREAMING AUDIO_STREAMING
Alternate Setting #0 Alternate Setting #0 Alternate Setting #0

Interface #1 Interface #2
AUDIO_STREAMING AUDIO_STREAMING
Alternate Setting #1 Alternate Setting #1

Standard
Endpoint #6 * Endpoint #8 *
MaxPacketSize: 196 MaxPacketSize: 98

Interface #2
AUDIO_STREAMING
Alternate Setting #2
* In the case of SuperSpeed and SuperSpeed+, this Descriptor may actually
consist of multiple standard Descriptors

Endpoint #8 *
MaxPacketSize: 196

Then for 2.0 BRL operation, the Audio Function exposes the Audio Function Descriptor Set (example) during
enumeration as illustrated in Figure 8-2. All the class-specific Interface and Endpoint Descriptors are added to
accurately describe the Audio Function in its ADC 2.0 compliant form.

Release 4.0 April 2023 210

Universal Serial Bus Device Class Definition for Audio Devices

Figure 8-2: ADC 2.0 Audio Function Descriptor Set

Interface Association

Interface #0 Interface #1 Interface #2

AUDIO_CONTROL AUDIO_STREAMING AUDIO_STREAMING
Alternate Setting #0 Alternate Setting #0 Alternate Setting #0
Protocol 2.0 Protocol 2.0 Protocol 2.0

2.0 Interface #1 Interface #2

CS_Interface
AUDIO_STREAMING AUDIO_STREAMING
Alternate Setting #1 Alternate Setting #1
Protocol 2.0 Protocol 2.0

2.0 2.0
CS_Interface CS_Interface
Standard

Traditional
Endpoint #6 * Endpoint #8 *
MaxPacketSize: 196 MaxPacketSize: 98
Class-Specific

2.0 2.0
CS_Endpoint CS_Endpoint

Interface #2
AUDIO_STREAMING
Alternate Setting #2
* In the case of SuperSpeed and SuperSpeed+, this Descriptor may actually
Protocol 2.0
consist of multiple standard Descriptors

2.0
CS_Interface

Endpoint #8 *
MaxPacketSize: 196

2.0
CS_Endpoint

For 4.0 HRL operation, the Audio Function exposes the Audio Function Descriptor Set as illustrated in Figure 8-3. All
the class-specific Interface and Endpoint Descriptors are added to accurately describe the Audio Function in its
ADC 4.0 compliant form.

Release 4.0 April 2023 211

Universal Serial Bus Device Class Definition for Audio Devices

Figure 8-3: ADC 4.0 Audio Function Descriptor Set

Interface Association

Interface #0 Interface #1 Interface #2

AUDIO_CONTROL AUDIO_STREAMING AUDIO_STREAMING
Alternate Setting #0 Alternate Setting #0 Alternate Setting #0
Protocol 4.0 Protocol 4.0 Protocol 4.0

CS_INTERFACE 4.0 Interface #1 Interface #2

AC_GENERIC AUDIO_STREAMING AUDIO_STREAMING
Alternate Setting #1 Alternate Setting #1
Protocol 4.0 Protocol 4.0

CS_INTERFACE 4.0 CS_INTERFACE 4.0

AS_GENERIC AS_GENERIC

Standard
Endpoint #6 * Endpoint #8 *
MaxPacketSize: 196 MaxPacketSize: 98
Traditional
Class-Specific

Extended
Class-Specific
Interface #2
AUDIO_STREAMING
Alternate Setting #2
Protocol 4.0
* In the case of SuperSpeed and SuperSpeed+, this Descriptor may actually
consist of multiple standard Descriptors CS_INTERFACE 4.0
AS_GENERIC
Extended Descriptor Store

EXT_INTERFACE
Endpoint #8 *
CS_Interface MaxPacketSize: 196
CS_Interface
CS_Interface
CS_Interface

8.2 DISCOVERY OF HIGHER VERSION LEVEL SUPPORT FOR AN AUDIO FUNCTION

If an Audio Function supports a mode at a higher revision level than advertised during enumeration, the Device
shall include a new BOS Capability Descriptor, the Higher Revision Level Function Capability Descriptor, as part of
its BOS Descriptor. Note that the BOS Descriptor may include multiple HRL Function Capability Descriptors, one for
each Higher Revision Level an embedded Audio Function supports, and this for each embedded Audio Function in
the Device. The bConfigurationValue field, together with the bFirstInterface field in the HRL Function Capability
Descriptor uniquely identify the Audio Function in the Device to which this HRL Audio Function Capability
Descriptor refers.

The BOS Capability Code is assigned by the USB-IF. The assigned code can be found in Appendix A.1, “BOS
Capability Codes.”

Note: For this 4.0 revision of the Specification, only one 4.0 HRL Function Capability Descriptor per Audio
Function is allowed.

Release 4.0 April 2023 212

Universal Serial Bus Device Class Definition for Audio Devices

Table 8-1: BOS HRL Function Capability Descriptor

Offset Field Size Value Description

0 bLength 1 Number Size of this Descriptor, in bytes: 11.
1 bDescriptorType 1 Constant DEVICE_CAPABILITY Descriptor type, as
defined in the USB 3 Specification.
2 bDevCapabilityType 1 Constant HRL_FUNCTION. Higher Revision Level
Audio Function Capability.
3 bConfigurationValue 1 Number Matches the Configuration Value of exactly
one Configuration Descriptor. Only that
Configuration will have an Audio Function
which may be replaced by this Capability.
4 bFirstInterface 1 Number Matches the Interface Number of the first
USB interface being replaced by this HRL
Audio Function.
5 bInterfaceCount 1 Number The number of USB Interfaces being
replaced by this HRL Audio Function.
6 bFunctionClass 1 Constant AUDIO. Audio Function Class
7 bFunctionSubClass 1 Constant Audio Function Subclass as defined in the
IAD of the HRL Interface Descriptor Set.
8 bFunctionProtocol 1 Number Audio Function Protocol as defined in the
IAD of the HRL Interface Descriptor Set. This
indicates the compliance level of the Audio
Function being described by this Descriptor.
9 wHRLFuncDescrID 2 Number The ID of the Extended Descriptor that
contains the HRL Audio Function Descriptor
Set for the Audio Function.

8.3 RETRIEVAL OF AN HRL AUDIO FUNCTION DESCRIPTOR SET

The HRL Audio Function Descriptor Set consists exclusively of traditional Descriptors, formatted exactly as they
would be if they were to be retrieved at enumeration time. The IAD is the first Descriptor in the Set, followed by
the standard AudioControl Interface Descriptor and so on.

To make the Set accessible for the Host during normal operation, the HRL Audio Function Descriptor Set is
encapsulated within the Extended Function Container Descriptor so that it can be retrieved from the AudioControl
Interface using the Get Extended Descriptor Command as defined in Section 5.3.5.2, “Extended Descriptor ”. Note
that the HRL Function Container Descriptor can only be retrieved from the Device after the Device is switched into
Higher Revision Level mode of operation. Whether the BRL Function Container Descriptor can be retrieved while
the device is in Base Revision Level mode of operation is implementation dependent.

The following table outlines the layout of this Function Container Descriptor.
Table 8-2: Function Container Descriptor

Offset Field Size Value Description

0 wLength 2 Number Size of this Descriptor, in bytes: 10+n.
2 wDescriptorType 2 Constant EXT_INTERFACE Descriptor Type.
4 wDescriptorSubType 2 Constant FUNCTION_CONTAINER Descriptor Subtype.

Release 4.0 April 2023 213

Universal Serial Bus Device Class Definition for Audio Devices

Offset Field Size Value Description

6 wDescriptorID 2 Number Unique ID for this Descriptor.
8 wStrDescriptorID 2 Number Unique ID for a String Descriptor.
10 HRLAFDescrSet n N/A The HRL Audio Function Descriptor Set.

8.4 SWITCHING THE AUDIO FUNCTION FROM BRL OPERATION TO HRL OPERATION
After inspecting the BOS Descriptor from the Device, the Host can decide to switch the Audio Function into a
Higher Revision Level operating mode. This is done through a Set Request operation to the Switch Function
Capability that shall be supported whenever the Audio Function supports one or more HRL modes of operation.
See Section 5.2.2, “Switch Function” for more details. Note that the switching operation can only be performed
once after power up or reset of the Audio Function.

Release 4.0 April 2023 214

Universal Serial Bus Device Class Definition for Audio Devices

APPENDIX A. AUDIO DEVICE CLASS CODES

This Appendix list all the Audio Device Class constant definitions used by this specification.

A.1 BOS CAPABILITY CODES

Table A-1: BOS Capability Codes

BOS Capability Codes Value

HRL_FUNCTION 0x12

A.2 AUDIO FUNCTION CLASS CODE

Table A-2: Audio Function Class Code

Audio Function Class Code Value

AUDIO 0x01

A.3 AUDIO FUNCTION SUBCLASS CODES

Table A-3: Audio Function Subclass Codes

Audio Function Subclass Code Value

FUNCTION_SUBCLASS_UNDEFINED 0x00
Reserved 0x01..0xFF

A.4 AUDIO FUNCTION PROTOCOL CODES

Table A-4: Audio Function Protocol Codes

Audio Function Protocol Code Value

FUNCTION_PROTOCOL_UNDEFINED 0x00
AF_VERSION_04_00 IP_VERSION_04_00

A.5 AUDIO INTERFACE CLASS CODE

Table A-5: Audio Interface Class Code

Audio Interface Class Code Value

AUDIO 0x01

A.6 AUDIO INTERFACE SUBCLASS CODES

Table A-6: Audio Interface Subclass Codes

Audio Interface Subclass Code Value

INTERFACE_SUBCLASS_UNDEFINED 0x00
AUDIOCONTROL 0x01
AUDIOSTREAMING 0x02
Reserved for MIDI use 0x03

Release 4.0 April 2023 215

Universal Serial Bus Device Class Definition for Audio Devices

Audio Interface Subclass Code Value

Reserved 0x04..0xFE

A.7 AUDIO INTERFACE PROTOCOL CODES

Table A-7: Audio Interface Protocol Codes

Audio Interface Protocol Code Value

IP_VERSION_04_00 0x40
Reserved All Other Values

A.8 AUDIO CLASS-SPECIFIC TRADITIONAL DESCRIPTOR TYPES

Table A-8: Audio Class-specific Traditional Descriptor Types

Descriptor Type Value

CS_UNDEFINED 0x20
CS_INTERFACE 0x21
CS_STRING 0x22
Reserved 0x23..0xFF

A.9 AUDIO CLASS-SPECIFIC TRADITIONAL DESCRIPTOR SUBTYPES

Table A-9: Audio Class-specific Traditional Descriptor Subtypes

Descriptor Subtype Value

SUBTYPE_UNDEFINED 0x00
AC_GENERIC 0x01
AS_GENERIC 0x02

A.10 AUDIO CLASS-SPECIFIC EXTENDED DESCRIPTOR TYPES

Table A-10: Audio Class-specific Extended Descriptor Types

Descriptor Type Value

EXT_UNDEFINED 0x0000
EXT_INTERFACE 0x0001
EXT_STRING 0x0002
Reserved 0x0003..0xFFFF

A.11 AUDIO CLASS-SPECIFIC EXTENDED DESCRIPTOR SUBTYPES

Table A-11: Audio Class-specific Extended Descriptor Subtypes

Descriptor Subtype Value

EXT_SUBTYPE_UNDEFINED 0x0000

Release 4.0 April 2023 216

Universal Serial Bus Device Class Definition for Audio Devices

Descriptor Subtype Value

AC_SELF 0x0001
INPUT_TERMINAL 0x0002
OUTPUT_TERMINAL 0x0003
TERMINAL_COMPANION 0x0004
MIXER_UNIT 0x0005
SELECTOR_UNIT 0x0006
FEATURE_UNIT 0x0007
EFFECT_UNIT 0x0008
PROCESSING_UNIT 0x0009
EXTENSION_UNIT 0x000A
CLOCK_SOURCE 0x000B
CLOCK_SELECTOR 0x000C
SAMPLE_RATE_CONVERTER 0x000D
CLUSTER 0x000E
CONNECTOR 0x000F
POWER_DOMAIN 0x0010
ENTITY_GROUP 0x0011
COMMIT_GROUP 0x0012
FUNCTION_CONTAINER 0x00FF
AS_SELF 0x0101
AS_VALID_FREQ_RANGE 0x0102
STRING 0x0201
Reserved All Other Values

A.12 DESCRIPTOR VARIANT TYPES

Table A-12: Descriptor Variant Types

Descriptor Variant Type Value

VARIANT_NONE 0x0000
VARIANT_INTERFACE 0x0001
VARIANT_ENTITIES 0x0002
Reserved 0x0003..0xFFFF

A.13 CLUSTER DESCRIPTOR SEGMENT TYPES

Table A-13: Cluster Descriptor Segment Types

Segment Type Value

CLUSTER_SEGMENT_UNDEFINED 0x0000

Release 4.0 April 2023 217

Universal Serial Bus Device Class Definition for Audio Devices

Segment Type Value

Reserved for Common Block Segment Types 0x0001..0x00FF
CHANNEL_INFORMATION 0x0101
CHANNEL_AMBISONIC 0x0102
CHANNEL_DESCRIPTION 0x0103
Reserved for Channel Block Segment Types 0x0104..0x01FF
CLUSTER_END_BLOCK 0xFFFF
Reserved All Other Values

A.14 CHANNEL PURPOSE DEFINITIONS

Table A-14: Channel Purpose Definitions

Channel Purpose Value

PURPOSE_UNDEFINED 0x0000
GENERIC_AUDIO 0x0001
VOICE 0x0002
SPEECH 0x0003
AMBIENT 0x0004
REFERENCE 0x0005
ULTRASONIC 0x0006
VIBROKINETIC 0x0007
SENSE 0x0008
SILENCE 0xFFFE
NON_AUDIO 0xFFFF
Reserved All Other Values

A.15 CHANNEL RELATIONSHIP DEFINITIONS

Table A-15: Channel Relationship Definitions

USB Audio Channel Relationship CEA-861.2 Channel Allocation

Description Acronym Value Description Acronym
RELATIONSHIP_UNDEFINED --- 0x0000 --- ---

Mono M 0x0001 --- ---

Left L 0x0002 --- ---
Right R 0x0003 --- ---
Array AR 0x0004 --- ---
Headset_Mic HM 0x0005 --- ---
Headset_Mic Left HML 0x0006 --- ---
Headset_Mic Right HMR 0x0007 --- ---

Release 4.0 April 2023 218

Universal Serial Bus Device Class Definition for Audio Devices

USB Audio Channel Relationship CEA-861.2 Channel Allocation

Headset_Mic Center HMC 0x0008 --- ---
Body_Mic BM 0x0009 --- ---
Body_Mic Left BML 0x000A --- ---
Body_Mic Right BMR 0x000B --- ---
Body_Mic Center BMC 0x000C --- ---
Lapel_Mic LM 0x000D --- ---
Lapel_Mic Left LML 0x000E --- ---
Lapel_Mic Right LMR 0x000F --- ---
Lapel_Mic Center LMC 0x0010 --- ---
Pattern_X PX 0x0011 --- ---
Pattern _Y PY 0x0012 --- ---
Pattern _A PA 0x0013 --- ---
Pattern _B PB 0x0014 --- ---
Pattern _M PM 0x0015 --- ---
Pattern _S PS 0x0016 --- ---
Reserved --- 0x0017.. --- ---
0x8000
Front Left FL 0x8001 Front Left FL
Front Right FR 0x8002 Front Right FR
Front Center FC 0x8003 Front Center FC
Front Left of Center FLC 0x8004 Front Left of Center FLc
Front Right of Center FRC 0x8005 Front Right of Center FRc
Front Wide Left FWL 0x8006 Front left Wide FLw
Front Wide Right FWR 0x8007 Front Right Wide FRw
Side Left SL 0x8008 Side Left SiL
Side Right SR 0x8009 Side Right SiR
Surround Array Left SAL 0x800A Left Surround LS
Surround Array Right SAR 0x800B Right Surround RS
Back Left BL 0x800C Back Left BL
Back Right BR 0x800D Back Right BR
Back Center BC 0x800E Back Center BC
Back Left of Center BLC 0x800F --- ---
Back Right of Center BRC 0x8010 --- ---
Back Wide Left BWL 0x8011 --- ---
Back Wide Right BWR 0x8012 --- ---
Top Center TC 0x8013 Top Center TpC

Release 4.0 April 2023 219

Universal Serial Bus Device Class Definition for Audio Devices

USB Audio Channel Relationship CEA-861.2 Channel Allocation

Top Front Left TFL 0x8014 Top Front Left TpFL
Top Front Right TFR 0x8015 Top Front Right TpFR
Top Front Center TFC 0x8016 Top Front Center TpFC
Top Front Left of Center TFLC 0x8017 --- ---
Top Front Right of Center TFRC 0x8018 --- ---
Top Front Wide Left TFWL 0x8019 --- ---
Top Front Wide Right TFWR 0x801A --- ---
Top Side Left TSL 0x801B Top Side Left TpSiL
Top Side Right TSR 0x801C Top Side Right TpSiR
Top Surround Array Left TSAL 0x801D Top Left Surround TpLS
Top Surround Array Right TSAR 0x801E Top Right Surround TpRS
Top Back Left TBL 0x801F Top Back Left TpBL
Top Back Right TBR 0x8020 Top Back Right TpBR
Top Back Center TBC 0x8021 Top Back Center TpBC
Top Back Left of Center TBLC 0x8022 --- ---
Top Back Right of Center TBRC 0x8023 --- ---
Top Back Wide Left TBWL 0x8024 --- ---
Top Back Wide Right TBWR 0x8025 --- ---
Bottom Center BC 0x8026 --- ---
Bottom Front Left BFL 0x8027 Bottom Front Left BtFL
Bottom Front Right BFR 0x8028 Bottom Front Right BtFR
Bottom Front Center BFC 0x8029 Bottom Front Center BtFC
Bottom Front Left of Center BFLC 0x802A --- ---
Bottom Front Right of Center BFRC 0x802B --- ---
Bottom Front Wide Left BFWL 0x802C --- ---
Bottom Front Wide Right BFWR 0x802D --- ---
Bottom Side Left BSL 0x802E --- ---
Bottom Side Right BSR 0x802F --- ---
Bottom Surround Array Left BSAL 0x8030 --- ---
Bottom Surround Array Right BSAR 0x8031 --- ---
Bottom Back Left BBL 0x8032 --- ---
Bottom Back Right BBR 0x8033 --- ---
Bottom Back Center BBC 0x8034 --- ---
Bottom Back Left of Center BBLC 0x8035 --- ---
Bottom Back Right of Center BBRC 0x8036 --- ---

Release 4.0 April 2023 220

Universal Serial Bus Device Class Definition for Audio Devices

USB Audio Channel Relationship CEA-861.2 Channel Allocation

Bottom Back Wide Left BBWL 0x8037 --- ---
Bottom Back Wide Right BBWR 0x8038 --- ---
Low Frequency Effects LFE 0x8039 Low Frequency Effects 1 LFE1
Low Frequency Effects Left LFEL 0x803A --- ---
Low Frequency Effects Right LFER 0x803B Low Frequency Effects 2 LFE2
Headphone Left HPL 0x803C --- ---
Headphone Right HPR 0x803D --- ---
Reserved --- 0x803E.. --- ---
0xFFFF

A.16 AMBISONIC COMPONENT ORDERING CONVENTION TYPES

Table A-16: Ambisonic Component Ordering Convention Types

Ambisonic Component Ordering Convention Type Value

ORD_TYPE_UNDEFINED 0x0000
AMBISONIC_CHANNEL_NUMBER (CAN) 0x0001
FURSE_MALHAM 0x0002
SINGLE_INDEX DESIGNATION (SID) 0x0003
Reserved 0x0004..0xFFFF

A.17 AMBISONIC NORMALIZATION TYPES

Table A-17: Ambisonic Normalization Types

Ambisonic Normalization Type Value

NORM_TYPE_UNDEFINED 0x0000
max00N 0x0001
SN3D 0x0002
N3D 0x0003
SN2D 0x0004
N2D 0x0005
Reserved 0x0006..0xFFFF

A.18 TERMINAL COMPANION SEGMENT TYPES

Table A-18: Terminal Companion Segment Types

Segment Type Value

TERMINAL_COMPANION_SEGMENT_UNDEFINED 0x0000
EN_50332-2_SPL_LEVEL 0x0001
EN_50332-2_VOLTAGE_LEVEL 0x0002

Release 4.0 April 2023 221

Universal Serial Bus Device Class Definition for Audio Devices

Segment Type Value

CHANNEL_BANDWIDTH 0x0021
CHANNEL_MAGNITUDE_RESPONSE 0x0022
CHANNEL_MAGNITUDE/PHASE_RESPONSE 0x0023
CHANNEL_POSITION_XYZ 0x0024
CHANNEL_POSITION_ R_THETA_PHI 0x0025
TERMINAL_COMPANION_END_BLOCK 0xFFFF
Reserved All Other Values

A.19 EFFECT UNIT EFFECT TYPES

Table A-19: Effect Unit Effect Types

wEffectType Value
EFFECT_UNDEFINED 0x0000
PARAM_EQ_SECTION_EFFECT 0x0001
REVERBERATION_EFFECT 0x0002
MOD_DELAY_EFFECT 0x0003
DYN_RANGE_COMP_EXP_EFFECT 0x0004
Reserved 0x0005..0xFFFF

A.20 PROCESSING UNIT PROCESS TYPES

Table A-20: Processing Unit Process Types

wProcessType Value
PROCESS_UNDEFINED 0x0000
UP/DOWNMIX_PROCESS 0x0001
CHANNEL_REMAP_PROCESS 0x0002
CLUSTER_MODIFICATION_PROCESS 0x0003
STEREO_EXTENDER_PROCESS 0x0004
MULTI_FUNCTION_PROCESS 0x0005
Reserved 0x0006..0xFFFF

A.21 CLASS-SPECIFIC REQUEST CODES

Table A-21: Class-specific Request Codes

Class-specific Request Code Value

REQUEST_UNDEFINED 0x00
PUSH 0x01
PULL 0x02
COMMIT 0xFE

Release 4.0 April 2023 222

Universal Serial Bus Device Class Definition for Audio Devices

Class-specific Request Code Value

SWITCH_FUNCTION 0xFF
Reserved 0x0003..0xFD

A.22 CLASS-SPECIFIC ATTRIBUTE CODES

Table A-22: Class-specific Attribute Codes

Class-specific Attribute Code Value

ATTRIBUTE_UNDEFINED 0x0000
CUR 0x0001
NEXT 0x0002
RANGE 0x0003
CAP 0x0004
STRING 0x0005
EXTENDED_DESCRIPTOR 0x0006
PAGED_EXTENDED_DESCRIPTOR 0x0007
Reserved 0x0008..0xFFFF

A.23 CONTROL SELECTOR CODES

A.23.1 TERMINAL CONTROL SELECTORS

Table A-23: Terminal Control Selectors

Control Selector Value

TE_CONTROL_UNDEFINED 0x0000
TE_CLUSTER_CONTROL 0x0001
TE_CLUSTER_ACTIVE_CONTROL 0x0002
TE_VOLTAGE_CONTROL 0x0003
TE_OVERLOAD_CONTROL 0x0004
TE_CLIPPING_CONTROL 0x0005
TE_LATENCY_CONTROL 0x0006
Reserved 0x0007..0xFFFF

A.23.2 MIXER UNIT CONTROL SELECTORS

Table A-24: Mixer Unit Control Selectors

Control Selector Value

MU_CONTROL_UNDEFINED 0x0000
MU_MIXER_CONTROL 0x0001
MU_CLUSTER_CONTROL 0x0002
MU_CLUSTER_ACTIVE_CONTROL 0x0003

Release 4.0 April 2023 223

Universal Serial Bus Device Class Definition for Audio Devices

Control Selector Value

MU_LATENCY_CONTROL 0x0004
Reserved 0x0005..0xFFFF

A.23.3 SELECTOR UNIT CONTROL SELECTORS

Table A-25: Selector Unit Control Selectors

Control Selector Value

SU_CONTROL_UNDEFINED 0x0000
SU_SELECTOR_CONTROL 0x0001
SU_LATENCY_CONTROL 0x0002
Reserved 0x0003..0xFFFF

A.23.4 FEATURE UNIT CONTROL SELECTORS

Table A-26: Feature Unit Control Selectors

Control Selector Value

FU_CONTROL_UNDEFINED 0x0000
FU_BYPASS_CONTROL 0x0001
FU_MUTE_CONTROL 0x0002
FU_GAIN_CONTROL 0x0003
FU_BASS_CONTROL 0x0004
FU_MID_CONTROL 0x0005
FU_TREBLE_CONTROL 0x0006
FU_GRAPHIC_EQUALIZER_CONTROL 0x0007
FU_AUTOMATIC_GAIN_CONTROL 0x0008
FU_DELAY_CONTROL 0x0009
FU_BASS_BOOST_CONTROL 0x000A
FU_LOUDNESS_CONTROL 0x000B
FU_INPUT_GAIN_PAD_CONTROL 0x000C
FU_PHASE_INVERTER_CONTROL 0x000D
FU_LATENCY_CONTROL 0x000E
Reserved 0x000F..0xFFFF

A.23.5 SRC UNIT CONTROL SELECTORS

Table A-27: SRC Unit Control Selectors

Control Selector Value

RU_CONTROL_UNDEFINED 0x0000
RU_LATENCY_CONTROL 0x0001

Release 4.0 April 2023 224

Universal Serial Bus Device Class Definition for Audio Devices

Control Selector Value

Reserved 0x0002..0xFFFF

A.23.6 EFFECT UNIT CONTROL SELECTORS

A.23.6.1 PARAMETRIC EQUALIZER SECTION EFFECT UNIT CONTROL SELECTORS

Table A-28: Parametric Equalizer Section Effect Unit Control Selectors

Control Selector Value

PE_CONTROL_UNDEFINED 0x0000
PE_BYPASS_CONTROL 0x0001
PE_CENTERFREQ_CONTROL 0x0002
PE_QFACTOR_CONTROL 0x0003
PE_GAIN_CONTROL 0x0004
PE_LATENCY_CONTROL 0x0005
Reserved 0x0006..0xFFFF

A.23.6.2 REVERBERATION EFFECT UNIT CONTROL SELECTORS

Table A-29: Reverberation Effect Unit Control Selectors

Control Selector Value

RV_CONTROL_UNDEFINED 0x0000
RV_BYPASS_CONTROL 0x0001
RV_TYPE_CONTROL 0x0002
RV_LEVEL_CONTROL 0x0003
RV_TIME_CONTROL 0x0004
RV_FEEDBACK_CONTROL 0x0005
RV_PREDELAY_CONTROL 0x0006
RV_DENSITY_CONTROL 0x0007
RV_HIFREQ_ROLLOFF_CONTROL 0x0008
RV_LATENCY_CONTROL 0x0009
Reserved 0x000A..0xFFFF

A.23.6.3 MODULATION DELAY EFFECT UNIT CONTROL SELECTORS

Table A-30: Modulation Delay Effect Unit Control Selectors

Control Selector Value

MD_CONTROL_UNDEFINED 0x0000
MD_BYPASS_CONTROL 0x0001
MD_BALANCE_CONTROL 0x0002

Release 4.0 April 2023 225

Universal Serial Bus Device Class Definition for Audio Devices

Control Selector Value

MD_RATE_CONTROL 0x0003
MD_DEPTH_CONTROL 0x0004
MD_TIME_CONTROL 0x0005
MD_FEEDBACK_CONTROL 0x0006
MD_LATENCY_CONTROL 0x0007
Reserved 0x0008..0xFFFF

A.23.6.4 DYNAMIC RANGE COMPRESSOR/EXPANDER EFFECT UNIT CONTROL SELECTORS

Table A-31: Dynamic Range Compressor/Expander Effect Unit Control Selectors

Control Selector Value

DR_CONTROL_UNDEFINED 0x0000
DR_BYPASS_CONTROL 0x0001
DR_RATIO_CONTROL 0x0002
DR_MAXAMPL_CONTROL 0x0003
DR_THRESHOLD_CONTROL 0x0004
DR_ATTACK_TIME_CONTROL 0x0005
DR_RELEASE_TIME_CONTROL 0x0006
DR_LATENCY_CONTROL 0x0007
Reserved 0x0008..0xFFFF

A.23.7 PROCESSING UNIT CONTROL SELECTORS

A.23.7.1 UP/DOWN-MIX PROCESSING UNIT CONTROL SELECTORS

Table A-32: Up/Down-mix Processing Unit Control Selectors

Control Selector Value

UD_CONTROL_UNDEFINED 0x0000
UD_BYPASS_CONTROL 0x0001
UD_CLUSTER_CONTROL 0x0002
UD_CLUSTER_ACTIVE_CONTROL 0x0003
UD_LATENCY_CONTROL 0x0004
Reserved 0x0005..0xFFFF

A.23.7.2 CHANNEL REMAP PROCESSING UNIT CONTROL SELECTORS

Table A-33: Channel Remap Processing Unit Control Selectors

Control Selector Value

CR_CONTROL_UNDEFINED 0x0000

Release 4.0 April 2023 226

Universal Serial Bus Device Class Definition for Audio Devices

Control Selector Value

CR_BYPASS_CONTROL 0x0001
CR_CLUSTER_CONTROL 0x0002
CR_CLUSTER_ACTIVE_CONTROL 0x0003
CR_LATENCY_CONTROL 0x0004
Reserved 0x0005..0xFFFF

A.23.7.3 STEREO EXTENDER PROCESSING UNIT CONTROL SELECTORS

Table A-34: Stereo Extender Processing Unit Control Selectors

Control Selector Value

ST_CONTROL_UNDEFINED 0x0000
ST_BYPASS_CONTROL 0x0001
ST_WIDTH_CONTROL 0x0002
ST_LATENCY_CONTROL 0x0003
Reserved 0x0004..0xFFFF

A.23.7.4 MULTI-FUNCTION PROCESSING UNIT CONTROL SELECTORS

Table A-35: Multi-Function Processing Unit Control Selectors

Control Selector Value

MF_CONTROL_UNDEFINED 0x0000
MF_BYPASS_CONTROL 0x0001
MF_CLUSTER_CONTROL 0x0002
MF_CLUSTER_ACTIVE_CONTROL 0x0003
MF_ALGO_PRESENT_CONTROL 0x0004
MF_ALGO_ENABLE_CONTROL 0x0005
MF_LATENCY_CONTROL 0x0006
Reserved 0x0007..0xFFFF

A.23.8 EXTENSION UNIT CONTROL SELECTORS

Table A-36: Extension Unit Control Selectors

Control Selector Value

XU_CONTROL_UNDEFINED 0x0000
XU_BYPASS_CONTROL 0x0001
XU_CLUSTER_CONTROL 0x0002
XU_CLUSTER_ACTIVE_CONTROL 0x0003
XU_LATENCY_CONTROL 0x0004
Reserved 0x0005..0xFFFF

Release 4.0 April 2023 227

Universal Serial Bus Device Class Definition for Audio Devices

A.23.9 CLOCK SOURCE CONTROL SELECTORS

Table A-37: Clock Source Control Selectors

Control Selector Value

CS_CONTROL_UNDEFINED 0x0000
CS_SAM_FREQ_CONTROL 0x0001
CS_CLOCK_VALID_CONTROL 0x0002
Reserved 0x0003..0xFFFF

A.23.10 CLOCK SELECTOR CONTROL SELECTORS

Table A-38: Clock Selector Control Selectors

Control Selector Value

CX_CONTROL_UNDEFINED 0x0000
CX_SELECTOR_CONTROL 0x0001
Reserved 0x0002..0xFFFF

A.23.11 CONNECTOR CONTROL SELECTORS

Table A-39: Connector Control Selectors

Control Selector Value

CO_CONTROL_UNDEFINED 0x0000
CO_INSERTION_CONTROL 0x0001
Reserved 0x0002..0xFFFF

A.23.12 POWER DOMAIN CONTROL SELECTORS

Table A-40: Power Domain Control Selectors

Control Selector Value

PD_CONTROL_UNDEFINED 0x0000
PD_POWER_STATE_CONTROL 0x0001
Reserved 0x0002..0xFFFF

A.23.13 AUDIOSTREAMING INTERFACE CONTROL SELECTORS

Table A-41: AudioStreaming Interface Control Selectors

Control Selector Value

AS_CONTROL_UNDEFINED 0x0000
AS_ACTIVE_ALT_SETTING_CONTROL 0x0001
AS_VALID_ALT_SETTINGS_CONTROL 0x0002
AS_AUDIO_DATA_FORMAT_CONTROL 0x0003

Release 4.0 April 2023 228

Universal Serial Bus Device Class Definition for Audio Devices

Control Selector Value

Reserved 0x0004..0xFFFF

A.24 CONNECTOR TYPES

Table A-42: Connector Types

Connector Type Value

UNDEFINED 0x0000
2.5 MM PHONE CONNECTOR 0x0001
3.5 MM PHONE CONNECTOR 0x0002
6.35 MM PHONE CONNECTOR 0x0003
XLR/6.35MM COMBO CONNECTOR 0x0004
XLR 0x0005
OPTICAL/3.5MM COMBO CONNECTOR 0x0006
RCA 0x0007
BNC 0x0008
BANANA 0x0009
BINDING POST 0x000A
SPEAKON 0x000B
SPRING CLIP 0x000C
SCREW TYPE 0x000D
DIN 0x000E
MINI DIN 0x000F
EUROBLOCK 0x0010
RJ-11 0x0011
RJ-45 0x0012
TOSLINK 0x0013
HDMI 0x0014
Mini-HDMI 0x0015
Micro-HDMI 0x0016
DP 0x0017
MINI-DP 0x0018
D-SUB 0x0019
THUNDERBOLT 0x001A
LIGHTNING 0x001B
WIRELESS 0x001C
USB LEGACY 0x001D
USB-C 0x001E

Release 4.0 April 2023 229

Universal Serial Bus Device Class Definition for Audio Devices

Connector Type Value

SMA 0x001F
SMB 0x0020
LEMO 0x0021
OTHER (CONNECTOR TYPE NOT IN THIS LIST) 0xFFFF
Reserved All Other Values

A.25 AUDIOCONTROL CAPABILITIES OVERVIEW

Table A-43: AudioControl Capabilities Overview

AudioControl Support Comments

AudioControl Interface
Commit CR Required when at least one AudioControl has a NEXT Attribute.
Switch CR Required when the Device support at least one HRL operation
mode.
Terminal
Cluster CR Required when the Input Terminal represents a non-
AudioStreaming Interface.
Cluster Active MAN Input Terminal Only
Overload OPT
Clipping OPT
Latency MAN*
Mixer Unit
Mixer MAN At least one Mixer Control shall be present.
Cluster OPT
Cluster Active MAN
Latency MAN*
Selector Unit
Selector MAN
Latency MAN*
Feature Unit
Bypass OPT
Mute OPT
OPT
Bass OPT
At least one of these AudioControls shall be implemented.
Mid OPT
Treble OPT
Graphic Equalizer OPT

Release 4.0 April 2023 230

Universal Serial Bus Device Class Definition for Audio Devices

AudioControl Support Comments

Automatic Gain OPT
Delay OPT
Bass Boost OPT
Loudness OPT
Input Gain Pad OPT
Phase Inverter OPT
Latency MAN*
SRC Unit
Latency MAN*
PEQ Effect Unit
Bypass OPT
Center Frequency OPT
QFactor OPT At least one of these AudioControls shall be implemented.
Gain OPT
Latency MAN*
Reverberation Effect Unit
Bypass OPT
Type OPT
Level OPT
Time OPT
Delay Feedback OPT At least one of these AudioControls shall be implemented.
Pre-Delay OPT
Density OPT
Hi_Freq Roll-Off OPT
Latency MAN*
Modulation Delay Effect Unit
Bypass OPT
Balance OPT
Rate OPT
Depth OPT At least one of these AudioControls shall be implemented.
Time OPT
Feedback Level OPT
Latency MAN*
Dynamic Range Compressor Effect Unit
Bypass OPT

Release 4.0 April 2023 231

Universal Serial Bus Device Class Definition for Audio Devices

AudioControl Support Comments

Compression Ratio OPT
MaxAmpl OPT
Threshold OPT At least one of these AudioControls shall be implemented.
Attack Time OPT
Release Time OPT
Latency MAN*
Up/Down-mix Processing Unit
Bypass OPT
Cluster OPT
Cluster Active MAN
Latency MAN*
Channel Remap Processing Unit
Bypass OPT
Cluster OPT
Cluster Active MAN
Latency MAN*
Stereo Extender Processing Unit
Bypass OPT
Width MAN
Latency MAN*
Multi-Function Processing Unit
Bypass MAN
Cluster OPT
Cluster Active MAN
Algo Present MAN
Enable Algo OPT
Latency MAN*
Extension Unit
Bypass MAN
Cluster OPT
Cluster Active MAN
Latency MAN*
Clock Source Entity
Sampling Frequency MAN
Clock Valid MAN

Release 4.0 April 2023 232

Universal Serial Bus Device Class Definition for Audio Devices

AudioControl Support Comments

Clock Selector Entity
Selector MAN
Connector Entity
Insertion OPT
Power Domain Entity
Power State MAN
AudioStreaming Interface
Active Alternate Settings OPT
Valid Alternate Settings OPT
AudioStreaming Endpoint
None -

* The Latency Control is either supported on all Terminals and Units or it is not supported anywhere.

A.26 INTERRUPT SOURCE TYPES

Table A-44: Interrupt Source Types

Connector Type Value

SOURCE_TYPE_UNDEFINED 0x0000
AUDIOCONTROL_CUR 0x0001
AUDIOCONTROL_RANGE 0x0002
EXTENDED_DESCRIPTOR 0x0003
CLASS_SPECIFIC_STRING_DESCRIPTOR 0x0004
Reserved 0x0005..0xFFFF

Release 4.0 April 2023 233

Universal Serial Bus Device Class Definition for Audio Devices

APPENDIX B. AUDIO DATA FORMAT CODES

This Appendix list all the Audio Data Format Code constants used by this specification.

B.1 AUDIO DATA FORMATS

Table B-1: Audio Data Formats in the wFormat Field and Usage

Name wFormat Value Type

PCM 0x0000 I
PCM8 0x0001 I
IEEE_FLOAT 0x0002 I
ALAW 0x0003 I
MULAW 0x0004 I
DSD 0x0005 I
Reserved 0x0006..0x00FF I
PCM_IEC60958 0x0100 III
AC-3 0x0101 III
MPEG-1_Layer1 0x0102 III
MPEG-1_Layer2/3 or MPEG-2_NOEXT 0x0103 III
MPEG-2_EXT 0x0104 III
MPEG-2_AAC_ADTS 0x0105 III
MPEG-2_Layer1_LS 0x0106 III
MPEG-2_Layer2/3_LS 0x0107 III
DTS-I 0x0108 III
DTS-II 0x0109 III
DTS-III 0x010A III
ATRAC 0x010B III
ATRAC2/3 0x010C III
WMA 0x010D III
E-AC-3 0x010E III
MAT 0x010F III
DTS-IV 0x0110 III
MPEG-4_HE_AAC 0x0111 III
MPEG-4_HE_AAC_V2 0x0112 III
MPEG-4_AAC_LC 0x0113 III
DRA 0x0114 III
MPEG-4_HE_AAC_SURROUND 0x0115 III

Release 4.0 April 2023 234

Universal Serial Bus Device Class Definition for Audio Devices

Name wFormat Value Type

MPEG-4_AAC_LC_SURROUND 0x0116 III

MPEG-H_3D_AUDIO 0x0117 III
AC4 0x0118 III
MPEG-4_AAC_ELD 0x0119 III
Reserved 0x011A..0x01FF III
Reserved 0x0200..0xFFFF ---

B.2 SUBHEADER CODES

Table B-2: SubHeader Codes

SubHeader Code Value

HEADER_UNDEFINED 0x0000
HDCP_ENCRYPTION 0x0001
Reserved 0x0002..0xFFFF

B.3 AUDIO FORMAT GENERAL CONSTANTS

Table B-3: General Constants

Constant Identifier Value

HDCP_PACKET_HEADER_TIME 512 (ms)

Release 4.0 April 2023 235