ADOBE SYSTEMS INCORPORATED

Corporate Headquarters
345 Park Avenue
San Jose, CA 95110- 2704
(408) 536-6000
http://www.adobe.com
Digital Negative (DNG)
Specification
Version 1.4.0.0
June 2012

Copyright 2004-2012 Adobe Systems Incorporated. All rights reserved.

NOTICE: All information contained herein is the property of Adobe Systems Incorporated. No part of this publication
(whether in hardcopy or electronic form) may be reproduced or transmitted, in any form or by any means, electronic,
mechanical, photocopying, recording, or otherwise, without the prior written consent of Adobe Systems Incorporated.
Adobe, the Adobe logo, and Photoshop are either registered trademarks or trademarks of Adobe Systems Incorporated
in the United States and/or other countries. All other trademarks are the property of their respective owners.
This publication and the information herein is furnished AS IS, is subject to change without notice, and should
not be construed as a commitment by Adobe Systems Incorporated. Adobe Systems Incorporated assumes no
responsibility or liability for any errors or inaccuracies, makes no warranty of any kind (express, implied, or
statutory) with respect to this publication, and expressly disclaims any and all warranties of merchantability,
fitness for particular purposes, and noninfringement of third party rights.
Digital Negative Specification iii
Table of Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
About This Document . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
Audience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
How This Document Is Organized . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
New Information for Version 1.4.0.0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x
Where to Go for More Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x
Chapter 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 11
The Pros and Cons of Raw Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
A Standard Format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
The Advantages of DNG. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Chapter 2 DNG Format Overview . . . . . . . . . . . . . . . . . . . . 13
File Extensions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
SubIFD Trees . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Byte Order . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Masked Pixels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Defective Pixels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Metadata . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Proprietary Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Camera Profiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Opcode Lists . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Floating Point Image Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Transparent Pixels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Proxy DNG Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Chapter 3 Restrictions and Extensions to Existing TIFF Tags . . . . . 18
NewSubFileType . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
BitsPerSample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Digital Negative Specification iv
Table of Contents
SampleFormat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Compression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Predictor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
PhotometricInterpretation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Orientation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Chapter 4 DNG Tags . . . . . . . . . . . . . . . . . . . . . . . . . . 22
DNGVersion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
DNGBackwardVersion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
UniqueCameraModel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
LocalizedCameraModel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
CFAPlaneColor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
CFALayout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
LinearizationTable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
BlackLevelRepeatDim . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
BlackLevel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
BlackLevelDeltaH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
BlackLevelDeltaV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
WhiteLevel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
DefaultScale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
BestQualityScale. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
DefaultCropOrigin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
DefaultCropSize . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
CalibrationIlluminant1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
CalibrationIlluminant2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
ColorMatrix1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
ColorMatrix2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
CameraCalibration1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
CameraCalibration2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
ReductionMatrix1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
ReductionMatrix2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
AnalogBalance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
AsShotNeutral . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
AsShotWhiteXY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Digital Negative Specification v
Table of Contents
BaselineExposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
BaselineNoise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
BaselineSharpness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
BayerGreenSplit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
LinearResponseLimit. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
CameraSerialNumber . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
LensInfo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
ChromaBlurRadius. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
AntiAliasStrength. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
ShadowScale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
DNGPrivateData . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
MakerNoteSafety. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
RawDataUniqueID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
OriginalRawFileName . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
OriginalRawFileData . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
ActiveArea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
MaskedAreas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
AsShotICCProfile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
AsShotPreProfileMatrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
CurrentICCProfile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
CurrentPreProfileMatrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
Additional Tags for Version 1.2.0.0. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
ColorimetricReference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
CameraCalibrationSignature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
ProfileCalibrationSignature. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
ExtraCameraProfiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
AsShotProfileName . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
NoiseReductionApplied . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
ProfileName . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
ProfileHueSatMapDims . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
ProfileHueSatMapData1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
ProfileHueSatMapData2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
ProfileToneCurve. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
ProfileEmbedPolicy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
Digital Negative Specification vi
Table of Contents
ProfileCopyright . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
ForwardMatrix1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
ForwardMatrix2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
PreviewApplicationName. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
PreviewApplicationVersion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
PreviewSettingsName . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
PreviewSettingsDigest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
PreviewColorSpace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
PreviewDateTime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
RawImageDigest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
OriginalRawFileDigest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
SubTileBlockSize. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
RowInterleaveFactor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
ProfileLookTableDims . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
ProfileLookTableData . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Additional Tags for Version 1.3.0.0. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
OpcodeList1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
OpcodeList2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
OpcodeList3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
NoiseProfile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Additional Tags for Version 1.4.0.0. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
DefaultUserCrop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
DefaultBlackRender . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
BaselineExposureOffset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
ProfileLookTableEncoding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
ProfileHueSatMapEncoding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
OriginalDefaultFinalSize . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
OriginalBestQualityFinalSize. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
OriginalDefaultCropSize . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
NewRawImageDigest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
RawToPreviewGain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
Chapter 5 Mapping Raw Values to Linear Reference Values . . . . . . 77
Linearization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Digital Negative Specification vii
Table of Contents
Black Subtraction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Rescaling. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Clipping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
Chapter 6 Mapping Camera Color Space to CIE XYZ Space . . . . . . 79
Camera Calibration Matrices. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
One or Two Color Calibrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
Definitions used in the following sections . . . . . . . . . . . . . . . . . . . . . . . . . . 80
Translating White Balance xy Coordinates to Camera Neutral Coordinates . . . . . . . . . 80
Translating Camera Neutral Coordinates to White Balance xy Coordinates . . . . . . . . . 80
Camera to XYZ (D50) Transform. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
If the ForwardMatrix tags are not included in the camera profile: . . . . . . . . . . . . . . 81
If the ForwardMatrix tags are included in the camera profile: . . . . . . . . . . . . . . . . 82
Applying the Hue/Saturation/Value Mapping Table . . . . . . . . . . . . . . . . . . . . . 82
Chapter 7 Opcode List Processing . . . . . . . . . . . . . . . . . . . 83
WarpRectilinear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
WarpFisheye . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
FixVignetteRadial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
FixBadPixelsConstant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
FixBadPixelsList . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
TrimBounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
MapTable. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
MapPolynomial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
GainMap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
DeltaPerRow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
DeltaPerColumn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
ScalePerRow. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
ScalePerColumn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
Appendix A: Compatibility with Previous Versions . . . . . . . . . . . . 99
Compatibility Issue 1: ActiveArea Tag . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
Compatibility Issue 2: 16-bit Lossless JPEG Encoding . . . . . . . . . . . . . . . . . . . 99
Digital Negative Specification viii
Table of Contents
Compatibility Issue 3: SubTileBlockSize . . . . . . . . . . . . . . . . . . . . . . . . . . .100
Compatibility Issue 4: RowInterleaveFactor . . . . . . . . . . . . . . . . . . . . . . . . .100
Compatibility Issue 5: PreviewColorSpace. . . . . . . . . . . . . . . . . . . . . . . . . .100
Compatibility Issue 6: CFALayout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .100
Compatibility Issue 7: Opcode Lists . . . . . . . . . . . . . . . . . . . . . . . . . . . . .101
Compatibility Issue 8: Floating Point Image Data . . . . . . . . . . . . . . . . . . . . . .101
Compatibility Issue 9: Transparent Pixels . . . . . . . . . . . . . . . . . . . . . . . . . .101
Compatibility Issue 10: Deflate Compression . . . . . . . . . . . . . . . . . . . . . . . .101
Compatibility Issue 11: Lossy JPEG Compression . . . . . . . . . . . . . . . . . . . . .101
Digital Negative Specification September 2012 ix
Preface
About This Document
The Digital Negative (DNG) Specification describes a non-proprietary file format for storing
camera raw files that can be used by a wide range of hardware and software vendors.
This section contains information about this document, including how it is organized and
where to go for additional information.
Audience
This document is intended for developers of hardware and software applications that will
generate, process, manage, or archive camera raw files.
How This Document Is Organized
This document has the following sections:
Chapter 1, Introduction explains what digital negatives are, gives an overview of the
DNG file format, and discusses the advantages of DNG.
Chapter 2, DNG Format Overview provides an overview of the DNG format, including
information on file extensions, SubIFD trees, byte order, masked pixels, defective pixels,
metadata, and proprietary data.
Chapter 3, Restrictions and Extensions to Existing TIFF Tags describes tag differences
between DNG and the TIFF 6.0 format on which DNG is based.
Chapter 4, DNG Tags lists all DNG-specific tags and describes how they are used.
Chapter 5, Mapping Raw Values to Linear Reference Values specifies DNG's processing
model for mapping stored raw sensor values into linear reference values.
Chapter 6, Mapping Camera Color Space to CIE XYZ Space describes DNG's
processing model for mapping between the camera color space coordinates (linear
reference values) and CIE XYZ coordinates.
Chapter 7, Opcode List Processing describes the concept of Opcode Lists, which allow
additional processing steps to be specified in an extensible manner.
Appendix A: Compatibility with Previous Versions documents compatibility between the
current and previous DNG versions.
Digital Negative Specification September 2012 x
Preface
New Information for Version 1.4.0.0
New Information for Version 1.4.0.0
The following changes have been made for the 1.4.0.0 version of this specification:
In Chapter 2, DNG Format Overview, additional tags have been added to the section on
Camera Profiles.
Sections on Floating Point Image Data, Transparent Pixels and Proxy DNG Files have been
added to Chapter 2, DNG Format Overview.
In Chapter 3, Restrictions and Extensions to Existing TIFF Tags, the section on
Compression has been updated.
New sections on SampleFormat and Predictor have been added to Chapter 3, Restrictions and
Extensions to Existing TIFF Tags.
The WhiteLevel tag has been updated in Chapter 4, DNG Tags.
Additional Tags were added to Chapter 4, DNG Tags for Version 1.4.0.0.
Appendix A: Compatibility with Previous Versions was updated.
Where to Go for More Information
DNG is an extension of TIFF 6.0 and is compatible with the TIFF-EP standard. See these
specifications for more information on TIFF and TIFF-EP:
TIFF 6.0 Specification, Adobe Systems, Inc.,
1992-06-03.
http://partners.adobe.com/public/developer/en/tiff/TIFF6.pdf
TIFF/EP Specification, ISO/DIS 12234-2,
2001-10-15.
http://www.iso.org
Adobe Photoshop TIFF Technical Notes http://partners.adobe.com/public/developer/en/tiff/TIFFphoto
shop.pdf
Adobe Photoshop TIFF Technical Note 3 http://chriscox.org/TIFFTN3d1.pdf
Digital Negative Specification September 2012 11
1
Introduction
The Pros and Cons of Raw Data
Seeking a greater degree of flexibility and artistic control, professional photographers
increasingly opt to manipulate raw data from their digital cameras. Unlike JPEG and TIFF
formats which store images that have been processed by the camera, camera raw files capture
unprocessed or minimally processed data directly from the camera sensor. Because they are
analogous to film negatives in a photographers workflow, camera raw formats are often
referred to as digital negatives.
Camera raw formats offer both advantages and disadvantages. One advantage is increased
artistic control for the end user. The user can precisely adjust a range of parameters, including
white balance, tone mapping, noise reduction, sharpening and others, to achieve a desired
look.
One disadvantage is that unlike JPEG and TIFF files which are ready for immediate use,
camera raw files must be processed before they can be used, typically through software
provided by the camera manufacturer or through a converter like the Camera Raw plug-in for
Adobe Photoshop software.
The challenge for end users and camera vendors alike is that there is no publicly-documented
and supported format for storing raw camera data. Every camera manufacturer that supports
raw data must create their own proprietary format, along with software for converting the
proprietary format into the standard JPEG and/or TIFF formats.
A Standard Format
The lack of a standard format for camera raw files creates additional work for camera
manufacturers because they need to develop proprietary formats along with the software to
process them. It also poses risks for end users. Camera raw formats vary from camera to
camera, even those produced by the same manufacturer. It is not uncommon for a camera
manufacturer to terminate support for a discontinued cameras raw format. This means users
have no guarantee they will be able to open archived camera raw files in the future.
To address these problems, Adobe has defined a new non-proprietary format for camera raw
files. The format, called Digital Negative or DNG, can be used by a wide range of hardware
and software developers to provide a more flexible raw processing and archiving workflow.
End users may also use DNG as an intermediate format for storing images that were originally
captured using a proprietary camera raw format.
Digital Negative Specification September 2012 12
Introduction
The Advantages of DNG
1
The Advantages of DNG
DNG has all the benefits of current camera raw formats; namely, increased flexibility and
artistic control. In addition, DNG offers several new advantages over proprietary camera raw
formats.
Self-Contained
With the current proprietary camera raw formats, software programs wishing to process
camera raw files must have specific information about the camera that created the file. As new
camera models are released, software manufacturers (and by extension users) must update
their software to accommodate the new camera raw formats.
Because DNG metadata is publicly documented, software readers such as the Adobe
Photoshop Camera Raw plug-in do not need camera-specific knowledge to decode and process
files created by a camera that supports DNG. That means reduced software maintenance and a
more self-contained solution for end users.
Archival
Camera manufacturers sometimes drop support for a propriety raw format a few years after a
camera is discontinued. Without continued software support, users may not be able to access
images stored in proprietary raw formats and the images may be lost forever. Since DNG is
publicly documented, it is far more likely that raw images stored as DNG files will be readable
by software in the distant future, making DNG a safer choice for archival.
TIFF Compatible
DNG is an extension of the TIFF 6.0 format, and is compatible with the TIFF-EP standard. It
is possible (but not required) for a DNG file to simultaneously comply with both the Digital
Negative specification and the TIFF-EP standard.
Digital Negative Specification September 2012 13
2
DNG Format Overview
This section describes the DNG format. As an extension of the TIFF 6.0 format, DNG should
follow all the formatting rules for TIFF 6.0. For more information, refer to the TIFF 6.0
specification.
The following topics are discussed in this section:
File Extensions
SubIFD Trees
Byte Order
Masked Pixels
Defective Pixels
Metadata
Proprietary Data
Camera Profiles
Opcode Lists
Floating Point Image Data
Transparent Pixels
Proxy DNG Files
File Extensions
The recommended file extension for Digital Negative is ".DNG". Readers should accept either
the ".DNG" or ".TIF" extensions for compatibility with TIFF-EP.
SubIFD Trees
DNG recommends the use of SubIFD trees, as described in the TIFF-EP specification. SubIFD
chains are not supported.
The highest-resolution and quality IFD should use NewSubFileType equal to 0. Reduced
resolution (or quality) thumbnails or previews, if any, should use NewSubFileType equal to 1
(for a primary preview) or 10001.H (for an alternate preview).
DNG recommends, but does not require, that the first IFD contain a low-resolution thumbnail,
as described in the TIFF-EP specification.
Digital Negative Specification September 2012 14
DNG Format Overview
Byte Order
2
Byte Order
DNG readers are required to support either byte order, even for files from a particular camera
model. Writers can write either byte order, whichever is easier and/or faster for the writer.
Masked Pixels
Most camera sensors measure the black encoding level using fully-masked pixels at the edges
of the sensor. These pixels can either be trimmed before storing the image in DNG, or they can
be included in the stored image. If the masked pixels are not trimmed, the area of the non-
masked pixels must be specified using the ActiveArea tag.
The black encoding level information extracted from these masked pixels should be used to
either pre-compensate the raw data stored in the file or they should be included in the file using
the DNG tags for specifying the black level.
This black encoding level information is required even if the masked pixels are not trimmed, to
allow DNG readers to process the image without requiring knowledge of the best way to
compute the black levels for any given camera model.
Defective Pixels
There are two ways to deal with defective pixels in DNG. The first is to map out (interpolate
over) the defective pixels before storing the raw data in DNG. The second is to include a bad
pixel fixing opcode in the OpcodeList1 tag.
Metadata
Additional metadata may be embedded in DNG in the following ways:
Using TIFF-EP or EXIF metadata tags
Using the IPTC metadata tag (33723)
Using the XMP metadata tag (700)
Note that TIFF-EP and EXIF use nearly the same metadata tag set, but TIFF-EP stores the tags
in IFD 0, while EXIF store the tags in a separate IFD. Either location is allowed by DNG, but
the EXIF location is preferred.
Digital Negative Specification September 2012 15
DNG Format Overview
Proprietary Data
2
Proprietary Data
Camera manufacturers may want to include proprietary data in a raw file for use by their own
raw converter. DNG allows proprietary data to be stored using private tags, private IFDs,
and/or a private MakerNote.
It is recommended that manufacturers use the DNGPrivateData and MakerNoteSafety tags to
ensure that programs that edit DNG files preserve this proprietary data. See Chapter 4, DNG
Tags on page 22 for more information on the DNGPrivateData and MakerNoteSafety tags.
Camera Profiles
DNG 1.2.0.0 and later formalizes the concept of a camera profile and allows multiple
camera profiles to be embedded in a single DNG file. A camera profile consists of a set of tags
(both existing in DNG versions earlier than 1.2.0.0 and newly defined in DNG version
1.2.0.0), some of which are optional.
The set of tags belonging to a camera profile includes the following:
BaselineExposureOffset
CalibrationIlluminant1
CalibrationIlluminant2
ColorMatrix1
ColorMatrix2
DefaultBlackRender
ForwardMatrix1
ForwardMatrix1
ProfileCalibrationSignature
ProfileCopyright
ProfileEmbedPolicy
ProfileHueSatMapData1
ProfileHueSatMapData2
ProfileHueSatMapDims
ProfileHueSatMapEncoding
ProfileLookTableData
ProfileLookTableDims
ProfileLookTableEncoding
Digital Negative Specification September 2012 16
DNG Format Overview
Opcode Lists
2
ProfileName
ProfileToneCurve
ReductionMatrix1
ReductionMatrix2
The primary camera profile is stored in IFD 0, as it was for DNG versions earlier than 1.2.0.0.
This allows backward compatibility with older DNG readers. DNG allows additional camera
profiles to be embedded using the ExtraCameraProfiles tag, which points to a list of Camera
Profile IFDs.
Opcode Lists
DNG 1.3.0.0 and later incorporates the concept of Opcode Lists, which allow additional
processing steps to be specified in an extensible manner.
This allows complex processing to be moved off the camera hardware, which often has limited
processing power, and into the DNG reader, which is often running on more powerful
hardware.
This also allows processing steps to be specified, such as lens corrections, which ideally
should be performed on the image data after it has been demosaiced, while still retaining the
advantages of a raw mosaic data format.
The set of tags controlling this feature are:
OpcodeList1
OpcodeList2
OpcodeList3
Floating Point Image Data
DNG 1.4.0.0 and later support floating point image data, in addition to the previous support for
unsigned integer image data.
Floating point bit depths of 16, 24 and 32 bits per sample are supported.
Transparent Pixels
DNG 1.4.0.0 and later incorporates the concept of Transparent Pixels, which allow the level
of transparency of pixels to be specified for the image data. This is useful when not all of the
Digital Negative Specification September 2012 17
DNG Format Overview
Proxy DNG Files
2
pixels in the image bounding rectangle are defined. For example, after combining several
source images into a panorama, the edge of the resulting combined image is often not perfectly
rectangular.
Transparency is supported by allowing masks to be stored in IFDs with the same width and
length as the image IFDs, but with a NewSubFileType of 4 or 5 (see the section on
NewSubFileType for details).
Fully transparent (undefined) pixels are stored with a mask value of zero. Fully opaque
(defined) pixels are stored with a mask value of 1.0 for floating point masks, and the maximum
possible unsigned integer for integer masks. Intermediate values are allowed for semi-
transparent pixels.
For semi-transparent pixels, the image values should not be matted to either white or black, but
instead stored with the same value as if the pixel was fully opaque.
Proxy DNG Files
DNG 1.4.0.0 adds the concept of proxy DNG files, which are lower resolution (and/or quality)
versions of original DNG files. These proxy DNGs are useful files as placeholders for original
DNGs in applications that have limited bandwidth or available storage space.
Proxy DNGs have information on the pixel size of the original DNG file to allow rendering
parameters to be transferred back to the original DNGs with similar visual results.
The size of the corresponding original DNG file is specified using the
OriginalDefaultFinalSize, OriginalBestQualitySize and OriginalCropSize tags.
Also added in DNG 1.4.0.0 is a new lossy JPEG compression option that is particularity useful
for proxy DNGs, greatly reducing their file size.
Digital Negative Specification September 2012 18
3
Restrictions and Extensions to
Existing TIFF Tags
This section describes the restrictions and extension to the following TIFF tags:
NewSubFileType
BitsPerSample
Compression
PhotometricInterpretation
Orientation
NewSubFileType
In DNG versions earlier than 1.2.0.0, full resolution raw images should use NewSubFileType
equal to 0. Rendered previews or reduced resolution versions of raw images should use
NewSubFileType equal to 1. DNG 1.2.0.0 allows a new value for NewSubFileType equal to
10001.H.
This value, used for alternative or non-primary rendered previews, allows for multiple
renderings (not just multiple sizes of a single rendering) to be stored in a DNG file. DNG
reading software that displays a preview for a DNG file should, by default, display a preview
from an IFD with NewSubFileType equal to 1. Alternative renderings should only be
displayed if requested by the user.
DNG 1.4.0.0 allows two new values for NewSubFileType. Transparency information for the
full resolution raw image should use NewSubFileType equal to 4. Transparency information
for reduced resolution versions of raw images should use NewSubFileType equal to 5.
BitsPerSample
Supported values are from 8 to 32 bits/sample. The depth must be the same for each sample if
SamplesPerPixel is not equal to 1. If BitsPerSample is not equal to 8 or 16 or 32, then the bits
must be packed into bytes using the TIFF default FillOrder of 1 (big-endian), even if the TIFF
file itself uses little-endian byte order.
Digital Negative Specification September 2012 19
Restrictions and Extensions to Existing TIFF Tags
SampleFormat
3
SampleFormat
DNG versions earlier than 1.4.0.0 always use the default SampleFormat value of 1 (unsigned
integer). DNG version 1.4.0.0 allows an additional SampleFormat value of 3 (IEEE floating
point). The allowed BitsPerSample values for floating point images are 16, 24 and 32.
Compression
Two Compression tag values are supported in DNG versions before 1.4.0.0:
Value = 1: Uncompressed data.
Value = 7: JPEG compressed data, either baseline DCT JPEG, or lossless JPEG
compression.
If PhotometricInterpretation = 6 (YCbCr) and BitsPerSample = 8/8/8, or if
PhotometricInterpretation = 1 (BlackIsZero) and BitsPerSample = 8, then the JPEG variant
must be baseline DCT JPEG.
Otherwise, the JPEG variant must be lossless Huffman JPEG. For lossless JPEG, the internal
width/length/components in the JPEG stream are not required to match the strip or tile's
width/length/components. Only the total sample counts need to match. It is common for CFA
images to be encoded with a different width, length or component count to allow the JPEG
compression predictors to work across like colors.
DNG Version 1.4.0.0 adds support for the following compression codes:
Value = 8: Deflate (ZIP)
Value = 34892: Lossy JPEG
Deflate (8) compression is allowed for floating point image data, 32-bit integer image data and
transparency mask data.
Lossy JPEG (34892) is allowed for IFDs that use PhotometricInterpretation = 34892
(LinearRaw) and 8-bit integer data. This new compression code is required to let the DNG
reader know to use a lossy JPEG decoder rather than a lossless JPEG decoder for this
combination of PhotometricInterpretation and BitsPerSample.
Predictor
The following predictor codes are allowed for IFDs using compression code 8 (deflate):
1 = Null Predictor. No predictor is used as part of the compression.
2 = Horizontal Difference. Documented in the TIFF 6.0 specification.
Digital Negative Specification September 2012 20
Restrictions and Extensions to Existing TIFF Tags
PhotometricInterpretation
3
3 = Floating Point. Documented in Adobe Photoshop TIFF Technical Note 3.
34892 = Horizontal Difference X2. Same as Horizontal Difference except the pixel
two to the left is used rather than the pixel one to the left.
34893 = Horizontal Difference X4. Same as Horizontal Difference except the pixel
four to the left is used rather than the pixel one to the left.
34894 = Floating Point X2. Same as Floating Point except the pixel two to the left is
used rather than the pixel one to the left.
34895 = Floating Point X4. Same as Floating Point except the pixel four to the left
is used rather than the pixel one to the left.
PhotometricInterpretation
The following values are supported for thumbnail and preview IFDs only:
1 = BlackIsZero. Assumed to be in a gamma 2.2 color space, unless otherwise
specified using PreviewColorSpace tag.
2 = RGB. Assumed to be in the sRGB color space, unless otherwise specified using
the PreviewColorSpace tag.
6 = YCbCr. Used for JPEG encoded preview images.
The following values are supported for the raw IFD, and are assumed to be the camera's native
color space:
32803 = CFA (Color Filter Array).
34892 = LinearRaw.
The CFA PhotometricInterpretation value is documented in the TIFF-EP specification. Its use
requires the use of the CFARepeatPatternDim and CFAPattern tags in the same IFD. The
origin of the repeating CFA pattern is the top-left corner of the ActiveArea rectangle.
The LinearRaw PhotometricInterpretation value is intended for use by cameras that do not use
color filter arrays, but instead capture all color components at each pixel. It can also be used
for CFA data that has already been de-mosaiced.
The LinearRaw value can be used in reduced resolution IFDs, even if the raw IFD uses the
CFA PhotometricInterpretation value.
Orientation
Orientation is a required tag for DNG. With the Orientation tag present, file browsers can
perform lossless rotation of DNG files by modifying a single byte of the file. DNG readers
Digital Negative Specification September 2012 21
Restrictions and Extensions to Existing TIFF Tags
Orientation
3
should support all possible orientations, including mirrored orientations. Note that the
mirrored orientations are not allowed by the TIFF-EP specification, so writers should not use
them if they want their files be compatible with both specifications.
Digital Negative Specification September 2012 22
4
DNG Tags
This section describes DNG-specific tags. Note that the tags listed here are not part of the
TIFF-EP specification.
DNGVersion
Tag 50706 (C612.H)
Type BYTE
Count 4
Value See below
Default Required tag
Usage IFD 0
Description
This tag encodes the DNG four-tier version number. For files compliant with this version of
the DNG specification (1.4.0.0), this tag should contain the bytes: 1, 4, 0, 0.
DNGBackwardVersion
Tag 50707 (C613.H)
Type BYTE
Count 4
Value See below
Default DNGVersion with the last two
bytes set to zero.
Usage IFD 0
Digital Negative Specification September 2012 23
DNG Tags
UniqueCameraModel
4
Description
This tag specifies the oldest version of the Digital Negative specification for which a file is
compatible. Readers should not attempt to read a file if this tag specifies a version number that
is higher than the version number of the specification the reader was based on.
In addition to checking the version tags, readers should, for all tags, check the types, counts,
and values, to verify it is able to correctly read the file.
For more information on compatibility with previous DNG versions, see Appendix A:
Compatibility with Previous Versions.
UniqueCameraModel
Tag 50708 (C614.H)
Type ASCII
Count String length including null
Value Null terminated string
Default Required tag
Usage IFD 0
Description
UniqueCameraModel defines a unique, non-localized name for the camera model that created
the image in the raw file. This name should include the manufacturer's name to avoid conflicts,
and should not be localized, even if the camera name itself is localized for different markets
(see LocalizedCameraModel).
This string may be used by reader software to index into per-model preferences and
replacement profiles.
Examples of unique model names are:
"Canon EOS 300D"
"Fujifilm FinePix S2Pro"
"Kodak ProBack645"
"Minolta DiMAGE A1"
"Nikon D1X"
"Olympus C-5050Z"
"Pentax *istD"
"Sony F828"
Digital Negative Specification September 2012 24
DNG Tags
LocalizedCameraModel
4
LocalizedCameraModel
Tag 50709 (C615.H)
Type ASCII or BYTE
Count Byte count including null
Value Null terminated UTF-8 encoded
Unicode string
Default Same as UniqueCameraModel
Usage IFD 0
Description
Similar to the UniqueCameraModel field, except the name can be localized for different
markets to match the localization of the camera name.
CFAPlaneColor
Tag 50710 (C616.H)
Type BYTE
Count ColorPlanes
Value See below
Default 0, 1, 2 (red, green, blue)
Usage Raw IFD
Description
CFAPlaneColor provides a mapping between the values in the CFAPattern tag and the plane
numbers in LinearRaw space. This is a required tag for non-RGB CFA images.
Digital Negative Specification September 2012 25
DNG Tags
CFALayout
4
CFALayout
Tag 50711 (C617.H)
Type SHORT
Count 1
Value See below
Default 1
Usage Raw IFD
Description
CFALayout describes the spatial layout of the CFA. The currently defined values are:
1 = Rectangular (or square) layout
2 = Staggered layout A: even columns are offset down by 1/2 row
3 = Staggered layout B: even columns are offset up by 1/2 row
4 = Staggered layout C: even rows are offset right by 1/2 column
5 = Staggered layout D: even rows are offset left by 1/2 column
6 = Staggered layout E: even rows are offset up by 1/2 row, even columns are offset left by
1/2 column
7 = Staggered layout F: even rows are offset up by 1/2 row, even columns are offset right by
1/2 column
8 = Staggered layout G: even rows are offset down by 1/2 row, even columns are offset left
by 1/2 column
9 = Staggered layout H: even rows are offset down by 1/2 row, even columns are offset
right by 1/2 column
Note that for the purposes of this tag, rows and columns are numbered starting with one.
Layout values 6 through 9 were added with DNG version 1.3.0.0.
Digital Negative Specification September 2012 26
DNG Tags
LinearizationTable
4
LinearizationTable
Tag 50712 (C618.H)
Type SHORT
Count N
Value See below
Default Identity table (0, 1, 2, 3, etc.)
Usage Raw IFD
Description
LinearizationTable describes a lookup table that maps stored values into linear values. This tag
is typically used to increase compression ratios by storing the raw data in a non-linear, more
visually uniform space with fewer total encoding levels.
If SamplesPerPixel is not equal to one, this single table applies to all the samples for each
pixel.
See Chapter 5, Mapping Raw Values to Linear Reference Values on page 77 for details of
the processing model.
BlackLevelRepeatDim
Tag 50713 (C619.H)
Type SHORT
Count 2
Value Value 0: BlackLevelRepeatRows
Value 1: BlackLevelRepeatCols
Default 1, 1
Usage Raw IFD
Description
This tag specifies repeat pattern size for the BlackLevel tag.
Digital Negative Specification September 2012 27
DNG Tags
BlackLevel
4
BlackLevel
Tag 50714 (C61A.H)
Type SHORT or LONG or
RATIONAL
Count BlackLevelRepeatRows *
BlackLevelRepeatCols *
SamplesPerPixel
Value See below
Default 0
Usage Raw IFD
Description
This tag specifies the zero light (a.k.a. thermal black or black current) encoding level, as a
repeating pattern. The origin of this pattern is the top-left corner of the ActiveArea rectangle.
The values are stored in row-column-sample scan order.
See Chapter 5, Mapping Raw Values to Linear Reference Values on page 77 for details of
the processing model.
BlackLevelDeltaH
Tag 50715 (C61B.H)
Type SRATIONAL
Count ActiveArea width
Value See below
Default All zeros
Usage Raw IFD
Description
If the zero light encoding level is a function of the image column, BlackLevelDeltaH specifies
the difference between the zero light encoding level for each column and the baseline zero
light encoding level.
If SamplesPerPixel is not equal to one, this single table applies to all the samples for each
pixel.
Digital Negative Specification September 2012 28
DNG Tags
BlackLevelDeltaV
4
See Chapter 5, Mapping Raw Values to Linear Reference Values on page 77 for details of
the processing model.
BlackLevelDeltaV
Tag 50716 (C61C.H)
Type SRATIONAL
Count ActiveArea length
Value See below
Default All zeros
Usage Raw IFD
Description
If the zero light encoding level is a function of the image row, this tag specifies the difference
between the zero light encoding level for each row and the baseline zero light encoding level.
If SamplesPerPixel is not equal to one, this single table applies to all the samples for each
pixel.
See Chapter 5, Mapping Raw Values to Linear Reference Values on page 77 for details of
the processing model.
Digital Negative Specification September 2012 29
DNG Tags
WhiteLevel
4
WhiteLevel
Tag 50717 (C61D.H)
Type SHORT or LONG
Count SamplesPerPixel
Value See below
Default See below
Usage Raw IFD
Description
This tag specifies the fully saturated encoding level for the raw sample values. Saturation is
caused either by the sensor itself becoming highly non-linear in response, or by the camera's
analog to digital converter clipping.
The default value for this tag is (2 ** BitsPerSample) -1 for unsigned integer images, and 1.0
for floating point images.
See Chapter 5, Mapping Raw Values to Linear Reference Values on page 77 for details of
the processing model.
DefaultScale
Tag 50718 (C61E.H)
Type RATIONAL
Count 2
Value Value 0: DefaultScaleH
Value 1: DefaultScaleV
Default 1.0, 1.0
Usage Raw IFD
Description
DefaultScale is required for cameras with non-square pixels. It specifies the default scale
factors for each direction to convert the image to square pixels. Typically these factors are
selected to approximately preserve total pixel count.
Digital Negative Specification September 2012 30
DNG Tags
BestQualityScale
4
For CFA images that use CFALayout equal to 2, 3, 4, or 5, such as the Fujifilm SuperCCD,
these two values should usually differ by a factor of 2.0.
BestQualityScale
Tag 50780 (C65C.H)
Type RATIONAL
Count 1
Value see below
Default 1.0
Usage Raw IFD
Description
For some cameras, the best possible image quality is not achieved by preserving the total pixel
count during conversion. For example, Fujifilm SuperCCD images have maximum detail when
their total pixel count is doubled.
This tag specifies the amount by which the values of the DefaultScale tag need to be multiplied
to achieve the best quality image size.
DefaultCropOrigin
Tag 50719 (C61F.H)
Type SHORT or LONG or
RATIONAL
Count 2
Value Value 0: DefaultCropOriginH
Value 1: DefaultCropOriginV
Default 0, 0
Usage Raw IFD
Description
Raw images often store extra pixels around the edges of the final image. These extra pixels
help prevent interpolation artifacts near the edges of the final image.
Digital Negative Specification September 2012 31
DNG Tags
DefaultCropSize
4
DefaultCropOrigin specifies the origin of the final image area, in raw image coordinates (i.e.,
before the DefaultScale has been applied), relative to the top-left corner of the ActiveArea
rectangle.
DefaultCropSize
Tag 50720 (C620.H)
Type SHORT or LONG or
RATIONAL
Count 2
Value Value 0: DefaultCropSizeH
Value 1: DefaultCropSizeV
Default ImageWidth, ImageLength
Usage Raw IFD
Description
Raw images often store extra pixels around the edges of the final image. These extra pixels
help prevent interpolation artifacts near the edges of the final image.
DefaultCropSize specifies the size of the final image area, in raw image coordinates (i.e.,
before the DefaultScale has been applied).
CalibrationIlluminant1
Tag 50778 (C65A.H)
Type SHORT
Count 1
Value See below
Default 0 (unknown)
Usage IFD 0 or Camera Profile IFD
Description
The illuminant used for the first set of color calibration tags. The legal values for this tag are
the same as the legal values for the LightSource EXIF tag.
Digital Negative Specification September 2012 32
DNG Tags
CalibrationIlluminant2
4
See Chapter 6, Mapping Camera Color Space to CIE XYZ Space on page 79 for details of
the color-processing model.
CalibrationIlluminant2
Tag 50779 (C65B.H)
Type SHORT
Count 1
Value See below
Default None
Usage IFD 0 or Camera Profile IFD
Description
The illuminant used for an optional second set of color calibration tags. The legal values for
this tag are the same as the legal values for the CalibrationIlluminant1 tag; however, if both are
included, neither is allowed to have a value of 0 (unknown).
See Chapter 6, Mapping Camera Color Space to CIE XYZ Space on page 79 for details of
the color-processing model.
ColorMatrix1
Tag 50721 (C621.H)
Type SRATIONAL
Count ColorPlanes * 3
Value See below
Default None
Usage IFD 0 or Camera Profile IFD
Description
ColorMatrix1 defines a transformation matrix that converts XYZ values to reference camera
native color space values, under the first calibration illuminant. The matrix values are stored in
row scan order.
Digital Negative Specification September 2012 33
DNG Tags
ColorMatrix2
4
The ColorMatrix1 tag is required for all non-monochrome DNG files.
See Chapter 6, Mapping Camera Color Space to CIE XYZ Space on page 79 for details of
the color-processing model.
ColorMatrix2
Tag 50722 (C622.H)
Type SRATIONAL
Count ColorPlanes * 3
Value See below
Default None
Usage IFD 0 or Camera Profile IFD
Description
ColorMatrix2 defines a transformation matrix that converts XYZ values to reference camera
native color space values, under the second calibration illuminant. The matrix values are stored
in row scan order.
See Chapter 6, Mapping Camera Color Space to CIE XYZ Space on page 79 for details of
the color-processing model.
Digital Negative Specification September 2012 34
DNG Tags
CameraCalibration1
4
CameraCalibration1
Tag 50723 (C623.H)
Type SRATIONAL
Count ColorPlanes * ColorPlanes
Value See below
Default Identity matrix
Usage IFD 0
Description
CameraCalibration1 defines a calibration matrix that transforms reference camera native space
values to individual camera native space values under the first calibration illuminant. The
matrix is stored in row scan order.
This matrix is stored separately from the matrix specified by the ColorMatrix1 tag to allow
raw converters to swap in replacement color matrices based on UniqueCameraModel tag,
while still taking advantage of any per-individual camera calibration performed by the camera
manufacturer.
See Chapter 6, Mapping Camera Color Space to CIE XYZ Space on page 79 for details of
the color-processing model.
CameraCalibration2
Tag 50724 (C624.H)
Type SRATIONAL
Count ColorPlanes * ColorPlanes
Value See below
Default Identity matrix
Usage IFD 0
Description
CameraCalibration2 defines a calibration matrix that transforms reference camera native space
values to individual camera native space values under the second calibration illuminant. The
matrix is stored in row scan order.
Digital Negative Specification September 2012 35
DNG Tags
ReductionMatrix1
4
This matrix is stored separately from the matrix specified by the ColorMatrix2 tag to allow
raw converters to swap in replacement color matrices based on UniqueCameraModel tag,
while still taking advantage of any per-individual camera calibration performed by the camera
manufacturer.
See Chapter 6, Mapping Camera Color Space to CIE XYZ Space on page 79 for details of
the color-processing model.
ReductionMatrix1
Tag 50725 (C625.H)
Type SRATIONAL
Count 3 * ColorPlanes
Value See below
Default None
Usage IFD 0 or Camera Profile IFD
Description
ReductionMatrix1 defines a dimensionality reduction matrix for use as the first stage in
converting color camera native space values to XYZ values, under the first calibration
illuminant. This tag may only be used if ColorPlanes is greater than 3. The matrix is stored in
row scan order.
See Chapter 6, Mapping Camera Color Space to CIE XYZ Space on page 79 for details of
the color-processing model.
Digital Negative Specification September 2012 36
DNG Tags
ReductionMatrix2
4
ReductionMatrix2
Tag 50726 (C626.H)
Type SRATIONAL
Count 3 * ColorPlanes
Value See below
Default None
Usage IFD 0 or Camera Profile IFD
Description
ReductionMatrix2 defines a dimensionality reduction matrix for use as the first stage in
converting color camera native space values to XYZ values, under the second calibration
illuminant. This tag may only be used if ColorPlanes is greater than 3. The matrix is stored in
row scan order.
See Chapter 6, Mapping Camera Color Space to CIE XYZ Space on page 79 for details of
the color-processing model.
AnalogBalance
Tag 50727 (C627.H)
Type RATIONAL
Count ColorPlanes
Value See below
Default All 1.0
Usage IFD 0
Description
Normally the stored raw values are not white balanced, since any digital white balancing will
reduce the dynamic range of the final image if the user decides to later adjust the white
balance; however, if camera hardware is capable of white balancing the color channels before
the signal is digitized, it can improve the dynamic range of the final image.
AnalogBalance defines the gain, either analog (recommended) or digital (not recommended)
that has been applied the stored raw values.
Digital Negative Specification September 2012 37
DNG Tags
AsShotNeutral
4
See Chapter 6, Mapping Camera Color Space to CIE XYZ Space on page 79 for details of
the color-processing model.
AsShotNeutral
Tag 50728 (C628.H)
Type SHORT or RATIONAL
Count ColorPlanes
Value See below
Default None
Usage IFD 0
Description
AsShotNeutral specifies the selected white balance at time of capture, encoded as the
coordinates of a perfectly neutral color in linear reference space values. The inclusion of this
tag precludes the inclusion of the AsShotWhiteXY tag.
AsShotWhiteXY
Tag 50729 (C629.H)
Type RATIONAL
Count 2
Value See below
Default None
Usage IFD 0
Description
AsShotWhiteXY specifies the selected white balance at time of capture, encoded as x-y
chromaticity coordinates. The inclusion of this tag precludes the inclusion of the
AsShotNeutral tag.
Digital Negative Specification September 2012 38
DNG Tags
BaselineExposure
4
BaselineExposure
Tag 50730 (C62A.H)
Type SRATIONAL
Count 1
Value See below
Default 0.0
Usage IFD 0
Description
Camera models vary in the trade-off they make between highlight headroom and shadow
noise. Some leave a significant amount of highlight headroom during a normal exposure. This
allows significant negative exposure compensation to be applied during raw conversion, but
also means normal exposures will contain more shadow noise. Other models leave less
headroom during normal exposures. This allows for less negative exposure compensation, but
results in lower shadow noise for normal exposures.
Because of these differences, a raw converter needs to vary the zero point of its exposure
compensation control from model to model. BaselineExposure specifies by how much (in EV
units) to move the zero point. Positive values result in brighter default results, while negative
values result in darker default results.
BaselineNoise
Tag 50731 (C62B.H)
Type RATIONAL
Count 1
Value See below
Default 1.0
Usage IFD 0
Description
BaselineNoise specifies the relative noise level of the camera model at a baseline ISO value of
100, compared to a reference camera model.
Digital Negative Specification September 2012 39
DNG Tags
BaselineSharpness
4
Since noise levels tend to vary approximately with the square root of the ISO value, a raw
converter can use this value, combined with the current ISO, to estimate the relative noise level
of the current image.
BaselineSharpness
Tag 50732 (C62C.H)
Type RATIONAL
Count 1
Value See below
Default 1.0
Usage IFD 0
Description
BaselineSharpness specifies the relative amount of sharpening required for this camera model,
compared to a reference camera model. Camera models vary in the strengths of their anti-
aliasing filters. Cameras with weak or no filters require less sharpening than cameras with
strong anti-aliasing filters.
BayerGreenSplit
Tag 50733 (C62D.H)
Type LONG
Count 1
Value See below
Default 0
Usage Raw IFD
Description
BayerGreenSplit only applies to CFA images using a Bayer pattern filter array. This tag
specifies, in arbitrary units, how closely the values of the green pixels in the blue/green rows
track the values of the green pixels in the red/green rows.
Digital Negative Specification September 2012 40
DNG Tags
LinearResponseLimit
4
A value of zero means the two kinds of green pixels track closely, while a non-zero value
means they sometimes diverge. The useful range for this tag is from 0 (no divergence) to about
5000 (quite large divergence).
LinearResponseLimit
Tag 50734 (C62E.H)
Type RATIONAL
Count 1
Value See below
Default 1.0
Usage IFD 0
Description
Some sensors have an unpredictable non-linearity in their response as they near the upper limit
of their encoding range. This non-linearity results in color shifts in the highlight areas of the
resulting image unless the raw converter compensates for this effect.
LinearResponseLimit specifies the fraction of the encoding range above which the response
may become significantly non-linear.
CameraSerialNumber
Tag 50735 (C62F.H)
Type ASCII
Count String length including null
Value Null terminated string
Default None
Usage IFD 0
Description
CameraSerialNumber contains the serial number of the camera or camera body that captured
the image.
Digital Negative Specification September 2012 41
DNG Tags
LensInfo
4
LensInfo
Tag 50736 (C630.H)
Type RATIONAL
Count 4
Value Value 0: Minimum focal length in mm.
Value 1: Maximum focal length in mm.
Value 2: Minimum (maximum aperture) f-stop at minimum focal length.
Value 3: Minimum (maximum aperture) f-stop at maximum focal length.
Default none
Usage IFD 0
Description
LensInfo contains information about the lens that captured the image. If the minimum f-stops
are unknown, they should be encoded as 0/0.
ChromaBlurRadius
Tag 50737 (C631.H)
Type RATIONAL
Count 1
Value Chroma blur radius in pixels
Default See below
Usage Raw IFD
Description
ChromaBlurRadius provides a hint to the DNG reader about how much chroma blur should be
applied to the image. If this tag is omitted, the reader will use its default amount of chroma
blurring.
Normally this tag is only included for non-CFA images, since the amount of chroma blur
required for mosaic images is highly dependent on the de-mosaic algorithm, in which case the
DNG reader's default value is likely optimized for its particular de-mosaic algorithm.
Digital Negative Specification September 2012 42
DNG Tags
AntiAliasStrength
4
AntiAliasStrength
Tag 50738 (C632.H)
Type RATIONAL
Count 1
Value Relative strength of the camera's
anti-alias filter
Default 1.0
Usage Raw IFD
Description
AntiAliasStrength provides a hint to the DNG reader about how strong the camera's anti-alias
filter is. A value of 0.0 means no anti-alias filter (i.e., the camera is prone to aliasing artifacts
with some subjects), while a value of 1.0 means a strong anti-alias filter (i.e., the camera
almost never has aliasing artifacts).
Note that this tag overlaps in functionality with the BaselineSharpness tag. The primary
difference is the AntiAliasStrength tag is used as a hint to the de-mosaic algorithm, while the
BaselineSharpness tag is used as a hint to a sharpening algorithm applied later in the
processing pipeline.
ShadowScale
Tag 50739 (C633.H)
Type RATIONAL
Count 1
Value See below
Default 1.0
Usage IFD 0
Description
This tag is used by Adobe Camera Raw to control the sensitivity of its "Shadows" slider.
Digital Negative Specification September 2012 43
DNG Tags
DNGPrivateData
4
DNGPrivateData
Tag 50740 (C634.H)
Type BYTE
Count Length of private data block in
bytes
Value See below
Default None
Usage IFD 0
Description
DNGPrivateData provides a way for camera manufacturers to store private data in the DNG
file for use by their own raw converters, and to have that data preserved by programs that edit
DNG files.
The private data must follow these rules:
The private data must start with a null-terminated ASCII string identifying the data. The
first part of this string must be the manufacturer's name, to avoid conflicts between
manufacturers.
The private data must be self-contained. All offsets within the private data must be offsets
relative to the start of the private data, and must not point to bytes outside the private data.
The private data must be byte-order independent. If a DNG file is converted from a big-
endian file to a little-endian file, the data must remain valid.
Digital Negative Specification September 2012 44
DNG Tags
MakerNoteSafety
4
MakerNoteSafety
Tag 50741 (C635.H)
Type SHORT
Count 1
Value 0 (unsafe) or 1 (safe)
Default 0
Usage IFD 0
Description
MakerNoteSafety lets the DNG reader know whether the EXIF MakerNote tag is safe to
preserve along with the rest of the EXIF data. File browsers and other image management
software processing an image with a preserved MakerNote should be aware that any thumbnail
image embedded in the MakerNote may be stale, and may not reflect the current state of the
full size image.
A MakerNote is safe to preserve if it follows these rules:
The MakerNote data must be self-contained. All offsets within the MakerNote must be
offsets relative to the start of the MakerNote, and must not point to bytes outside the
MakerNote.
The MakerNote data must be byte-order independent. Moving the data to a file with a
different byte order must not invalidate it.
RawDataUniqueID
Tag 50781 (C65D.H)
Type BYTE
Count 16
Value See below
Default Optional
Usage IFD 0
Description
Digital Negative Specification September 2012 45
DNG Tags
OriginalRawFileName
4
This tag contains a 16-byte unique identifier for the raw image data in the DNG file. DNG
readers can use this tag to recognize a particular raw image, even if the file's name or the
metadata contained in the file has been changed.
If a DNG writer creates such an identifier, it should do so using an algorithm that will ensure
that it is very unlikely two different images will end up having the same identifier.
OriginalRawFileName
Tag 50827 (C68B)
Type ASCII or BYTE
Count Byte count including null
Value Null terminated UTF-8 encoded
Unicode string
Default Optional
Usage IFD 0
Description
If the DNG file was converted from a non-DNG raw file, then this tag contains the file name of
that original raw file.
OriginalRawFileData
Tag 50828 (C68C.H)
Type UNDEFINED
Count Byte count of embedded data
Value See below
Default Optional
Usage IFD 0
Description
If the DNG file was converted from a non-DNG raw file, then this tag contains the compressed
contents of that original raw file.
Digital Negative Specification September 2012 46
DNG Tags
OriginalRawFileData
4
The contents of this tag always use the big-endian byte order.
The tag contains a sequence of data blocks. Future versions of the DNG specification may
define additional data blocks, so DNG readers should ignore extra bytes when parsing this tag.
DNG readers should also detect the case where data blocks are missing from the end of the
sequence, and should assume a default value for all the missing blocks.
There are no padding or alignment bytes between data blocks. The sequence of data blocks is:
1. Compressed data fork of original raw file.
2. Compressed Mac OS resource fork of original raw file.
3. Mac OS file type (4 bytes) of original raw file.
4. Mac OS file creator (4 bytes) of original raw file.
5. Compressed data fork of sidecar ".THM" file.
6. Compressed Mac OS resource fork of sidecar ".THM" file.
7. Mac OS file type (4 bytes) of sidecar ".THM" file.
8. Mac OS file creator (4 bytes) of sidecar ".THM" file.
If the Mac OS file types or creator codes are unknown, zero is stored.
If the Mac OS resource forks do not exist, they should be encoded as zero length forks.
Each fork (data or Mac OS resource) is compressed and encoded as:
ForkLength = first four bytes. This is the uncompressed length of this fork. If this value is
zero, then no more data is stored for this fork.
From ForkLength, compute the number of 64K compression blocks used for this data (the last
block is usually smaller than 64K):
ForkBlocks = Floor ((ForkLength + 65535) / 65536)
The next (ForkBlocks + 1) 4-byte values are an index into the compressed data. The first
ForkBlock values are offsets from the start of the data for this fork to the start of the
compressed data for the corresponding compression block. The last value is an offset from the
start of the data for this fork to the end of the data for this fork.
Following this index is the ZIP compressed data for each 64K compression block.
Digital Negative Specification September 2012 47
DNG Tags
ActiveArea
4
ActiveArea
Tag 50829 (C68D.H)
Type SHORT or LONG
Count 4
Value See below
Default 0, 0, ImageLength, ImageWidth
Usage Raw IFD
Description
This rectangle defines the active (non-masked) pixels of the sensor. The order of the rectangle
coordinates is: top, left, bottom, right.
MaskedAreas
Tag 50830 (C68E)
Type SHORT or LONG
Count 4 * number of rectangles
Value See below
Default None
Usage Raw IFD
Description
This tag contains a list of non-overlapping rectangle coordinates of fully masked pixels, which
can be optionally used by DNG readers to measure the black encoding level.
The order of each rectangle's coordinates is: top, left, bottom, right.
If the raw image data has already had its black encoding level subtracted, then this tag should
not be used, since the masked pixels are no longer useful.
Note that DNG writers are still required to include estimate and store the black encoding level
using the black level DNG tags. Support for the MaskedAreas tag is not required of DNG
readers.
Digital Negative Specification September 2012 48
DNG Tags
AsShotICCProfile
4
AsShotICCProfile
Tag 50831 (C68F.H)
Type UNDEFINED
Count Length of ICC profile in bytes
Value See below
Default Optional
Usage IFD 0
Description
This tag contains an ICC profile that, in conjunction with the AsShotPreProfileMatrix tag,
provides the camera manufacturer with a way to specify a default color rendering from camera
color space coordinates (linear reference values) into the ICC profile connection space.
The ICC profile connection space is an output referred colorimetric space, whereas the other
color calibration tags in DNG specify a conversion into a scene referred colorimetric space.
This means that the rendering in this profile should include any desired tone and gamut
mapping needed to convert between scene referred values and output referred values.
DNG readers that have their own tone and gamut mapping controls (such as Adobe Camera
Raw) will probably ignore this tag pair.
Digital Negative Specification September 2012 49
DNG Tags
AsShotPreProfileMatrix
4
AsShotPreProfileMatrix
Tag 50832 (C690.H)
Type SRATIONAL
Count 3 * ColorPlanes or ColorPlanes *
ColorPlanes
Value See below
Default Identity matrix
Usage IFD 0
Description
This tag is used in conjunction with the AsShotICCProfile tag. It specifies a matrix that should
be applied to the camera color space coordinates before processing the values through the ICC
profile specified in the AsShotICCProfile tag.
The matrix is stored in the row scan order.
If ColorPlanes is greater than three, then this matrix can (but is not required to) reduce the
dimensionality of the color data down to three components, in which case the
AsShotICCProfile should have three rather than ColorPlanes input components.
CurrentICCProfile
Tag 50833 (C691.H)
Type UNDEFINED
Count Length of ICC profile in bytes
Value See below
Default Optional
Usage IFD 0
Description
This tag is used in conjunction with the CurrentPreProfileMatrix tag.
The CurrentICCProfile and CurrentPreProfileMatrix tags have the same purpose and usage as
the AsShotICCProfile and AsShotPreProfileMatrix tag pair, except they are for use by raw file
editors rather than camera manufacturers.
Digital Negative Specification September 2012 50
DNG Tags
CurrentPreProfileMatrix
4
CurrentPreProfileMatrix
Tag 50834 (C692.H)
Type SRATIONAL
Count 3 * ColorPlanes or ColorPlanes *
ColorPlanes
Value See below
Default Identity matrix
Usage IFD 0
Description
This tag is used in conjunction with the CurrentICCProfile tag.
The CurrentICCProfile and CurrentPreProfileMatrix tags have the same purpose and usage as
the AsShotICCProfile and AsShotPreProfileMatrix tag pair, except they are for use by raw file
editors rather than camera manufacturers.
Additional Tags for Version 1.2.0.0
The following tags have been added for the 1.2.0.0 version of this specification.
ColorimetricReference
Tag 50879 (C6BF.H)
Type SHORT
Count 1
Value 0 or 1
Default 0
Usage IFD 0
Description
The DNG color model documents a transform between camera colors and CIE XYZ values.
This tag describes the colorimetric reference for the CIE XYZ values.
Digital Negative Specification September 2012 51
DNG Tags
CameraCalibrationSignature
4
0 = The XYZ values are scene-referred.
1 = The XYZ values are output-referred, using the ICC profile perceptual dynamic range.
This tag allows output-referred data to be stored in DNG files and still processed correctly by
DNG readers.
CameraCalibrationSignature
Tag 50931 (C6F3.H)
Type ASCII or BYTE
Count Length of string including null
Value Null terminated string
Default Empty string
Usage IFD 0
Description
A UTF-8 encoded string associated with the CameraCalibration1 and CameraCalibration2
tags. The CameraCalibration1 and CameraCalibration2 tags should only be used in the DNG
color transform if the string stored in the CameraCalibrationSignature tag exactly matches the
string stored in the ProfileCalibrationSignature tag for the selected camera profile.
ProfileCalibrationSignature
Tag 50932 (C6F4.H)
Type ASCII or BYTE
Count Length of string including null
Value Null terminated string
Default Empty string
Usage IFD 0 or Camera Profile IFD
A UTF-8 encoded string associated with the camera profile tags. The CameraCalibration1 and
CameraCalibration2 tags should only be used in the DNG color transfer if the string stored in
the CameraCalibrationSignature tag exactly matches the string stored in the
ProfileCalibrationSignature tag for the selected camera profile.
Digital Negative Specification September 2012 52
DNG Tags
ExtraCameraProfiles
4
ExtraCameraProfiles
Tag 50933 (C6F5.H)
Type LONG
Count Number of extra camera profiles
Value Offsets to Camera Profile IFDs
Default Empty list
Usage IFD 0
Description
A list of file offsets to extra Camera Profile IFDs. The format of a camera profile begins with a
16-bit byte order mark (MM or II) followed by a 16-bit "magic" number equal to 0x4352
(CR), a 32-bit IFD offset, and then a standard TIFF format IFD. All offsets are relative to the
start of the byte order mark. Note that the primary camera profile tags should be stored in IFD
0, and the ExtraCameraProfiles tag should only be used if there is more than one camera
profile stored in the DNG file.
AsShotProfileName
Tag 50934 (C6F6.H)
Type ASCII or BYTE
Count Length of string including null
Value Null terminated string
Default Optional
Usage IFD 0
Description
A UTF-8 encoded string containing the name of the "as shot" camera profile, if any.
Digital Negative Specification September 2012 53
DNG Tags
NoiseReductionApplied
4
NoiseReductionApplied
Tag 50935 (C6F7.H)
Type RATIONAL
Count 1
Value See below
Default 0/0
Usage Raw IFD
Description
This tag indicates how much noise reduction has been applied to the raw data on a scale of 0.0
to 1.0. A 0.0 value indicates that no noise reduction has been applied. A 1.0 value indicates
that the "ideal" amount of noise reduction has been applied, i.e. that the DNG reader should
not apply additional noise reduction by default. A value of 0/0 indicates that this parameter is
unknown.
ProfileName
Tag 50936 (C6F8.H)
Type ASCII or BYTE
Count Length of string including null
Value Null terminated string
Default None
Usage IFD 0 or Camera Profile IFD
Description
A UTF-8 encoded string containing the name of the camera profile. This tag is optional if there
is only a single camera profile stored in the file but is required for all camera profiles if there is
more than one camera profile stored in the file.
Digital Negative Specification September 2012 54
DNG Tags
ProfileHueSatMapDims
4
ProfileHueSatMapDims
Tag 50937 (C6F9.H)
Type LONG
Count 3
Value HueDivisions
SaturationDivisions
ValueDivisions
Default None
Usage IFD 0 or Camera Profile IFD
Description
This tag specifies the number of input samples in each dimension of the hue/saturation/value
mapping tables. The data for these tables are stored in ProfileHueSatMapData1 and
ProfileHueSatMapData2 tags. Allowed values include the following:
HueDivisions >= 1
SaturationDivisions >= 2
ValueDivisions >=1
The most common case has ValueDivisions equal to 1, so only hue and saturation are used as
inputs to the mapping table.
Digital Negative Specification September 2012 55
DNG Tags
ProfileHueSatMapData1
4
ProfileHueSatMapData1
Tag 50938 (C6FA.H)
Type FLOAT
Count HueDivisions *
SaturationDivisions *
ValueDivisions * 3
Value See below
Default None
Usage IFD 0 or Camera Profile IFD
Description
This tag contains the data for the first hue/saturation/value mapping table. Each entry of the
table contains three 32-bit IEEE floating-point values. The first entry is hue shift in degrees;
the second entry is saturation scale factor; and the third entry is a value scale factor. The table
entries are stored in the tag in nested loop order, with the value divisions in the outer loop, the
hue divisions in the middle loop, and the saturation divisions in the inner loop. All zero input
saturation entries are required to have a value scale factor of 1.0. The hue/saturation/value
table application is described in detail in Chapter 6.
ProfileHueSatMapData2
Tag 50939 (C6FB.H)
Type FLOAT
Count HueDivisions *
SaturationDivisions *
ValueDivisions * 3
Value See below
Default None
Usage IFD 0 or Camera Profile IFD
Description
This tag contains the data for the second hue/saturation/value mapping table. Each entry of the
table contains three 32-bit IEEE floating-point values. The first entry is hue shift in degrees;
Digital Negative Specification September 2012 56
DNG Tags
ProfileToneCurve
4
the second entry is a saturation scale factor; and the third entry is a value scale factor. The table
entries are stored in the tag in nested loop order, with the value divisions in the outer loop, the
hue divisions in the middle loop, and the saturation divisions in the inner loop. All zero input
saturation entries are required to have a value scale factor of 1.0. The hue/saturation/value
table application is described in detail in Chapter 6.
ProfileToneCurve
Tag 50940 (C6FC.H)
Type FLOAT
Count Samples * 2
Value See below
Default None
Usage IFD 0 or Camera Profile IFD
Description
This tag contains a default tone curve that can be applied while processing the image as a
starting point for user adjustments. The curve is specified as a list of 32-bit IEEE floating-
point value pairs in linear gamma. Each sample has an input value in the range of 0.0 to 1.0,
and an output value in the range of 0.0 to 1.0. The first sample is required to be (0.0, 0.0), and
the last sample is required to be (1.0, 1.0). Interpolated the curve using a cubic spline.
Digital Negative Specification September 2012 57
DNG Tags
ProfileEmbedPolicy
4
ProfileEmbedPolicy
Tag 50941 (C6FD.H)
Type LONG
Count 1
Value See below
Default 0
Usage IFD 0 or Camera Profile IFD
Description
This tag contains information about the usage rules for the associated camera profile. The valid
values and meanings are:
0 = allow copying. The camera profile can be used to process, or be embedded in, any
DNG file. It can be copied from DNG files to other DNG files, or copied from DNG files
and stored on the users system for use in processing or embedding in any DNG file. The
camera profile may not be used to process non-DNG files.
1 = embed if used. This value applies the same rules as allow copying, except it does
not allow copying the camera profile from a DNG file for use in processing any image
other than the image in which it is embedded, unless the profile is already stored on the
users system.
2 = embed never. This value only applies to profiles stored on a users system but not
already embedded in DNG files. These stored profiles can be used to process images but
cannot be embedded in files. If a camera profile is already embedded in a DNG file, then
this value has the same restrictions as embed if used.
3 = no restrictions. The camera profile creator has not placed any restrictions on the use
of the camera profile.
Digital Negative Specification September 2012 58
DNG Tags
ProfileCopyright
4
ProfileCopyright
Tag 50942 (C6FE.H)
Type ASCII or BYTE
Count Length of string including null
Value Null terminated string
Default Optional
Usage IFD 0 or Camera Profile IFD
Description
A UTF-8 encoded string containing the copyright information for the camera profile. This
string always should be preserved along with the other camera profile tags.
ForwardMatrix1
Tag 50964 (C714.H)
Type SRATIONAL
Count 3 * ColorPlanes
Value See below
Default Optional
Usage IFD 0 or Camera Profile IFD
Description
This tag defines a matrix that maps white balanced camera colors to XYZ D50 colors.
Application is described in detail in Chapter 6.
Digital Negative Specification September 2012 59
DNG Tags
ForwardMatrix2
4
ForwardMatrix2
Tag 50965 (C715.H)
Type SRATIONAL
Count 3 * ColorPlanes
Value See below
Default Optional
Usage IFD 0 or Camera Profile IFD
Description
This tag defines a matrix that maps white balanced camera colors to XYZ D50 colors.
Application is described in detail in Chapter 6.
PreviewApplicationName
Tag 50966 (C716.H)
Type ASCII or BYTE
Count Length of string including null
Value Null terminated string
Default Optional
Usage Preview IFD
Description
A UTF-8 encoded string containing the name of the application that created the preview stored
in the IFD.
Digital Negative Specification September 2012 60
DNG Tags
PreviewApplicationVersion
4
PreviewApplicationVersion
Tag 50967 (C717.H)
Type ASCII or BYTE
Count Length of string including null
Value Null terminated string
Default Optional
Usage Preview IFD
Description
A UTF-8 encoded string containing the version number of the application that created the
preview stored in the IFD.
PreviewSettingsName
Tag 50968 (C718.H)
Type ASCII or BYTE
Count Length of string including null
Value Null terminated string
Default Optional
Usage Preview IFD
Description
A UTF-8 encoded string containing the name of the conversion settings (for example, snapshot
name) used for the preview stored in the IFD.
Digital Negative Specification September 2012 61
DNG Tags
PreviewSettingsDigest
4
PreviewSettingsDigest
Tag 50969 (C719.H)
Type BYTE
Count 16
Value See below
Default Optional
Usage Preview IFD
Description
A unique ID of the conversion settings (for example, MD5 digest) used to render the preview
stored in the IFD.
PreviewColorSpace
Tag 50970 (C71A.H)
Type LONG
Count 1
Value See below
Default See below
Usage Preview IFD
Description
This tag specifies the color space in which the rendered preview in this IFD is stored. The valid
values include:
0 = Unknown
1 = Gray Gamma 2.2
2 = sRGB
3 = Adobe RGB
4 = ProPhoto RGB
The default value for this tag is sRGB for color previews and Gray Gamma 2.2 for
monochrome previews.
Digital Negative Specification September 2012 62
DNG Tags
PreviewDateTime
4
PreviewDateTime
Tag 50971 (C71B.H)
Type ASCII
Count Length of string including null
Value Null terminated string
Default Optional
Usage Preview IFD
Description
This tag is an ASCII string containing the name of the date/time at which the preview stored in
the IFD was rendered. The date/time is encoded using ISO 8601 format.
RawImageDigest
Tag 50972 (C71C.H)
Type BYTE
Count 16
Value See below
Default Optional
Usage IFD 0
Description
This tag is an MD5 digest of the raw image data. All pixels in the image are processed in row-
scan order. Each pixel is zero padded to 16 or 32 bits deep (16-bit for data less than or equal to
16 bits deep, 32-bit otherwise). The data for each pixel is processed in little-endian byte order.
Digital Negative Specification September 2012 63
DNG Tags
OriginalRawFileDigest
4
OriginalRawFileDigest
Tag 50973 (C71D.H)
Type BYTE
Count 16
Value See below
Default Optional
Usage IFD 0
Description
This tag is an MD5 digest of the data stored in the OriginalRawFileData tag.
SubTileBlockSize
Tag 50974 (C71E.H)
Type SHORT or LONG
Count 2
Value SubTileBlockRows
SubTileBlockCols
Default 1, 1
Usage Raw IFD
Description
Normally, the pixels within a tile are stored in simple row-scan order. This tag specifies that
the pixels within a tile should be grouped first into rectangular blocks of the specified size.
These blocks are stored in row-scan order. Within each block, the pixels are stored in row-scan
order. The use of a non-default value for this tag requires setting the DNGBackwardVersion
tag to at least 1.2.0.0.
Digital Negative Specification September 2012 64
DNG Tags
RowInterleaveFactor
4
RowInterleaveFactor
Tag 50975 (C71F.H)
Type SHORT or LONG
Count 1
Value RowInterleaveFactor
Default 1
Usage Raw IFD
Description
This tag specifies that rows of the image are stored in interleaved order. The value of the tag
specifies the number of interleaved fields. The use of a non-default value for this tag requires
setting the DNGBackwardVersion tag to at least 1.2.0.0.
ProfileLookTableDims
Tag 50981 (C725.H)
Type LONG
Count 3
Value HueDivisions
SaturationDivisions
ValueDivisions
Default none
Usage IFD 0 or Camera Profile IFD
Description
This tag specifies the number of input samples in each dimension of a default "look" table. The
data for this table is stored in the ProfileLookTableData tag. Allowed values include:
HueDivisions >= 1
SaturationDivisions >= 2
ValueDivisions >= 1
Digital Negative Specification September 2012 65
DNG Tags
ProfileLookTableData
4
ProfileLookTableData
Tag 50982 (C726.H)
Type FLOAT
Count HueDivisions *
SaturationDivisions *
ValueDivisions * 3
Value See below
Default none
Usage IFD 0 or Camera Profile IFD
Description
This tag contains a default "look" table that can be applied while processing the image as a
starting point for user adjustment. This table uses the same format as the tables stored in the
ProfileHueSatMapData1 and ProfileHueSatMapData2 tags, and is applied in the same color
space. However, it should be applied later in the processing pipe, after any exposure
compensation and/or fill light stages, but before any tone curve stage.
Each entry of the table contains three 32-bit IEEE floating-point values. The first entry is hue
shift in degrees, the second entry is a saturation scale factor, and the third entry is a value scale
factor.
The table entries are stored in the tag in nested loop order, with the value divisions in the outer
loop, the hue divisions in the middle loop, and the saturation divisions in the inner loop.
All zero input saturation entries are required to have a value scale factor of 1.0.
Digital Negative Specification September 2012 66
DNG Tags
Additional Tags for Version 1.3.0.0
4
Additional Tags for Version 1.3.0.0
The following tags have been added for the 1.3.0.0 version of this specification.
OpcodeList1
Tag 51008 (C740.H)
Type UNDEFINED
Count Variable
Value Opcode List
Default Empty List
Usage Raw IFD
Description
Specifies the list of opcodes that should be applied to the raw image, as read directly from the
file. The format and processing details of an opcode list are described in Chapter 7, "Opcode
List Processing".
OpcodeList2
Tag 51009 (C741.H)
Type UNDEFINED
Count Variable
Value Opcode List
Default Empty List
Usage Raw IFD
Description
Specifies the list of opcodes that should be applied to the raw image, just after it has been
mapped to linear reference values. The format and processing details of an opcode list are
described in Chapter 7, "Opcode List Processing".
Digital Negative Specification September 2012 67
DNG Tags
OpcodeList3
4
OpcodeList3
Tag 51022 (C74E.H)
Type UNDEFINED
Count Variable
Value Opcode List
Default Empty List
Usage Raw IFD
Description
Specifies the list of opcodes that should be applied to the raw image, just after it has been
demosaiced. The format and processing details of an opcode list are described in Chapter 7,
"Opcode List Processing".
NoiseProfile
Tag 51041 (C761.H)
Type DOUBLE
Count 2 or 2 * ColorPlanes
Value See Below
Default Values are estimated from
BaselineNoise tag (see below).
Usage Raw IFD
Description
NoiseProfile describes the amount of noise in a raw image. Specifically, this tag models the
amount of signal-dependent photon (shot) noise and signal-independent sensor readout noise,
two common sources of noise in raw images. The model assumes that the noise is white and
spatially independent, ignoring fixed pattern effects and other sources of noise (e.g., pixel
response non-uniformity, spatially-dependent thermal effects, etc.).
This tag is intended to be used to describe the amount of noise present in unprocessed raw
image data. When noise reduction has already been applied to the raw data
(i.e., NoiseReductionApplied 0 > ), this tag may be used to estimate the white component of
the residual noise.
Digital Negative Specification September 2012 68
DNG Tags
NoiseProfile
4
For the purposes of this tag, noise is defined as the standard deviation of a random variable x,
where xrepresents a recorded linear signal in the range x 01 [ , ] e . The two-parameter noise
model is
N
i
x ( ) S
i
x O
i
+ =
for parameters (S
i
O
i
, ), where S
i
is a scale term that models the amount of sensor
amplification, and O
i
is an offset term that models the amount of sensor readout noise. A more
detailed explanation of this model is given below.
The data elements for this tag are the n sets of noise model parameters:
S
1
O
1
S
2
O
2
. S
n
O
n
, , , , , ,
Note that n must be 1 (i.e., tag count is 2) or equal to the number of color planes in the image
(i.e., tag count is 2 ColorPlanes ). When n 1 = , the two specified parameters (S
1
O
1
, )
define the same noise model for all image planes. Whenn is equal to the number of image
planes, the parameters (S
i
O
i
, ) define the noise model for the i th image plane, e.g., (S
1
O
1
, )
correspond to the first image plane, (S
2
O
2
, ) correspond to the second image plane, etc. The
order of the parameters follows the plane order specified by the CFAPlaneColor tag.
Each S
i
term must be positive (S
i
0 > ), and each O
i
term must be non-negative (O
i
0 > ).
A BaselineNoise tag value of 1.0 at ISO 100 corresponds approximately to NoiseProfile
parameter values of S
i
2 10
5
= and O
i
4.5 10
7
= (e.g., standard deviation of ~ 0.00201
when x 0.18 = ); these values may be used to estimate absolute noise levels in an image
when the NoiseProfile tag is missing. When both tags are present, however, DNG readers
should prefer using the NoiseProfile data, since it describes noise levels more precisely than
BaselineNoise.
A more detailed description of the noise model is given below. This tag models two common
sources of noise:
1. Photon (shot) noise p, which has a white Poisson distribution, and
2. Electronic readout noise r, which is present even in the absence of light and is assumed to
have an approximately white normal (Gaussian) distribution.
Assuming that p and r are independent random variables, the square of the total noise (i.e.,
the variance) can be expressed as the sum of the squares of the individual sources of noise:
N
2
p
2
r
2
+ = 1 ( )
In this expression, the variables N, p, and r are expressed in B-bit recorded digital values,
where common values of B include 12, 14, and 16 bits. If x

is the average signal level

Digital Negative Specification September 2012 69
DNG Tags
NoiseProfile
4
expressed in photons, then its variance will also be x

, since a random variable with a Poisson

distribution has a variance equal to its mean:
p

2
x

= 2 ( )
wherep

denotes the photon noise, expressed in photons. The conversion factor between
photons (x

, ) and B-bit digital values (x p , ) is the gain factor g:

x g x

= 3 ( )
p g p

= 4 ( )
Substituting equations 2, 3, and 4 into equation 1 yields:
N
2
g p

( )
2
r
2
+ =
g
2
x

( ) r
2
+ =
g x ( ) r
2
+ =
Therefore the total noiseN can be expressed as a two-parameter function of the signal x:
N x ( ) g x r
2
+ =
S
i
s O
i
+ =
for model parameters S
i
g O
i
r
2
= , = .
This tag uses the convention of a normalized noise model, i.e., N x ( ) is the standard deviation
(i.e., noise) of a random variable x, where x represents a recorded linear signal in the range
x 01 [ , ] e . The specified parameters (S
i
O
i
, ) must also be appropriately normalized.
Digital Negative Specification September 2012 70
DNG Tags
Additional Tags for Version 1.4.0.0
4
Additional Tags for Version 1.4.0.0
The following tags have been added for the 1.4.0.0 version of this specification.
DefaultUserCrop
Tag 51125 (C7B5.H)
Type RATIONAL
Count 4
Value Value 0: 0.0
Value 1: 0.0
Value 2: 1.0
Value 3: 1.0
Default Value 0: 0.0
Value 1: 0.0
Value 2: 1.0
Value 3: 1.0
Usage Raw IFD
Description
Specifies a default user crop rectangle in relative coordinates.
The values must satisfy:
0.0 <= top < bottom <= 1.0
0.0 <= left < right <= 1.0
The default values of (top = 0, left = 0, bottom = 1, right = 1) correspond exactly to the default
crop rectangle (as specified by the DefaultCropOrigin and DefaultCropSize tags).
Digital Negative Specification September 2012 71
DNG Tags
DefaultBlackRender
4
DefaultBlackRender
Tag 51110 (C7A6.H)
Type LONG
Count 1
Value See below
Default 0
Usage IFD 0 or Camera Profile IFD
Description
This optional tag in a color profile provides a hint to the raw converter regarding how to handle
the black point (e.g., flare subtraction) during rendering. The currently defined values are:
0 = Auto
1 = None
If set to Auto, the raw converter should perform black subtraction during rendering. The
amount and method of black subtraction may be automatically determined and may be image-
dependent.
If set to None, the raw converter should not perform any black subtraction during rendering.
This may be desirable when using color lookup tables (e.g., LookTable) or tone curves in
camera profiles to perform a fixed, consistent level of black subtraction.
BaselineExposureOffset
Tag 51109 (C7A5.H)
Type RATIONAL
Count 1
Value See below
Default 0.0
Usage IFD 0 or Camera Profile IFD
Description
Provides a way for color profiles to increase or decrease exposure during raw conversion.
Digital Negative Specification September 2012 72
DNG Tags
ProfileLookTableEncoding
4
BaselineExposureOffset specifies the amount (in EV units) to add to the BaselineExposure tag
during image rendering. For example, if the BaselineExposure value for a given camera model
is +0.3, and the BaselineExposureOffset value for a given camera profile used to render an
image for that camera model is -0.7, then the actual default exposure value used during
rendering will be +0.3 - 0.7 = -0.4.
ProfileLookTableEncoding
Tag 51108 (C7A4.H)
Type LONG
Count 1
Value See below
Default 0
Usage IFD 0 or Camera Profile IFD
Description
Provides a way for color profiles to specify how indexing into a 3D LookTable is performed
during raw conversion. This tag is not applicable to a 2.5D LookTable (i.e., where the Value
dimension is 1).
The currently defined values are:
0 = Linear encoding
1 = sRGB encoding
If set to 0 (Linear encoding), then the method used to apply a 3D LookTable is as follows:
1. Convert linear ProPhoto RGB values to HSV.
2. Use the HSV coordinates to index into the color table.
3. Apply color table result to the original HSV values.
4. Convert modified HSV values back to linear ProPhoto RGB.
If set to 1 (sRGB encoding), then the method used to apply a 3D LookTable is as follows:
1. Convert linear ProPhoto RGB values to HSV.
2. Encode V coordinate using the sRGB encoding curve.
3. Use the encoded HSV coordinates to index into the color table.
Digital Negative Specification September 2012 73
DNG Tags
ProfileHueSatMapEncoding
4
4. Apply color table result to the encoded values from step b.
5. Decode V coordinate using the sRGB decoding curve (inverse of step b).
6. Convert HSV values back to linear ProPhoto RGB (inverse of step a).
The second method (sRGB encoding) may be desirable to provide additional table precision to
dark (shadow) image values.
ProfileHueSatMapEncoding
Tag 51107 (C7A3.H)
Type LONG
Count 1
Value See below
Default 0
Usage IFD 0 or Camera Profile IFD
Description
Provides a way for color profiles to specify how indexing into a 3D HueSatMap is performed
during raw conversion. This tag is not applicable to 2.5D HueSatMap tables (i.e., where the
Value dimension is 1).
The currently defined values are:
0 = Linear encoding
1 = sRGB encoding
If set to 0 (Linear encoding), then the method used to apply a 3D HueSatMap is as follows:
1. Convert linear ProPhoto RGB values to HSV.
2. Use the HSV coordinates to index into the color table.
3. Apply color table result to the original HSV values.
4. Convert modified HSV values back to linear ProPhoto RGB.
If set to 1 (sRGB encoding), then the method used to apply a 3D HueSatMap is as follows:
1. Convert linear ProPhoto RGB values to HSV.
2. Encode V coordinate using the sRGB encoding curve.
Digital Negative Specification September 2012 74
DNG Tags
OriginalDefaultFinalSize
4
3. Use the encoded HSV coordinates to index into the color table.
4. Apply color table result to the encoded values from step b.
5. Decode V coordinate using the sRGB decoding curve (inverse of step b).
6. Convert HSV values back to linear ProPhoto RGB (inverse of step a).
The second method (sRGB encoding) may be desirable to provide additional table precision to
dark (shadow) image values.
OriginalDefaultFinalSize
Tag 51089 (C791.H)
Type SHORT or LONG
Count 2
Value Value 0: width
Value 1: length
Default See below
Usage IFD
Description
If this file is a proxy for a larger original DNG file, this tag specifics the default final size of
the larger original file from which this proxy was generated.
The default value for this tag is default final size of the current DNG file, which is
DefaultCropSize * DefaultScale.
Digital Negative Specification September 2012 75
DNG Tags
OriginalBestQualityFinalSize
4
OriginalBestQualityFinalSize
Tag 51090 (C792.H)
Type SHORT or LONG
Count 2
Value Value 0: width
Value 1: length
Default See below
Usage IFD
Description
If this file is a proxy for a larger original DNG file, this tag specifics the best quality final size
of the larger original file from which this proxy was generated.
The default value for this tag is the OriginalDefaultFinalSize, if specified. Otherwise the
default value for this tag is the best quality size of the current DNG file, which is
DefaultCropSize * DefaultScale * BestQualityScale.
OriginalDefaultCropSize
Tag 51091 (C793.H)
Type SHORT or LONG or
RATIONAL
Count 2
Value Value 0: width
Value 1: length
Default See below
Usage IFD 0
Description
If this file is a proxy for a larger original DNG file, this tag specifics the DefaultCropSize of
the larger original file from which this proxy was generated.
The default value for this tag is the OriginalDefaultFinalSize, if specified. Otherwise, the
default value for this tag is the DefaultCropSize of the current DNG file.
Digital Negative Specification September 2012 76
DNG Tags
NewRawImageDigest
4
NewRawImageDigest
Tag 51111 (C7A7.H)
Type BYTE
Count 16
Value See below
Default Optional
Usage IFD 0
Description
This tag is a modified MD5 digest of the raw image data. It has been updated from the
algorithm used to compute the RawImageDigest tag be more multi-processor friendly, and to
support lossy compression algorithms. The details of the algorithm used to compute this tag
are documented in the Adobe DNG SDK source code.
RawToPreviewGain
Tag 51112 (C7A8.H)
Type DOUBLE
Count 1
Value See below
Default 1.0
Usage Preview IFD
Description
The gain (what number the sample values are multiplied by) between the main raw IFD and
the preview IFD containing this tag.
Digital Negative Specification September 2012 77
5
Mapping Raw Values to Linear
Reference Values
The section describes DNG's processing model for mapping stored raw sensor values into
linear reference values.
Linear reference values encode zero light as 0.0, and the maximum useful value (limited by
either sensor saturation or analog to digital converter clipping) as 1.0. If SamplesPerPixel is
greater than one, each sample plane should be processed independently.
The processing model follows these steps:
Linearization
Black Subtraction
Rescaling
Clipping
Linearization
The first step is to process the raw values through the look-up table specified by the
LinearizationTable tag, if any. If the raw value is greater than the size of the table, it is mapped
to the last entry of the table.
Black Subtraction
The black level for each pixel is then computed and subtracted. The black level for each pixel
is the sum of the black levels specified by the BlackLevel, BlackLevelDeltaH and
BlackLevelDeltaV tags.
Rescaling
The black subtracted values are then rescaled to map them to a logical 0.0 to 1.0 range. The
scale factor is the inverse of the difference between the value specified in the WhiteLevel tag
and the maximum computed black level for the sample plane.
Digital Negative Specification September 2012 78
Mapping Raw Values to Linear Reference Values
Clipping
5
Clipping
The rescaled values are then clipped to a 0.0 to 1.0 logical range.
Digital Negative Specification September 2012 79
6
Mapping Camera Color Space to
CIE XYZ Space
This section describes the DNG processing model for mapping between the camera color
space coordinates (linear reference values) and CIE XYZ (with a D50 white point).
Camera Calibration Matrices
DNG 1.2.0.0 and later supports different companies creating the camera calibration tags using
different reference cameras.
When rendering a DNG file using a camera profile, it is important to know if the selected
camera profile was designed using the same reference camera used to create the camera
calibration tags. If so, then the camera calibration tags should be used. If not, then it is
preferable to ignore the camera calibration tags and use identity matrices instead in order to
minimize the worse case calibration mismatch error.
This matching is done by comparing the CameraCalibrationSignature tag and the
ProfileCalibrationSignature tag for the selected camera profile. If they match, then use the
camera calibration tags. If not, then use identity matrices.
One or Two Color Calibrations
DNG provides for one or two sets of color calibration tags, each set optimized for a different
illuminant. If both sets of color calibration tags are included, then the raw converter should
interpolate between the calibrations based on the white balance selected by the user.
If two calibrations are included, then it is recommended that one of the calibrations be for a
low color temperature illuminant (e.g., Standard-A) and the second calibration illuminant be
for a higher color temperature illuminant (e.g., D55 or D65). This combination has been found
to work well for a wide range of real-world digital camera images.
DNG versions earlier than 1.2.0.0 allow the raw converter to choose the interpolation
algorithm. DNG 1.2.0.0 and later requires a specific interpolation algorithm: linear
interpolation using inverse correlated color temperature.
To find the interpolation weighting factor between the two tag sets, find the correlated color
temperature for the user-selected white balance and the two calibration illuminants. If the
white balance temperature is between two calibration illuminant temperatures, then invert all
the temperatures and use linear interpolation. Otherwise, use the closest calibration tag set.
Digital Negative Specification September 2012 80
Mapping Camera Color Space to CIE XYZ Space
Definitions used in the following sections
6
Definitions used in the following sections
Let n be the dimensionality of the camera color space (usually 3 or 4).
Let CM be the n-by-3 matrix interpolated from the ColorMatrix1 and ColorMatrix2 tags.
Let CC be the n-by-n matrix interpolated from the CameraCalibration1 and
CameraCalibration2 tags (or identity matrices, if the signatures dont match).
Let AB be the n-by-n matrix, which is zero except for the diagonal entries, which are defined
by the AnalogBalance tag.
Let RM be the 3-by-n matrix interpolated from the ReductionMatrix1 and ReductionMatrix2
tags.
Let FM be the 3-by-n matrix interpolated from the ForwardMatrix1 and ForwardMatrix2 tags.
Translating White Balance xy Coordinates to Camera Neutral
Coordinates
If the white balance is specified in terms of a CIE xy coordinate, then a camera neutral
coordinate can be derived by first finding the correlated color temperature for the xy value.
This value determines the interpolation weighting factor between the two sets of color
calibration tags.
The XYZ to camera space matrix is:
XYZtoCamera = AB * CC * CM
The camera neutral can be found by expanding the xy value to a 3-by-1 XYZ matrix (assuming
Y = 1.0) and multiplying it by the XYZtoCamera matrix:
CameraNeutral = XYZtoCamera * XYZ
Translating Camera Neutral Coordinates to White Balance xy
Coordinates
This process is slightly more complex than the transform in the other direction because it
requires an iterative solution.
1. Guess an xy value. Use that guess to find the interpolation weighting factor between the
color calibration tags. Find the XYZtoCamera matrix as above.
Digital Negative Specification September 2012 81
Mapping Camera Color Space to CIE XYZ Space
Camera to XYZ (D50) Transform
6
2. Find a new xy value by computing:
XYZ = Inverse (XYZtoCamera) * CameraNeutral
(If the XYZtoCamera matrix is not square, then use the pseudo inverse.)
3. Convert the resulting XYZ to a new xy value.
4. Iterate until the xy values converge to a solution.
Camera to XYZ (D50) Transform
DNG 1.2.0.0 and later support two methods of specifying the camera to XYZ (D50) transform,
depending on whether or not the forward matrix tags are included in the camera profile.
The use of the forward matrix tags is recommended for two reasons. First, it allows the camera
profile creator to control the chromatic adaptation algorithm used to convert between the
calibration illuminant and D50. Second, it causes the white balance adjustment (if the user
white balance does not match the calibration illuminant) to be done by scaling the camera
coordinates rather than by adapting the resulting XYZ values, which has been found to work
better in extreme cases.
If the ForwardMatrix tags are not included in the camera profile:
1. First, invert the XYZtoCamera matrix.
If n = 3, this is:
CameraToXYZ = Inverse (XYZtoCamera)
If n > 3, and the reduction matrix tags are included, then:
CameraToXYZ = Inverse (RM * XYZtoCamera) * RM
Otherwise:
CameraToXYZ = PseudoInverse (XYZtoCamera)
2. The white balanced transform is computed:
CameraToXYZ_D50 = CA * CameraToXYZ
CA, above, is a chromatic adaptation matrix that maps from the white balance xy value to
the D50 white point. The recommended method for computing this chromatic adaptation
matrix is to use the linear Bradford algorithm.
Digital Negative Specification September 2012 82
Mapping Camera Color Space to CIE XYZ Space
If the ForwardMatrix tags are included in the camera profile:
6
If the ForwardMatrix tags are included in the camera profile:
CameraToXYZ_D50 = FM * D * Inverse (AB * CC)
D, above, is a diagonal n-by-n matrix, computed so that the CameraToXYZ_D50 matrix maps
the selected camera neutral to XYZ D50. The forward matrix is required to map a unit vector
to XYZ D50 by definition, so D can be computed by finding the neutral for the reference
camera:
ReferenceNeutral = Inverse (AB * CC) * CameraNeutral
And then: D = Invert (AsDiagonalMatrix (ReferenceNeutral))
Applying the Hue/Saturation/Value Mapping Table
After the camera colors have been converted to XYZ (D50) values, the Hue/Saturation/Value
mapping table, if any, is applied. If there are two Hue/Saturation/Value mapping tables, then
they are interpolated in the same way that color calibration tags are interpolated. If only one
Hue/Saturation/Value table is included, then it is used regardless of the selected white balance.
1. First, the XYZ (D50) values are converted to linear RGB coordinates, using the ProPhoto
RGB primaries. (This is also known as RIMM space).
2. The linear RGB coordinates are converted to HSV coordinates (Hue-Saturation-Value).
3. The HSV coordinates are used to index the mapping table using tri-linear interpolation,
resulting in three values: hue shift (in degrees); saturation scale factor; value scale factor. If
the division count in a dimension is 1, then the table is constant for that dimension.
4. Hue is indexed using wrap-around math. For example, if HueDivisions is equal to 3, then
the table samples are at 0 degrees (red), 120 degrees (green), and 240 degrees (blue).
5. The hue coordinate is modified by adding the hue shift.
6. The saturation coordinate is modified by multiplying by the saturation scale factor, and
then clipping to no more than 1.0.
7. The value coordinate is modified by multiplying by the value scale factor, and then clipping
to no more than 1.0.
8. The HSV coordinates are converted to linear RGB coordinates, and then back to XYZ
(D50) values.
It is recommended that these tables be limited to use a ValueDivisions equal to 1, so the table
is only indexed by hue and saturation. In this way, all colors with the same hue and saturation,
but with different values, map to the same new hue and saturation while preserving their value
ratios.
Digital Negative Specification September 2012 83
7
Opcode List Processing
An opcode list is a list of opcodes, each of which does some specified image processing
operation. Each opcode is performed in sequence.
Opcode lists are always stored in big-endian byte order, no matter what the files main byte
order is. This allows DNG utility programs to copy opcode lists from file to file, without
needing to understand their detailed internal structure.
At the start of opcode list, there is a 32-bit unsigned integer count, which contains the number
of opcodes in the list. This is followed by the data for each opcode.
Each opcode starts with a 32-bit unsigned integer, which contains the opcode ID. The opcode
ID identifies the specific opcode. Documentation for each supported opcode ID in provided
later in this chapter.
Next is a 32-bit unsigned integer, which contains the DNG specification version in which the
opcode ID was defined. It is expected that new opcode IDs will be defined in future DNG
specification versions. A DNG reader should never attempt to process an opcode with a
version higher than DNG specification it was written to support.
Next is a 32-bit unsigned integer, which contains various flag bits. There are two defined flag
bits. If bit 0 (the least significant bit) is set to 1, the opcode is considered optional, and the
DNG reader may decide to not apply this opcode if it wishes, or it does not understand the
opcode ID. If bit 1 (the second to least significant bit) is set to 1, the opcode can be skipped
when doing preview quality processing, and only needs to be applied when doing full
quality processing.
Next is a 32-bit unsigned integer, containing the number of bytes in a variable size parameter
area for the opcode. The format of this variable size parameter area is dependent on the
specific opcode ID, and is documented later in this chapter, along with each supported opcode
ID.
When processing an opcode list, image values are clipped after the application of each opcode
to the logical range of the image being modified. For OpcodeList1, this range is 0 to 2
32
1 for
images with a bit depth greater than 16, otherwise 0 to 2
16
1 . For OpcodeList2 and
OpcodeList3, the logical range is 0.0 to 1.0.
The rest of this chapter contains information on each supported opcode ID.
Digital Negative Specification September 2012 84
Opcode List Processing
WarpRectilinear
7
WarpRectilinear
Opcode ID 1
DNG Version 1.3.0.0
Parameters
N LONG
k
r
0
i ,
DOUBLE
k
r
1
i ,
DOUBLE
k
r
2
i ,
DOUBLE
k
r
3
i ,
DOUBLE
k
t
0
i ,
DOUBLE
k
t
1
i ,
DOUBLE
For each coefficient set i 1 2 . N , , , e :
c

x
DOUBLE
c

y
DOUBLE
Description
This opcode applies a warp to an image and can be used to correct geometric distortion and
lateral (transverse) chromatic aberration for rectilinear lenses. The warp function supports
both radial and tangential distortion correction.
Let K
i
k
r
0
i ,
k
r
1
i ,
k
r
2
i ,
k
r
3
i ,
k
t
0
i ,
k
t
1
i ,
, , , , , { } = denote the i th coefficient set, where
i 1 2 . N

. , , , e

Parameter N is the number of coefficient sets. N must be 1 or the total number of image
planes. If N 1 = , then a single set of warp coefficients (i.e., K
1
) is applied to all image
planes, i.e., all planes undergo the same transformation. If N 1 > , coefficient set K
i
is used to
process the i th image plane, e.g., K
1
defines the warp function for the first image plane, K
2

defines the warp function for the second image plane, etc.
Parameters K
i
k
r
0
i ,
k
r
1
i ,
k
r
2
i ,
k
r
3
i ,
k
t
0
i ,
k
t
1
i ,
, , , , , { } = are the radial and tangential
coefficients that define the warp function for the i th image plane; see below for
Digital Negative Specification September 2012 85
Opcode List Processing
WarpRectilinear
7
implementation details and restrictions. Note that if k
r
0
i ,
1 = and the remaining terms are
zero, then the warp function is the identity (i.e., no warp will be applied).
Parameters c

x
c

y
, ( ) are the normalized x- and y-coordinates of the optical center, relative to
the top-left pixel of the image. Example 1: specifying (0.5, 0.5) means that the optical center
lies exactly at the image center. Example 2: specifying (1, 0) means that the optical center lies
at the top-right pixel of the image.
Processing of this opcode is performed as follows.
Let I'
i
x' y' , ( ) be the pixel value of the i th plane of the original unwarped image at pixel
position x' y' , ( ) i.e., before opcode processing.
Let I
i
x y , ( ) be the pixel value of the i th plane of the warped image at pixel position x y , ( )
i.e., after opcode processing.
For each pixel x y , ( ) of the i th plane of the warped image, compute:
i.e., the pixel at position x' y' , ( ) in the original unwarped image is effectively moved to
position x y , ( ) in the final warped image, where
x' c
x
m x
r
x
t
A + A ( ) + =
y' c
y
m y
r
y
t
A + A ( ) + =
c
x
x
0
c

x
x
1
x
0
( ) + =
c
y
y
0
c

y
y
1
y
0
( ) + =
m
x
max x
0
c
x
x
1
c
x
, ( ) =
m
y
max y
0
c
y
y
1
c
y
, ( ) =
m m
x
2
m
y
2
+ =
r x
2
y
2
A + A =
f k
r
0
i ,
k
r
1
i ,
r
2
k
r
2
i ,
r
4
k
r
3
i ,
r
6
+ + + =
x A x c
x
( ) m =
y A y c
y
( ) m =
Digital Negative Specification September 2012 86
Opcode List Processing
WarpRectilinear
7
[Continued from the previous page...]
x
r
A f x A =
y
r
A f y A =
x
t
A k
t
0
i ,
2 x y A A ( ) k
t
1
i ,
r
2
2 x
2
A + ( ) + =
y
t
A k
t
1
i ,
2 x y A A ( ) k
t
0
i ,
r
2
2 y
2
A + ( ) + =
x
0
y
0
, ( ) pixel coordinates of the top-left pixel of the warped image =
x
1
y
1
, ( ) pixel coord. of the bottom-right pixel of the warped image =
Notes and Restrictions
x
r
A y
r
A , ( ) and x
t
A y
t
A , ( ) are the radial and tangential warp components, respectively.
m is the Euclidean distance (in pixels) from the optical center to the farthest pixel in the
warped image.
r is the normalized Euclidean distance in 0 1 , | | ( ) from the optical center to a given pixel in
the warped image.
It is recommended that implementations use a suitable resampling kernel, such as a cubic
spline.
This opcode can be used to correct lateral (transverse) chromatic aberration by specifying the
appropriate coefficients for each image plane separately.
Each coefficient set K
i
must satisfy the following constraints. Let x' y' , ( ) F x y , ( ) = be the
2D warp function defined above. Let x' F
x
x y , ( ) = and y' F
y
x y , ( ) = be the x-component
and y-component of F x y , ( ) , respectively. Let w r ( ) k
r
0
i ,
r k
r
1
i ,
r
3
k
r
2
i ,
r
5
k
r
3
i ,
r
7
+ + + = .
The constraints are:
F x y , ( ) must be invertible.
Fx x y , ( ) must be an increasing function of x for all x x
0
x
1
, | | e , i.e., cF
x
x y , ( ) cx 0 > .
Fy x y , ( ) must be an increasing function of y for all y y
0
y
1
, | | e , i.e., cF
y
x y , ( ) cy 0 > .
w r ( ) must be an increasing function of r for all r 0 1 , | | e , i.e., w' r ( ) 0 > .
Digital Negative Specification September 2012 87
Opcode List Processing
WarpFisheye
7
WarpFisheye
Opcode ID 2
DNG Version 1.3.0.0
Parameters
N LONG
k
r
0
i ,
DOUBLE
k
r
1
i ,
DOUBLE
k
r
2
i ,
DOUBLE
k
r
3
i ,
DOUBLE
For each coefficient set i 1 2 . N , , , e :
c

x
DOUBLE
c

y
DOUBLE
Description
This opcode applies a warp to an image and can be used to unwrap an image captured with a
fisheye lens and map it instead to a perspective projection. It can also be used to correct
geometric distortion and lateral (transverse) chromatic aberration for both fisheye and
rectilinear lenses.
Let K
i
k
r
0
i ,
k
r
1
i ,
k
r
2
i ,
k
r
3
i ,
, , , { } = denote the i th coefficient set, where i 1 2 . N , , , e .
Parameter N is the number of coefficient sets. N must be 1 or the total number of image
planes. If N 1 = , then a single set of warp coefficients (i.e., K
1
) is applied to all image
planes, i.e., all planes undergo the same transformation. If N 1 > , coefficient set K
i
is used to
process the i th image plane, e.g., K
1
defines the warp function for the first image plane, K
2

defines the warp function for the second image plane, etc.
Parameters K
i
k
r
0
i ,
k
r
1
i ,
k
r
2
i ,
k
r
3
i ,
, , , { } = are the coefficients that define the warp function
for the i th image plane; see below for implementation details and restrictions.
Parameters c

x
c

y
, ( ) are the normalized x- and y-coordinates of the optical center, relative to
the top-left pixel of the image. Example 1: specifying (0.5, 0.5) means that the optical center
lies exactly at the image center. Example 2: specifying (1, 0) means that the optical center lies
at the top-right pixel of the image.
Digital Negative Specification September 2012 88
Opcode List Processing
WarpFisheye
7
Processing of this opcode is performed as follows.
Let I'
i
x' y' , ( ) be the pixel value of the i th plane of the original unwarped image at pixel
position x' y' , ( ) i.e., before opcode processing.
Let I
i
x y , ( ) be the pixel value of the i th plane of the warped image at pixel position x y , ( )
i.e., after opcode processing.
For each pixel x y , ( ) of the i th plane of the warped image, compute:
i.e., the pixel at position x' y' , ( ) in the original unwarped image is effectively moved to
position x y , ( ) in the final warped image, where
x' c
x
m f x A ( ) + =
y' c
y
m f y A ( ) + =
c
x
x
0
c

x
x
1
x
0
( ) + =
c
y
y
0
c

y
y
1
y
0
( ) + =
m
x
max x
0
c
x
x
1
c
x
, ( ) =
m
y
max y
0
c
y
y
1
c
y
, ( ) =
m m
x
2
m
y
2
+ =
r x
2
y
2
A + A =
u arctan r ( ) =
f
1
r
--- k
r
0
i ,
u k
r
1
i ,
u
3
k
r
2
i ,
u
5
k
r
3
i ,
u
7
+ + + ( ) =
x A x c
x
( ) m =
y A y c
y
( ) m =
x
0
y
0
, ( ) pixel coordinates of the top-left pixel of the warped image =
x
1
y
1
, ( ) pixel coord. of the bottom-right pixel of the warped image =
Notes and Restrictions
m is the Euclidean distance (in pixels) from the optical center to the farthest pixel in the
warped image.
r is the normalized Euclidean distance in 0 1 , | | ( ) from the optical center to a given pixel in
the warped image.
Digital Negative Specification September 2012 89
Opcode List Processing
FixVignetteRadial
7
It is recommended that implementations use a suitable resampling kernel, such as a cubic
spline.
This opcode can be used to correct lateral (transverse) chromatic aberration by specifying the
appropriate coefficients for each image plane separately.
Each coefficient set K
i
must satisfy the following constraints. Let
w r ( ) k
r
0
i ,
u k
r
1
i ,
u
3
k
r
2
i ,
u
5
k
r
3
i ,
u
7
+ + + = , where u arctan r ( ) = . w r ( ) must be an
increasing function of r for all r 0 1 , | | e , i.e., w' r ( ) 0 > .
FixVignetteRadial
Opcode ID 3
DNG Version 1.3.0.0
Parameters
k
0
DOUBLE
k
1
DOUBLE
k
2
DOUBLE
k
3
DOUBLE
k
4
DOUBLE
c

x
DOUBLE
c

y
DOUBLE
Description
This opcode applies a gain function to an image and can be used to correct vignetting.
Parameters (k
0
k
1
k
2
k
3
k
4
, , , , ) define a radially-symmetric gain function, explained below.
Note that if all k
i
terms are zero, then the gain function is the identity (i.e., no gain will be
applied).
Parameters (c

x
c

y
, ) are the normalized x- and y-coordinates of the optical center, relative to
the top-left pixel of the image. Example 1: specifying (0.5; 0.5) means that the optical center
lies exactly at the image center. Example 2: specifying (1, 0) means that the optical center lies
at the top-right pixel of the image.
Processing of this opcode is performed as follows.
Let I'
i
x y , ( ) be the uncorrected pixel value of the i th plane of the image at pixel position
(x y , ) (i.e., before opcode processing).
Digital Negative Specification September 2012 90
Opcode List Processing
FixBadPixelsConstant
7
Let I
i
x y , ( ) be the corrected pixel value of the i th plane of the image at pixel position (x y , )
(i.e., after opcode processing).
For each pixel (x y , ) in the i th plane of the image, compute:
I
i
x y , ( ) g I'
i
x y , ( ) =
where
g 1 k
0
r
2
k
1
r
4
k
2
r
6
k
3
r
8
k
4
r
10
+ + + + + =
c
x
x
0
c

x
x
1
x
0
( ) + =
c
y
y
0
c

y
y
1
y
0
( ) + =
m
x
max x
0
c
x
x
1
c
x
, ( ) =
m
y
max y
0
c
y
y
1
c
y
, ( ) =
m m
x
2
m
y
2
+ =
r
1
m
---- x c
x
( )
2
y c
y
( )
2
+ =
x
0
y
0
, ( ) pixel coordinates of the top-left pixel of the image =
x
1
y
1
, ( ) pixel coordinates of the bottom-right pixel of the image =
Note that m represents the Euclidean distance (in pixels) from the optical center to the farthest
pixel in the image, and r represents the normalized Euclidean distance in 0 1 , | | ( ) from the
optical center to a given pixel.
FixBadPixelsConstant
Opcode ID 4
DNG Version 1.3.0.0
Parameters
Constant - LONG
BayerPhase - LONG
Description
This opcode patches (interpolates over) bad pixels in a Bayer pattern CFA image. The bad
pixels are marked in the image by setting the bad pixels to a value of Constant.
Digital Negative Specification September 2012 91
Opcode List Processing
FixBadPixelsList
7
If BayerPhase is 0, then the top-left pixel of the image is red pixel.
If BayerPhase is 1, then the top-left pixel of the image is green pixel in a green/red row.
If BayerPhase is 2, then the top-left pixel of the image is green pixel in a green/blue row.
If BayerPhase is 3, then the top-left pixel of the image is blue pixel.
FixBadPixelsList
Opcode ID 5
DNG Version 1.3.0.0
Parameters
BayerPhase - LONG
BadPointCount - LONG
BadRectCount - LONG
For Each BadPointCount:
BadPointRow - LONG
BadPointColumn - LONG
For Each BadRectCount:
BadRectTop - LONG
BadRectLeft - LONG
BadRectBottom - LONG
BadRectRight - LONG
Description
This opcode patches (interpolates over) bad pixels and rectangles in a Bayer pattern CFA
image. The bad pixels and rectangles are specified as parameters for this opcode.
If BayerPhase is 0, then the top-left pixel of the image is red pixel.
If BayerPhase is 1, then the top-left pixel of the image is green pixel in a green/red row.
If BayerPhase is 2, then the top-left pixel of the image is green pixel in a green/blue row.
If BayerPhase is 3, then the top-left pixel of the image is blue pixel.
BadPointCount is the number of bad pixels to patch. BadPointRow and BadPointColumn
specify the coordinates of the bad pixels to patch.
Digital Negative Specification September 2012 92
Opcode List Processing
TrimBounds
7
BadRectCount is the number of bad rectangles to patch. BadRectTop, BadRectLeft,
BadRectBottom, and BadRectRight specify the coordinates of the bad rectangles to patch.
TrimBounds
Opcode ID 6
DNG Version 1.3.0.0
Parameters
Top - LONG
Left - LONG
Bottom - LONG
Right - LONG
Description
This opcode trims the image to the rectangle specified by Top, Left, Bottom, and Right.
MapTable
Opcode ID 7
DNG Version 1.3.0.0
Parameters
Top - LONG
Left - LONG
Bottom - LONG
Right - LONG
Plane - LONG
Planes - LONG
RowPitch - LONG
ColPitch - LONG
TableSize - LONG
For Each TableSize:
Digital Negative Specification September 2012 93
Opcode List Processing
MapPolynomial
7
TableEntry - SHORT
Description
This opcode maps a specified area and plane range of an image through a 16-bit LUT.
The bounds of the area of the image to be affected is specified by the Top, Left, Bottom, Right
parameters. The first plane, and the number of planes, to be modified are specified by the
Plane and Planes parameters. If RowPitch not equal to one, then only every RowPitch rows
starting at the Top are affected. If ColPitch is not equal to one, then only every ColPitch
columns starting at Left are affected.
TableSize specifies the number of entries in the 16-bit LUT. TableEntry specifies each table
entry.
MapPolynomial
Opcode ID 8
DNG Version 1.3.0.0
Parameters
Top - LONG
Left - LONG
Bottom - LONG
Right - LONG
Plane - LONG
Planes - LONG
RowPitch - LONG
ColPitch - LONG
Degree - LONG
For Each Value 0...Degree:
Coefficient - DOUBLE
Description
This opcode maps a specified area and plane range of an image through a polynomial function.
The bounds of the area of the image to be affected is specified by the Top, Left, Bottom, Right
parameters. The first plane, and the number of planes, to be modified are specified by the
Plane and Planes parameters. If RowPitch not equal to one, then only every RowPitch rows
Digital Negative Specification September 2012 94
Opcode List Processing
GainMap
7
starting at the Top are affected. If ColPitch is not equal to one, then only every ColPitch
columns starting at Left are affected.
The mapping function is a polynomial of degree Degree. The maximum allowed value for
Degree is 8. The coefficients are stored in increasing order, starting with the zero degree
coefficient (constant term).
GainMap
Opcode ID 9
DNG Version 1.3.0.0
Parameters
Top - LONG
Left - LONG
Bottom - LONG
Right - LONG
Plane - LONG
Planes - LONG
RowPitch - LONG
ColPitch - LONG
MapPointsV - LONG
MapPointsH - Long
MapSpacingV - DOUBLE
MapSpacingH - DOUBLE
MapOriginV - DOUBLE
MapOriginH - DOUBLE
MapPlanes - LONG
For Each MapPointsV
For Each MapPointsH
For Each MapPlanes
MapGain - FLOAT
Digital Negative Specification September 2012 95
Opcode List Processing
DeltaPerRow
7
Description
This opcode multiplies a specified area and plane range of an image by a gain map.
The bounds of the area of the image to be affected is specified by the Top, Left, Bottom, Right
parameters. The first plane, and the number of planes, to be modified are specified by the
Plane and Planes parameters. If RowPitch not equal to one, then only every RowPitch rows
starting at the Top are affected. If ColPitch is not equal to one, then only every ColPitch
columns starting at Left are affected.
The gain map is a sub-sampled 2-D 32-bit floating-point image. MapPointsV is the number of
samples in vertical direction. MapPointsH is the number of samples in the horizontal direction.
The gain map is not required to cover the entire image being modified. Inside the gain map
bounds, values are interpolated using bi-linear interpolation. Outside the gain map bounds,
values are replicated from the edges of the gain map.
The origin of the gain map relative to image being modified is specified by MapOriginV
(vertical direction) and MapOriginH (horizontal direction), which are in relative coordinates,
where 1.0 is equal to the height or width of the image being modified, respectively.
The spacing between gain map points is specified by MapSpacingV (vertical direction) and
MapSpacingH (horizontal direction). Again these are in relative coordinates, where 1.0 is
equal to the height or width of the image being modified, respectively.
MapPlanes specifies the number of planes in the gain map. If Planes > MapPlanes, the last
gain map plane is used for any remaining planes being modified.
DeltaPerRow
Opcode ID 10
DNG Version 1.3.0.0
Parameters
Top - LONG
Left - LONG
Bottom - LONG
Right - LONG
Plane - LONG
Planes - LONG
RowPitch - LONG
ColPitch - LONG
Digital Negative Specification September 2012 96
Opcode List Processing
DeltaPerColumn
7
Count - LONG
For Each Count:
Delta - FLOAT
Description
This opcode applies a per-row delta (constant offset) to a specified area and plane range of an
image.
The bounds of the area of the image to be affected is specified by the Top, Left, Bottom, Right
parameters. The first plane, and the number of planes, to be modified are specified by the
Plane and Planes parameters. If RowPitch not equal to one, then only every RowPitch rows
starting at the Top are affected. If ColPitch is not equal to one, then only every ColPitch
columns starting at Left are affected.
Count is the number of deltas, and is required to match the number of affected rows in the
specified area.
The delta to add to each affected row is specified by Delta.
DeltaPerColumn
Opcode ID 11
DNG Version 1.3.0.0
Parameters
Top - LONG
Left - LONG
Bottom - LONG
Right - LONG
Plane - LONG
Planes - LONG
RowPitch - LONG
ColPitch - LONG
Count - LONG
For Each Count:
Delta - FLOAT
Digital Negative Specification September 2012 97
Opcode List Processing
ScalePerRow
7
Description
This opcode applies a per-column delta (constant offset) to a specified area and plane range of
an image.
The bounds of the area of the image to be affected is specified by the Top, Left, Bottom, Right
parameters. The first plane, and the number of planes, to be modified are specified by the
Plane and Planes parameters. If RowPitch not equal to one, then only every RowPitch rows
starting at the Top are affected. If ColPitch is not equal to one, then only every ColPitch
columns starting at Left are affected.
Count is the number of deltas, and is required to match the number of affected columns in the
specified area.
The delta to add to each affected column is specified by Delta.
ScalePerRow
Opcode ID 12
DNG Version 1.3.0.0
Parameters
Top - LONG
Left - LONG
Bottom - LONG
Right - LONG
Plane - LONG
Planes - LONG
RowPitch - LONG
ColPitch - LONG
Count - LONG
For Each Count:
Scale - FLOAT
Description
This opcode applies a per-row scale to a specified area and plane range of an image.
The bounds of the area of the image to be affected is specified by the Top, Left, Bottom, Right
parameters. The first plane, and the number of planes, to be modified are specified by the
Plane and Planes parameters. If RowPitch not equal to one, then only every RowPitch rows
Digital Negative Specification September 2012 98
Opcode List Processing
ScalePerColumn
7
starting at the Top are affected. If ColPitch is not equal to one, then only every ColPitch
columns starting at Left are affected.
Count is the number of scale values, and is required to match the number of affected rows in
the specified area.
The scale to multiply each affected row by is specified by Scale.
ScalePerColumn
Opcode ID 13
DNG Version 1.3.0.0
Parameters
Top - LONG
Left - LONG
Bottom - LONG
Right - LONG
Plane - LONG
Planes - LONG
RowPitch - LONG
ColPitch - LONG
Count - LONG
For Each Count:
Scale - FLOAT
Description
This opcode applies a per-column scale to a specified area and plane range of an image.
The bounds of the area of the image to be affected is specified by the Top, Left, Bottom, Right
parameters. The first plane, and the number of planes, to be modified are specified by the
Plane and Planes parameters. If RowPitch not equal to one, then only every RowPitch rows
starting at the Top are affected. If ColPitch is not equal to one, then only every ColPitch
columns starting at Left are affected.
Count is the number of scale values, and is required to match the number of affected columns
in the specified area.
The scale to multiply each affected column by is specified by Scale.
Digital Negative Specification September 2012 99
Appendix A: Compatibility with
Previous Versions
This appendix documents only the differences between this and previous versions of the DNG
specification that are relevant to compatibility. Differences that are not relevant to
compatibility (e.g., new optional tags that DNG readers are not required to support) are not
documented in this appendix.
This information is useful in enabling DNG readers to correctly read DNG files with older
version numbers. It also helps determine what version DNG writers can include in the
DNGBackwardVersion tag.
The first version of the DNG speciation that was published was version 1.0.0.0.
Compatibility Issue 1: ActiveArea Tag
The ActiveArea tag was added to the DNG specification in version 1.1.0.0. Previous versions
of the DNG specification do not support storing masked pixels.
DNG writers should set the DNGBackwardVersion to a minimum of 1.1.0.0 if the masked
pixels are stored in the DNG file.
Compatibility Issue 2: 16-bit Lossless JPEG Encoding
The Lossless JPEG encoder/decoder used by Adobe applications to read and write DNG files
before version 1.1.0.0 incorrectly deviated from the JPEG specification when dealing with 16-
bit data. Since both the encoder and decoder deviated in the same way, no data was lost;
however the data stream did not exactly match the data stream specified in the Lossless JPEG
specification.
Because the vast majority of DNG 1.0.0.0 files using 16-bit Lossless JPEG encoding were
created by Adobe applications, it is strongly recommended that software that reads or writes
DNG files with version numbers less than 1.1.0.0 incorporate this deviation. Software that
reads or writes DNG files with version 1.1.0.0 or later can safely assume that the Lossless
JPEG stream is fully compliant with the Lossless JPEG specification.
Description of deviation
Lossless JPEG encodes the difference between a predicted value and the actual value for each
pixel. With 16-bit data, these differences are computed modulo 16-bits, so the range of
possible differences is -32768 to + 32767. Two values are stored for the difference. First the
number of bits required to store the difference (encoded via a Huffman code), and then the
actual difference.
Digital Negative Specification September 2012 100
Appendix A: Compatibility with Previous Versions
Compatibility Issue 3: SubTileBlockSize
Using the difference encoding scheme in the Lossless JPEG specification, only one difference
value would take 16-bits to store: -32768. The Lossless JPEG specification special cases this
difference bit length, and since there is only one possible difference value it does not bother to
use any bits to store the actual difference.
In earlier versions of DNG the special case logic is not present, and the difference value for -
32768 is stored in the compressed data stream as with all other difference bit lengths.
Compatibility Issue 3: SubTileBlockSize
The SubTileBlockSize tag was added to the DNG specification in version 1.2.0.0. Earlier
versions of the DNG specification did not support this feature.
DNG writers should set the DNGBackwardVersion tag to a minimum of 1.2.0.0 if non-default
values are used for this tag.
Compatibility Issue 4: RowInterleaveFactor
The RowInterleaveFactor tag was added to the DNG specification in version 1.2.0.0. Earlier
versions of the DNG specification did not support this feature.
DNG writers should set the DNGBackwardVersion tag to a minimum of 1.2.0.0 if non-default
values are used for this tag.
Compatibility Issue 5: PreviewColorSpace
The PreviewColorSpace tag was added to the DNG specification in version 1.2.0.0. Earlier
versions of the DNG specification did not support this feature.
DNG writers should set the DNGBackwardVersion tag to a minimum of 1.2.0.0 if non-default
values are used for this tag.
Compatibility Issue 6: CFALayout
CFALayout values 6 through 9 were added to the DNG specification in version 1.3.0.0. Earlier
versions of the DNG specification did not support these CFALayout values.
DNG writers should set the DNGBackwardVersion tag to a minimum of 1.3.0.0 if CFALayout
values 6 through 9 are used.
Digital Negative Specification September 2012 101
Appendix A: Compatibility with Previous Versions
Compatibility Issue 7: Opcode Lists
Compatibility Issue 7: Opcode Lists
The OpcodeList1, OpcodeList2, and OpcodeList3 tags, plus the initial set of supported
Opcode IDs, were added to the DNG specification in version 1.3.0.0.
DNG writers should not set the DNGBackwardVersion to a version less than the DNG version
for any opcode that is not marked optional.
Compatibility Issue 8: Floating Point Image Data
Support for floating point image data was added in DNG Version 1.4.0.0.
DNG writers should set the DNGBackwardVersion tag to a minimum of 1.4.0.0 if the main
image IFD uses floating point data.
Compatibility Issue 9: Transparent Pixels
Support for transparent pixels was added in DNG Version 1.4.0.0.
DNG writers should set the DNGBackwardVersion tag to a minimum of 1.4.0.0 if the DNG
file contains an IFD with NewSubFileType equal to 4.
Compatibility Issue 10: Deflate Compression
Support for compression code 8 (Deflate) was added in DNG Version 1.4.0.0.
DNG writers should set the DNGBackwardVersion tag to a minimum of 1.4.0.0 if the main
image IFD uses compression code 8.
Compatibility Issue 11: Lossy JPEG Compression
Support for compression code 34892 (Lossy JPEG) was added in DNG Version 1.4.0.0.
DNG writers should set the DNGBackwardVersion tag to a minimum of 1.4.0.0 if the main
image IFD uses compression code 34892.