Help index - Stereo Tool 7.

83
Introduction
Introduction: What is Stereo Tool? (outdated)
Stereo Tool configuration
Configuration Language and startup settings
License
CPU & Latency
Parameter scheduling to use different presets at different moments
Web interface for remote control via the built-in web server
Sound cards settings
Stereo Tool audio restoration
PNR Noise & Hum removes disturbing sounds.
Declipper repairs clipping: removes digital clipping distortion and restores dynamics.
Dequantizer increases the bit depth of audio.
Declipper repairs MP3 files.
Noise removal
Stereo Tool processing
General processing settings
Natural Dynamics restores dynamics
Phase rotation and Phase Delay
Automatic Gain Control (AGC)
Power Bass / Power Highs adds deep bass and highs to tracks that have little bass or highs in them
Stereo increases or decreases stereo separation
Equalizer
True Bass
Multiband Compressor
Multiband Compressor 2
Bandpass removes very high and very low frequencies
Bass Boost
Singleband Compressor
Limiting & Clipping
Advanced Clipper clips without distortion
Hard Limit, final limiter
Settings that improve the sound of streams
FM and AM transmitter settings
Using FM stereo and RDS encoding (outdated)
FM transmitter requirements
Sound card requirements (includes a list of supported sound cards)
Windows versions
Configuring the FM transmitter settings in Stereo Tool (outdated)
Overview of FM transmitter settings
Pre-emphasis, stereo & RDS coding, dynamic RDS texts, Stokkemask & BS412 compliance, composite clipping and
synchronization between FM transmitters
Overview of AM transmitter settings
Asymmetrical limiting and clipping
Using Stereo Tool with other software
Using multiple DSP plugins (for example Stereo Tool and SHOUTcast)
Stereo Tool and SAM Broadcaster (SAM3 or SAM4)
Stereo Tool and RadioBOSS
Help: What is Stereo Tool? - Stereo Tool 3.0
Stereo Tool is a plugin that can be used to influence the sound of audio recordings. It can make recordings sound equal and consistent in
volume and sound color, bring out the details, and increase existing stereo effects. Because the processed audio makes better use of the
available equipment, it often sounds better and richer and richer than the original - more details can be heard. Stereo widening is provided
to further improve the listening experience.

Special support is provided for (FM) radio stations: Extra loudness (+5.5 dB), FM pre-emphasis, and software stereo and RDS encoding.
Using Stereo Tool, even a very cheap (15 / $20) mono transmitter can sound like the big commercial stations, broadcasting in full stereo
with RDS.
Stereo Tool contains a dualband pre-limiter, 10-band multiband compressor/limiter with clipping, 10-band equalizer, 3-level overshoot
protection, extra loudness filter, AZIMUTH corrector, stereo image manipulator, lowpass filter, FM pre-emphasis, stereo and RDS encoder.
The dualband pre-limiter and multiband compressor can be used to get equal sound levels in different files, or to limit the maximum
sound level. The overshoot protection filters make sure that peaks in the sound stay below a configurable level.
The stereo image manipulator can be used to convert stereo to mono without getting any of the artifacts that are normally caused by this, to
repair recordings with phasing problems, to increase the stereo image, and even to create a stereo sound such that both speakers still
play all the instruments.
The lowpass filter can be used to filter out certain high frequencies, such as the 19 kHz pilot tone if you want to send the Winamp output
into an FM transmitter.
History

Back in 2001, I started my own internet radio station (Weird Titan Radio). At the time the highest stream quality that my provider offered was
56 kbit/s MP3. After some tests I quickly decided that the only way to get a decent quality was in mono.

Unfortunately, converting music to mono often causes the sound to get distorted, and even if that doesn't happen, the end result often
sounds very "thin". To solve this, I created some software to convert stereo to mono, without this quality effect. I also wrote some separate
tools for volume compression etc.
In 2004, I decided to convert all my processing software to C++ and create a WinAmp plugin: Stereo Tool.

In 2006, I created a new version with more features, a better interface and much lower CPU usage: Stereo Tool 2.0
In 2008, I added a lot of extra features that can be found on professional (FM) radio stations, which often cost thousands of Euros (or
dollars). Because of that, some of those features are no longer free in this new Stereo Tool 3.0 version - but that's only the features that are
interesting to (some) webcasters and FM radio stations, normal users should be able to use Stereo Tool 3.0 without ever noticing that
some parts are not free.
Configuration section
Non-audio settings.
Configuration and its sub-screens control the non-audio parts of Stereo Tool, such as performance, foreign language support, scheduling,
interfacing, sound cards and FM/AM transmitter settings.

Bypass panel
Turn all processing off.

Bypass processing
Turn all processing off.

If Bypass is enabled, Stereo Tool does nothing - the audio goes through unmodified. Bypass is useful to compare the original input with
the output, or to turn Stereo Tool off without having to unload it.

Auto start panel

Automatically start Stereo Tool when Windows starts.

Auto-start with Windows

Automatically start Stereo Tool when Windows starts.

Note that Stereo Tool is only started when the user under who's account this setting was enabled logs in. So, to make this work properly,
make sure that this happens automatically when Windows starts.

Watchdog panel
Automatic restart behavior when problems occur.
Stereo Tool is made to run unattended for months or even years without human interference. And - as many people have reported over the
years - it usually runs for years without a single glitch.
Audio processing is very different from other types of programs: Once the processing runs, as long as the program is untouched it does not
allocate or deallocate memory, it does not swap to or from the disk, it does not attempt to access servers, load or save files. All it does is
continuously getting a same-sized chunk of audio from a sound card, run the same processing over it and send it to another sound card.
This means that the vast majority of programming errors will cause issues immediately in the first second - if it runs for one second it will
probably run for years.
But if something goes wrong anyway, these settings determine what action is taken.
Note that if something bad happens, you will always get a popup window with a notification, and if the directory C:\temp exists and is writable
it is also written to a file, C:\temp\StereoTool_Exceptions_Log.txt.

Detect and attempt to fix sound card/VLC timeouts

Automatically calls Restart sound cards if an input or output does not work.

If no audio passes through a sound card input or output for over 10 seconds, if this setting is enabled Restart sound cards is triggered
automatically.

If after 4 restarts no audio has passed through it at all, Stereo Tool is restarted completely if Restart on crashes or unsolvable sound
card timeouts is also enabled.

Note that VLC inputs or outputs are not included in this (they might disappear for longer periods due to network issues etc).

Restart on crashes or unsolvable sound card timeouts

Restarts Stereo Tool after an unrecoverable issue.

If an exception occurs in Stereo Tool, it will first attempt to recover from it without restarting. If this keeps failing (more than 10 exceptions
in a 10 minute period, for some parts of code immediately), Stereo Tool will close itself and start a new instance.

The automatic recovery will be attempted regardless of whether this setting is enabled or disabled.

See also Detect and attempt to fix sound card/VLC timeouts; if an audio input or output does not function after restarting it several
times, Stereo Tool will also restart.

TEST ERROR (CRASH!)

CRASHES STEREO TOOL. Used to test watchdog and logging behavior.

Pressing this button will cause an exception in Stereo Tool that would normally cause it to crash. This can be used to test if the logging
works (see description above) and if the Restart on crashes or unsolvable sound card timeouts behavior functions properly.
 When this button is pressed, an attempt is made to write data to memory position 0. You will see an exception with code 0xc0000005,
read/write=1, r/w location=0x0.

User interface panel

Switch between simple and more extensive settings.

Startup minimized in system tray

Hides the Stereo Tool window upon startup.

Normally, the Stereo Tool window is shown when you start Stereo Tool. Once you have configured it to your satisfaction, if you only want to
hear Stereo Tool without seeing it, enable this option to auto-hide the window every time Stereo Tool is started.

Operating mode
[Select how extensive the user interface is.]

With this pull-down menu you can choose how extensive (and hence complicated to use) the user interface is.

if you want to load a preset and only change some simple things, use Simple. For normal users, Basic should be fine. Advanced and
Expert show a lot more settings that are difficult to use and can easily ruin the sound. Extreme Tweaker contains lots of obscure settings
that you should probably never touch.

Skin
Selects a built-in skin for the Stereo Tool GUI.

Full screen mode

Switches to Full Screen mode.

In Full Screen mode, the window bars are removed and the Windows start menu (if available) is hidden. Mainly useful if you have a small
touch screen, or if you use nothing else on your pc except Stereo Tool.

Full height mode

Switches to Full Height mode.

Full Height mode is similar to Full screen mode, but only in vertical direction. Useful for example when you want to show a window
beside the Stereo Tool window, like a music player.

Draw scopes at half height

Draws the scopes with more headroom.
If you have overshoots, this allows you to better see them.

Title bar info

Extra text shown in the title bar.
If you are running multiple Stereo Tool instances on a pc, for example if you have a streaming server, this allows you to change the title of
each Stereo Tool window to indicate which signal it is processing. The text is shown in the title bar and when you hover over the tray icon
in Windows.

Language panel
Settings for multi-language support.

Language
Language selection.
.stl (Stereo Tool Language) files must be placed in the same directory where Stereo Tool is installed. You can find language files in the
Language/Translation files section of the forum. In the future, they will be included in the Stereo Tool installation.

Export language file Export.stl

Used to translate Stereo Tool to another language.

This button exports the current texts to a file named Export.stl. This file is generated in the same directory where Stereo Tool is
installed. If that's not possible due to access rights, it is placed in the Users directory (normally C:\Users\yourusername\).

This file consists of lines with the original text in English, an = sign, and the translated text. Example:

Configuration=Configuratie
Language=Taal

If you want to edit this file, rename it to the language that you want to support (for example Nederlands.stl) and if needed copy it to the
directory where Stereo Tool is installed. After restarting Stereo Tool, you will see Nederlands as a selection under Link error ''. Now you
can start editing the file.
Some texts are too long to fit in on the screen if you have a small display. For such texts, multiple versions are available:

Startup minimized in system tray.0=Start geminimaliseerd in system tray

Startup minimized in system tray.1=Start geminimaliseerd
Startup minimized in system tray.2=Start min.

Upto 3 texts can be provided. In case your translation is longer than the original text and you need more alternative texts than are provided
in the generated file, you can just add them. If there's only one line, you can add the .0 yourself and create a .1 (and .2) version.

If a new version of Stereo Tool comes out, new texts may have been added or texts may have been changed. If you load your old
 language file and then press the Export button, you get a file where everything that you have translated previously is available in the file,
but things that you have not yet translated are there too - in English.

Password protection panel

Password protection settings.
Stereo Tool can be password protected. This makes it impossible to change or export any settings. Note: The protection is not very strong, if
someone has access to the .ini file they can still make changes or disable the password protection.

Password
The password to access Stereo Tool settings.

Warning: The password is stored without encoding, so don't reuse passwords that you use elsewhere.

Turn password protection on

Enables password protection.

Log in
Password entry field to log in.

Log out
Logs out.
License section
Activation and overview of licensed features.

Register panel
Registering Stereo Tool

Reg. key including < and >

Field to past a bought Stereo Tool license key in.
Use Control+V to paste a code in this field.

Confirm
Register using the key entered in Reg. key including < and >.

Paste
Paste clipboard to Reg. key including < and >.

Buy or upgrade license

Visit the Stereo Tool website to buy or upgrade a license.

Active license info panel

Overview of licensed features.
CPU & Latency section
Settings that affect global CPU usage and latency.

CPU and latency panel

Settings that affect global CPU load and latency.

Multicore processing
Enables multicore processing.

This should always be enabled, unless you really need to let Stereo Tool use only one core. By using multiple cores, about 70% more
audio can be processed in the same amount of time.
This option has no effect on a PC with only one CPU core (and no hyperthreading), or (on Windows) if the environment flag
NUMBER_OF_PROCESSORS is set to 1.

Process priority
Give Stereo Tool's processing priority over other running programs.

Normally, all programs get a share of the available CPU resources. Which means that if you run a lot of heavy programs, Stereo Tool
might not get enough CPU time to process everything in time and it might start to hiccup. Setting a higher priority here helps against that,
but it may also cause other programs to be slowed down more.

If you use Stereo Tool for background listening, you should probably leave this turned off. If the audio is 'mission critical', this should be
set to a high level.

Latency
Controls audio output delay.

This setting controls how much audio Stereo Tool gets to work with. If it can work on bigger blocks of audio, the resulting output quality is
usually better. But if you are talking through a microphone and listening to yourself on a headphone, too much delay can be annoying.
For the best quality, and definitely in all cases where the delay does not matter, use the maximum setting (4096 samples, around 93 ms,
depending on the sample rate). A good compromise between latency and quality is setting 1024 (23 ms), which gives a total latency of
about 28 ms (this includes sound card delays and the time needed to process the audio).
Each reduction of the latency by half makes the artifacts caused at lower latencies 4 times as loud. Which means that the step from 4096
to 1024 is smaller than that from 1024 to 512! At latency 512 (12 ms, 17 ms total latency) the audio quality really suffers.

The stand alone version of Stereo Tool has an extra LQ Low Latency monitoring output which offers a total latency of 11 ms.

Note: The quality is also very strongly affected by the Dynamically reduce deep bass to setting, especially at lower latencies.

Quality (CPU load)

Global CPU load control

Allows lowering the CPU load at the expense of a degradation in audio quality. For every 10% the Quality slider is below 100%, the CPU
load is reduced by 5% - so at the lowest setting of 20% the CPU load is 60% of that at Quality 100%.

At the maximum Latency setting, Quality has very little effect and a value of around 50% is acceptable for most people. At lower Latency
settings, the effect on the audio (especially bass) is much bigger.

Display refresh speed (CPU load)

Controls how often the display is refreshed.

Lowering the refresh rate slightly lowers the CPU load (more so if the window is very big), and it also helps a lot when using a remote
connection to a pc running Stereo Tool.

Auto
Automatically lowers Display refresh speed (CPU load) when the CPU load is high.

Ignore high frequencies

Throws high frequencies away to reduce CPU load.

Doing this for FM or for streaming has no effect on the audio quality if the Frequency is set higher than the Lowpass frequency. One
exception is that for Stokkemask, a minimum frequency of 19200 must be set. The CPU load reduction can be upto 30%, depending on
the settings. Using this setting does increase the latency a bit (also upto 30%).

Frequency
The frequency above which tones are ignored.
CPU core affinities panel
Controls the CPU core affinities.
Setting the CPU core affinities is useful for low latency processing, because it makes the behavior predictable. Beside that, it is also usable
when you run a lot of different Stereo Tool instances or other software on one pc, in that case you can give each instance its own core or
cores.
By default, if these settings are left to Any, the OS (Windows, Linux, Mac OS X) will control what runs on which core.
For optimal performance and low latencies, use these guidelines:
Try to avoid Core 0. Core 0 is used by drivers, which means that when using core 0 at very low latencies, you run a risk of getting hiccups.
When using a system that has Hyperthreading, assign Processing main and Processing 2nd to virtual cores that don't share the same
physical cores. On most systems, that means that you should avoid combining core 0 with 1, core 2 with 3, core 4 with 5 etc. Also note
that the fact that drivers use core 0 might also makes the core 1 performance less constant.
So, if you have a system with 4 cores with Hyperthreading, then Windows will see 8 cores, and you should normally select for example cores
2 and 4 for Processing main and Processing 2nd, and core 0 for Link error '3007' and Link error '3008' things that are non-critical.

Processing main
Main processing core affinity.

This core will perform about 60-70% of all the processing tasks.
Windows only: Set to Mask if you want to use an affinity Mask that leaves part of the control to Windows.

Mask
Main processing affinity mask.

Only used when Processing main is set to Mask.

See Windows MSDN documentation. You should normally not need this.

Processing 2nd
2nd processing core affinity.

This core will perform about 30-40% of all the processing tasks.

GUI
GUI core affinity.

This core handles the Stereo Tool GUI display, which is non-critical. Make sure that it doesn't interfere with processing.

Server
HTTP server core affinity.

This core handles the Stereo Tool Web, at least the internal part. Non-critical, you should normally just set this to the same core as the
GUI.

Percentage of time in processing (top is incl. overhead) panel

How much of the available time is spent for processing.
This is not a CPU load display. Here are the differences between an actual CPU display and what this graph shows:
Shows time between start and end of processing.If this bar is nearly full, there's a performance problem; if it's not too close to 100% the
processing will run fine.Not that easy to translate, time could also be spent waiting (even on other programs that occupy the same CPU
core).
CPU load Percentage of time
Shows total load or load per core
Since load is spread over cores, can show load below 100% and still
have performance problems.
Higher load from Stereo Tool means that other programs will run
slower.

Basically, if this bar is nearly full (around 90% or higher) you can expect performance problems.
Please note that overhead such as resampling (used when enabling Synchronize with different output sound card (not ASIO) or
Synchronize FM transmitters) is not displayed in this graph, but does take extra time.

Percentage of time in GUI display panel

Shows time spent to display the GUI.
See Percentage of time in processing (top is incl. overhead). This graph is less useful because it doesn't show actual CPU usage - in
fact, since the GUI thread runs at a lower priority a lot of the time that's displayed might be spent waiting for other programs.
If only Stereo Tool is running on a system which has enough CPU cores available it will show something that's close to the actual CPU
usage of the GUI.
Scheduler section
Schedule loading of different .sts settings files

Schedule panel
Scheduling loading of different .sts settings file
This feature continuously looks for a Poll STS File name on disk. It is loaded upon startup, and every time its date/time changes. This can be
used for example to have different processing during the day and night. The date/time is checked 4 times per second.
It is a good idea to first create a file with the necessary changes, and then rename it to the file that Stereo Tool checks. This avoids problems
when the file is being loaded while you are still writing to it. The modification date is only accurate op to seconds. So changes less than a
second after the previous changes may be ignored.
The new file only needs to contain the changes - it is read "on top of" the already existing data. For example, to change the Pre Amp value,
use:
[Common]
Pre amplifier=10
Warning: Certain abrupt parameter changes can give audible effects.

Poll STS Enabled

Enable file polling.
See the overall description of Schedule.

Poll STS File name

File name of the file that's used to update the settings.
See Schedule for details.
Web interface section
Web interface to configure Stereo Tool remotely.
Stereo Tool contains a web server that shows all the Stereo Tool settings in a browser, enabling configuring it from a remote pc, phone or
tablet.
Important: At this moment the web interface is NOT protected against unauthorized use. Anyone can get in and change settings. So DO NOT
USE THIS on a system with an open internet connection.
Metering is not visible in this interface, so if you can, it's often easier to use a program like TeamViewer to login to the pc remotely.

Web panel
Setting for the Stereo Tool web interface.

Enable web interface

Enables the web interface.

Important: At this moment the web interface is NOT protected against unauthorized use. Anyone can get in and change settings. So DO
NOT USE THIS on a system with an open internet connection.

Both enabling and disabling the web interface requires a restart!

Port
Web interface port number.
The port number at which the web interface can be reached. To access it, go to http://ip_address_of_pc:port/

So for example, on a local pc with the web host configured at port 8080 you can use http://127.0.0.1:8080/ .

Small buttons
Makes buttons in the web interface smaller.
Useful for very small screens such as phones.
Sound cards section
Configuration of input and output sound cards, and synchronization.
Besides choosing sound cards, this section also lets you configure FM transmitter synchronization, which can be used to synchronize the sound at
multiple FM transmitter sites, using a normal Shoutcast or other stream as input.
A few notes:
If you want low latency audio, use ASIO for both input and output. If you don't use ASIO, Windows gives you an extra 100-300 ms of latency, which is far
too much if you are listening to yourself for example on a headphone.
If you use multiple sound cards which don't share the same clock, the output buffer will at some point get underruns (causing drops in audio) or
overruns (causing some pieces of audio to get lost). See Synchronize to output for a (partial) solution.

Display synchronization panel

Synchronize the metering to the audio.

Synchronize with
Select what to synchronize the display to.

If you choose input, events are shown as fast as possible, to match the incoming levels even if the output is sent out with a delay.

Restart sound cards panel

Restart the whole sound card section.

Restart
Restarts the whole sound card section. Try this is you are running into

trouble (buffer underruns, sound cards that refuse to connect).

Normally, in case of trouble this is done automatically, so you should never need this.

Sample rate section

Selects the sample rate used for input and output.

General panel
Sample rate settings.

Sample rate
The sample rate to use for input.

Normally, the output sample rate will be identical to the input sample rate. The only exception is when you are not using ASIO, the samle rate is lower
than 128 kHz and FM output is used.

If you use samle rates above 48 kHz, the audio will be downsampled before processing and upsamled again afterwards. 88.2 and 176.4 kHz are
downsampled to 44.1 kHz, 96 and 192 kHz are downsampled to 48 kHz.

The sample rate has a small effect on the latency and CPU load. At 48 kHz there is a bit more data that needs to be processed, which increases the
CPU load slightly. The latency is a bit smaller because a block of the same number of samples is a bit smaller. If you use 32 kHz or a multiple thereof,
the CPU load is a lot smaller - but you should only use it if you don't need any audio above approx. 14 kHz.

If you want to use FM output with stereo and RDS encoding, and you are using ASIO, then the sample rate must be set to at least 128 kHz - 176.4 or
192 are preferred. If you don't use ASIO for the output, the output sound card will automatically be opened with a high enough sample rate to send out
the stereo and RDS signals.
Important: If you are not using ASIO and you are using Windows Vista, 7 or 8, you need to make sure that the sound card driver is configured to
use the same sample rate that you are setting in Stereo Tool, otherwise Windows will resample it, which causes artifacts any may cause the FM
output to not work at all. Go to Controls Panel -> Sound -> The sound card that you want to use -> Advanced -> Standard setting.

ASIO section
Use ASIO for input and output. Use this if possible!
This gives better control over the sound card leading to less potential problems, and it greatly reduces the delay between input and output.
ASIO panel
The ASIO settings.

ASIO
Use ASIO. ASIO still needs to be enabled per input or output.

See Input, SCA Input, Normal, FM Output, LQ Low Latency.

ASIO Device ID
The ASIO device to be used.
Only one ASIO device can be used in a program at once.

Marian Win7 driver

The Marian Trace Alpha sound card driver for Windows 7

(probably also Windows Vista and 8) causes some systems to lock up completely after running for some time - typically a few days.

The cause of this appears to be something that's timing-related: If the ASIO code in Stereo Tool returns control to the driver very quickly, it happens
more frequently (within a few hours instead of days).

If this slider is set higher, the time before control is given back to the sound card driver is increased, which seems to completely remove the lock-ups.
For non-Marian cards, and for Marian cards on Windows XP, you can safely set this slider to 0 (which slightly reduces the CPU load). The default
setting (3) fixes the problem for all the people who have reported this issue so far; if you run into it anyway, you can try a higher value. Please contact
us if this does not fix the issue.

Open ASIO Control Panel

Fire up the sound card driver's ASIO control panel.

The ASIO control panel is a window provided by the sound card driver which lets you configure certain things, such as the ASIO block size (which
affects latency).

If you want an as low as possible latency, find the lowest latency value here that works without hiccups.

Not all sound card drivers support this button (although you can usually change the settings elsewhere if they don't).

Reduce buffer clicking

Protects against clicks and pops due to buffer underruns.

When the ASIO buffer size is set very small, to minimize latency, in some cases a block of audio may not yet be available when it needs to be sent to
the ASIO driver, which normally results in audible clicks.

This setting lets Stereo Tool predict the data that would be sent to the sound card, and lets it send that instead. This masks most of the clicks and
pops, but not entirely.

It would make sense to have this option enabled at all times, but to more easily hear when the buffer setting is correct, it is advised to turn this
checkbox off when you are setting the buffer sizes, and on again afterwards as some extra protection.

Input panel
Main input sound card settings.

ASIO input Left

ASIO left channel input.

If ASIO is enabled, selecting an ASIO input port here overrules the setting in Input Device ID. As soon as an ASIO input port is selected, ASIO will be
used and the buffer filling display will turn from blue to green.

ASIO input Right

ASIO right channel input.

See ASIO input Left . You can select two sound card input channels for stereo input.

SCA panel
Second audio input, mostly used for SCA audio encoded at 67 kHz.

ASIO input 2 Left

ASIO SCA input #1
If ASIO is enabled, selecting an ASIO input port here overrules the setting in Input 2 Device ID. As soon as an ASIO input port is selected, ASIO will be
used and the buffer filling display will turn from blue to green.

ASIO input 2 Right

ASIO SCA input #2
See ASIO input 2 Left .

Normal panel
Sound card settings for normal (non-FM) output.
 ASIO output Left
ASIO left channel normal output.
If ASIO is enabled, selecting an ASIO output port here overrules the setting in Device ID. As soon as an ASIO output port is selected, ASIO will be used
and the buffer filling display will turn from blue to green.

ASIO output Right

ASIO right channel normal output.
See [[511].

FM panel
Sound card settings for FM output.
Note that if you use ASIO for FM output and the sample rate is not at least 176.4 kHz, some parts of the FM spectrum cannot be sent to the sound card.

Latency panel
Sound card settings for Low Latency Monitoring Output.

Input section
Input audio settings.
You can receive input using ASIO (preferred if your sound card supports it), the standard Windows audio layer, and if you have installed VLC, an audio
stream.

Input panel
General input settings.

Input Device ID
Sound card to use for audio input.
If this is set to Streaming (via VLC), a stream is used for input instead. This setting can be overruled by ASIO input Left and ASIO input Right.

Stream URL
Stream address, if Input Device ID is set to Stream (via VLC)

A valid address of a stream that VLC can decode. This requires that VLC (32 bit version for 32 bit Stereo Tool, 64 bit version for 64 bit Stereo Tool) is
installed on your system. Currently only works in Windows.

If you have a problem with a stream, please try if you can open it in VLC Media Player directly.

If necessary you can supply extra VLC command line arguments after the URL. For example --extraintf logger -vvv will show a logging window that
might be useful to find out what's happening exactly (do not turn this on during normal usage because it can make things unstable). Other options can
be used to control sample rate, buffering and more. For a full list of options (some can cause big problems when using them from within Stereo Tool)
see VLC command line help.

Input Buffer size

Do not use this, leave at 0.

Low input level correction panel

Adjusts the input level if the input level is low.

Input gain
Adjusts the input level to reach around 0 dB peak level.

Many studios use a lot of headroom in their signal, they feed the processor at levels like -24 dB. The built-in presets in Stereo Tool were designed for
input that reaches levels upto about 0 dB regularly. If you feed the audio at -24 dB, certain filters (Declipper, Noise removal, AGC) don't function
properly and the sound will be bad.

Of course, it should still be possible for studios to use a large amount of headroom. With this slider you can adjust the level such that under normal
circumstances the peaks are at about 0 dB. If peaks are occasionally louder, that's no problem - no cliping is performed and no distortion is created.

Balance
Adjusts for different input levels on left and right channel.

If the left and right channel input levels aren't perfectly equal (which might happen for example if you're using some analog equiment), this slider
allows you to correct it. A negative value means that the left channel is boosted, a positive value means that the right channel is boosted. Don't look too
much at the value, instead play some mono audio and adjust this slider until the input levels (which you can see in the VU meters) are equal for left
and right.

Most other devices (including amplifiers) will reduce the level of a channel instead of increasing it. The reason to increase the level is that otherwise
it's more difficult to show both the original and the modified (by Input gain and Balance) volume levels - one channel could be made louder while the
other could be made softer.

Synchronize with different output sound card (not ASIO) panel

Avoids buffer underruns or overruns if different sound cards are used for input and output.
 Synchronize to output
Resamples the input to match the output sample rate.

If you use a different sound card for input and output, chances are that the sample rates do not match perfectly and after some time (hours to days) the
output buffers suffer from underruns (buffer is empty, moment of silence gets inserted) or overruns (buffer is too full, input is ignored and some audio
is lost).

The same problem occurs if you feed the input through a virtual audio cable (VAC, VB Cable).

Enabling this setting matches the input sample rate to the output sample rate.

Note that this uses extra processing power, so if it is not needed you should keep it turned off.

Relative adjust
Controls how fast synchronization works
Synchronization is done by slightly changing the sample rate of the input signal. This means that it has a (small) effect on the pitch of audio. Using a
faster speed means that synchronization works faster, but gets more audible.

The default setting is 1%, which means that the maximum possible adjustment if things are really wrong is 1%.

Resampling quality
Quality of the resampling algorithm. Affects CPU load and quality.

Input tilt panel

Corrects tilt problems in the input signal.
If you have an analog signal path from the studio to Stereo Tool, you might have some tilt issues (a square wave does not look like a square wave
anymore.

Tilt correction lets you correct this, which slightly improves the audio quality (mainly the bass) and also helps the declipper to function optimally.

Correction enabled
Enables Tilt correction.

RC
RC value of the first highpass filter.
Tilt is usually caused by a Resistor and Capacitor (RC) circuit, which acts as a highpass filter. For example, a sound card might use an RC circuit to
remove DC offset from a signal. Unfortunately, this also reduces the audio quality and makes processing more difficult.
The tilt correction filter in Stereo Tool inverts the effect of an RC circuit, thus restoring the original signal (except for the DC offset).

This slider sets the RC value of the RC circuit to be inverted.

RC 2 same as RC
Makes RC 2 setting identical to RC.

In all the cases that we have seen so far, this gives the best results.

RC 2
Second RC circuit inversion value

See RC. This slider allows to invert a second RC circuit.

Note: In many cases, the best results are obtained by using the same value for both RC inversion sliders.

Fade speed
Protection against restoring DC offsets.
To avoid restoring a DC offset, if the inversion of the RC circuits causes a large offset, this slider controls how fast this DC offset is removed. Default
value is 100%, you should probably leave it there.

Input & detilted input panel

Display of the input before (left) and after (right) detilting.

SCA Input section

Second audio input, mostly used for SCA audio encoded at 67 kHz.
Note that this is NOT intended to feed an external RDS encoder.

SCA panel
SCA sound card settings.

Input 2
Enable the secondary (usually SCA) input.
 Input 2 Device ID
Sound card to use for SCA audio input.

If this is set to Streaming (via VLC), a stream is used for input instead. This setting can be overruled by ASIO input 2 Left and ASIO input 2 Right.

If the input is coming from a different sound card than Input Device ID, hiccups may occur on one of the inputs.

Input 2 Separate thread

Use a separate thread for SCA input.

Enable this to avoid hiccups on the main audio if there are differences in sample rate between the sound cards for Input Device ID and SCA Input.

Only enable this if it's really - if possible use the same sound card for both inputs. While this checkbox protects the main audio (Input Device ID) from
hiccups, the SCA signal may still get hiccups.

Non-SCA / RDS panel

Use the SCA Input for other purposes than SCA.

Use SCA Input as external RDS input

Enables external RDS input via SCA.

Internally, the pilot will be synchronized to the incoming RDS signal - any incoming pilot signal will be ignored. If this RDS signal is of really bad quality,
it might cause problems, and if the RDS signal is weak the internal stereo pilot generator will take over. Except for that, the input level does not matter.

No SCA, add this sound card input to normal Input

Combine both inputs and use the result as processing input.

This option was added for the following situation: One location where multiple signals are generated, which consist of a 'shared' part coming from one
studio (which is input via one of the sound cards) and different regional programming coming from different studios (input via the other sound card).

Connecting both signals to the same sound card would cause the regional audio signals to get 'mixed', this option circumvents that problem.

Normal Output section

Settings for normal (non-FM) output.
If FM processing is used, then the normal output contains the signal after demodulating and de-emphasis. It should be similar to the sound that people
hear on their FM radios.

Normal panel
Sound card settings for normal (non-FM) output.

Normal output
Enables the normal (non-FM) output.

Device ID
Sound card to which the normal (non-FM) output should be sent.

If this is set to Streaming (via VLC), a stream is used for input instead. This can be overruled by ASIO output Left and ASIO output Right.

VLC SOUT=
VLC --sout string for stream encoding, if Device ID is set to VLC.
This is a bit tricky and needs some explanation. It also requires that VLC (32 bit version for 32 bit Stereo Tool, 64 bit version for 64 bit Stereo Tool) is
installed on your system. Currently only works in Windows.
VLC is a powerful program which includes a library that makes most of the options available to other programs. While receiving a stream is relatively
simple (see Stream URL, you can just enter the URL and VLC can determine what to do with it), to send an output stream much more information
must be given: Which codec do you want to use, which sample rate and bit rate, stereo or mono, the type of stream and more.
Because VLC is very powerful, it's nearly impossible to include all the necessary settings in Stereo Tool. Instead, we have chosen to use the VLC
command line string to offer the full range of options in VLC, and to automatically take advantage of future VLC releases that might have more options.
How to generate a command line string using VLC. This section describes how to use VLC to generate a command line string that you can past into
Stereo Tool. Some standard strings will be provided below this section.
Steps:
Open VLC Media Player
Open Media &arrow; Streaming.
Select an audio file to play.
Click 'Stream' at the bottom.
You should now see a page showing the file that you selected earlier. Press Next.
Click on the pull-down field right of 'New target', and select the type of stream you want. For example HTTP.
Make sure that Transcoding is enabled.
Select a transcoding profile - for example 'Audio - MP3'. You can click the edit button to fine-tune the settings.
Click Next.
You will now see a command line rule that looks like this:
:sout=#transcode{vcodec=none,acodec=mp3,ab=200,channels=2,samplerate=44100} :sout-keep
The meaning is: No video codec, MP3 audio codec, 200 kbit/s, 2 channel audio, 44100 Hz. :sout-keep keeps the stream open if you switch to a new
file, for Stereo Tool this is irrelevant so you can remove it.
Paste this line in the SOUT field of Stereo Tool, and it should start streaming immediately.
 Some example strings that you can adjust manually to match your own situation are listed below:
OGG/Vorbis Shoutcast streaming: --
sout=#transcode{vcodec=none,acodec=vorb,ab=128,channels=2,samplerate=44100}:std{access=shout,mux=ogg,dst=user:pass@ip:port/access}
access is usually empty. Make sure that you do not remove the trailing / though or it won't work.
MP3 Shoutcast streaming: --
sout=#transcode{vcodec=none,acodec=mp3,ab=128,channels=2,samplerate=44100}:std{access=shout,mux=raw,dst=user:pass@ip:port/access}
--sout-shout-mp3
Please not the extra --sout-shout-mp3 option, if you omit this the stream will open but no data will ever be sent to it.
HTTP MP3 streaming to port 8080 on localhost: --
sout=#transcode{vcodec=none,acodec=mp3,ab=200,channels=2,samplerate=44100}:std{access=http,mux=raw,dst=127.0.0.1:8080}
In case of problems, you can add --extraintf logger -vvv to the options. This will open a logging window that might be useful to find out what's
happening exactly (do not turn this on during normal usage because it can make things unstable). For a full list of options (some can cause big
problems when using them from within Stereo Tool) see VLC command line help.

Signal selection
Selection of the sound that will be played though Normal.
If FM Output plays FM or AM transmitter output, with this setting you can select what Normal will play.
If you select a separately processed output, clipping and limiting will be done separately for the FM Output and Normal outputs. This means that
you can generate a signal suitable for streaming without suffering from the loss in high frequency content caused by FM pre-emphasis, for
example.
Options are:
Input without processing Plays the original input. Can be useful for monitoring.
De-emphasized version of FM output What listeners will hear on an FM receiver.
Separately processed streaming input Separate clipping and limiting for the stream.
Separately processed pre-emphasized left/right FM output Pre-emphasized stereo FM output, without composite clipper.
Separately processed de-emphasized left/right FM output De-emphasized stereo FM output, without composite clipper.

If separate processing is needed, the CPU load will go up. When that's the case, the botton next to the Signal selection pulldown menu will be lit.

Volume
Normal (non-FM) output volume.

Buffer size
Normal (non-FM) output buffer size.
Using a bigger buffer size reduces the chance of getting audio dropouts if the pc is very busy, but it also increases the delay between input and
output. If you want to do other things on the pc that's running Stereo Tool (starting other programs while processing is running etc.), you might need
a big buffer here.
If you use ASIO (ASIO output Left , ASIO output Right), the buffer size can be much lower than if you're not. If you're not using ASIO there are also
buffers in Windows that add a lot of extra delay.

To get the lowest possible latency (see also Latency), set this value as low as possible without getting audio drops, and avoid running other
programs that cause hiccups.

Restart on buffer issues panel

Controls buffer restarting.
If the sound card buffer is nearly empty or nearly full for a longer period of time, which indicates that some sort of problem has occurred and might lead
to hiccups in the audio, this will automatically act as if Restart is pressed.

Auto-restart
Enables the automatic restart feature.

Restart if below
Restarts if the output buffer is filled below this percentage for a longer period.

Restart if above
Restarts if the output buffer is filled above this percentage for a longer period.

Synchronize to input or FM output panel

Synchronizes the Normal to the FM.
If different sound cards are used for FM and non-FM (streaming, HD, DAB, listening) outputs, they will drift after a while, which leads to hiccups or lost
audio. By enabling synchronization, the Normal will be synchronized to match the FM sample rate. This is also needed when one of the sound cards is
a virtual cable.

Synchronize to output
Enables synchronization.

If this is set to Auto, Stereo Tool will compare sound card names to 'guess' if synchronization is needed. If the latency is very low, Auto will never
enable synchronization.

Extra delay (added to buffer size)

Extra delay for synchronization.

This can be used for example to synchronize a DAB or HD signal to the FM signal, with the amount of shift in time that's needed to make them arrive
 simultaneously at a receiver.

FM Output section
Settings for FM output.

FM panel
Sound card settings for FM output.

FM output
Enables FM output.

Device ID
Sound card to which the normal (non-FM) output should be sent.

This can be overruled by Link error '' and Link error ''.

VLC SOUT=
VLC --sout string for stream encoding, if Device ID is set to VLC.
See VLC SOUT=.

Volume (MPX level)

FM output volume. Need to be calibrated for a compliant FM signal!

Separate channel 2
Enables separate volume control for the 2nd (right) channel.
Used when you have multiple transmitters connected to a single Stereo Tool instance.

Volume channel 2
The relative volume of the right channel.

Buffer size
FM output buffer size.
Using a bigger buffer size reduces the chance of getting audio dropouts if the pc is very busy, but it also increases the delay between input and
output. If you want to do other things on the pc that's running Stereo Tool (starting other programs while processing is running etc.), you might need
a big buffer here.
If you use ASIO (Link error '', Link error ''), the buffer size can be much lower than if you're not. If you're not using ASIO there are also buffers in
Windows that add a lot of extra delay.
To get the lowest possible latency (see also Latency), set this value as low as possible without getting audio drops, and avoid running other
programs that cause hiccups.

Test signals panel

Shows the output before (left) and after (right) tilt correction.
Important: You need to make sure that there's enough headroom for the tilt correction. So, if you feed a 15 Hz square wave, the waveform should not
clip. If it does, you need to lower the output level to the sound card (and adjust the transmitter to handle the lower level).

Generate test tone

Send out test tones to calibrate an FM transmitter.

Frequency
Test tone frequency.

The frequency of the test tone if the Link error 'type' is set to Sine, Square or Smooth Square. Otherwise this slider is ignored.

FM Tilt correction panel

Corrects tilt caused by sound card, FM transmitter or cables.
Almost all sound cards (except those by Marian) use a highpass filter in their output to remove and DC offset from the signal. This causes overshoots
when broadcasting a tightly clipped MPX signal. Similarly, some transmitters slowly adjust their frequency when there's a DC offset, which has a
similar effect.
This filter reverses the effect to make it look like there is no RC circuit at all.
Some values measured with this filter on the built-in sound card of a laptop:
Tranmitter with the volume calibrated to give 75 kHz modulation for a 1000 Hz sine wave at maximum level:
15 Hz sine wave: 74 kHz modulation
1000 Hz sine wave: 75 kHz modulation
60000 Hz sine wave: 73 kHz modulation
15 Hz square wave: 91 kHz modulation
After calibrating Tilt Correction:
15 Hz sine wave: 75 kHz modulation
1000 Hz sine wave: 75 kHz modulation
60000 Hz sine wave: 73 kHz modulation
15 Hz square wave: 75 kHz modulation
This YouTube video describes how to set it up:
See this YouTube video for a detailed explanation of how to set this up correctly:

Correction enabled
Enables input tilt correction.

RC
Tilt correction value.

Tilt is usually caused by a Resistor and Capacitor (RC) circuit, which acts as a highpass filter. For example, a sound card might use an RC circuit
to remove DC offset from a signal. Unfortunately, this also reduces the audio quality and makes processing more difficult.

The tilt correction filter in Stereo Tool inverts the effect of an RC circuit, thus restoring the original signal (except for the DC offset).

This slider sets the RC value of the RC circuit to be inverted.

RC 2 same as RC
Makes RC 2 value identical to RC.

In all the cases that we have seen, using identical values gives the best result.

RC 2
Second RC circuit inversion value
See RC. This slider allows to invert a second RC circuit.

Fade speed
Protection against restoring DC offsets.

To avoid restoring a DC offset, if the inversion of the RC circuits causes a large offset, this slider controls how fast this DC offset is removed.
Default value is 100%, you should probably leave it there.

Synchronize FM transmitters panel

Synchronizes audio on FM transmitters using standard Shoutcast etc. streams.
If an FM or AM station has multiple transmitters and RDS is used to automatically switch between those frequencies, it's very important that the signals
are synchronized so that switching between the frequencies is not noticeable for a listener.
This is often achieved using specific hardware on both the sending (studio) and receiving (transmitter) end. This hardware is usually expensive, and it
would be much simpler if a station could just use an exising (free) streaming protocol to send the audio to the transmitters - there's already a PC there
that runs Stereo Tool which could receive the signals.
The problem with most streaming mechanisms is that there is nothing built in to keep the signals synchronized. In fact, even if you start two players on
the same pc, there's often a difference of multiple seconds between the two players. And due to minimal differences in sample rate between the
sound cards in pc's, after a long period (weeks-months) the signal that comes out of the different pc's could be multiple minutes apart.
Stereo Tool can synchronize any type of stream that you can feed to it, as long as it gets the audio as soon as it arrives (it should not be buffered
elsewhere). This means that the plugin version of Stereo Tool can do it, and the stand alone version can do it if you have configured it to be able to
receive streams.
 This video shows how well it works on a station in Belgium
that uses Stereo Tool on their 13 FM frequencies.

Synchronize to output
Enables synchronization.
This increases the CPU load.

Max speed adjustment

Maximum amount of speed adjustment.
See also Relative adjust. This limits the maximum amount of speed adjustment.

Turbo: Speed on start

Faster synchronization after startup of after loss of signal.
During normal operation, the synchronization should be done very prudently. But directly after connecting or reconnecting to a stream, to avoid
taking multiple hours to reach synchronization, it can be done a bit faster. Turbo controls how fast the speed can be adjusted in this situation. Note
that the maximum amount of speedup is still controlled by Max speed adjustment, so if that is not set too high it should still be nearly unnoticeable.
The Relative adjust effect is increased in Turbo mode, so there is no big slowdown when the target point is reached.

Extra delay (added to buffer size)

Compensation of small constant delay differences between sites.

If there's a constant small difference in timing between sites (this is usually not the case), you can correct it here.

Single Frequency Networks (SFN) panel

Sample-exact synchronization between 2 transmitters at the same frequency.

Delay one of the MPX outputs

Enables SFN support.

Channel
Selects which channel to delay.

Invert
Phase inverts one of the channels.

Delay
Time to delay Channel.

Time is measured in samples here.

Restart on buffer issues panel

See Restart on buffer issues.

Calibration section
This is obsolete and should normally not be needed anymore.

LQ Low Latency section

Settings for Low Latency Monitoring Output.

Latency panel
Low Latency Monitoring output settings.
Low Latency output
Enables the Low Latency Monitoring Output.

Device ID
The sound card to use for Low Latency Monitoring Output.
Note that using Low Latency Monitoring Output without ASIO is useless because the delay that Windows adds is far too big.

Volume
The Low Latency Monitoring Output volume.
Must be set low enough to handle spikes because no clipping is performed on this output.

Buffer size
Controls the delay. Set as low as possible.
PNR Noise & Hum section
Automatically learning noise and hiss remover.
Some radio stations have a problem with constant sounds, such as a 50/60 Hz hum from a bad cable (which can be hard to find), a high
pitch tone from a fan or airconditioner, a constantly present hiss etc. PNR Noise & Hiss can learn what the disturbing sound sounds like,
and then remove it.
For constant tones, the removal is nearly perfect, with nearly no side effects, even if a tone is as loud as the actual programming. For non-
constant tones, including hiss, PNR can reduce it by a few dB without causing noticeable artifacts.
See this YouTube video for an example of how to set it up and what it does:

How it works:
First, we measure at least a few seconds of silence (noise only, by enabling Collect Data) to determine what the noise sounds like.
From the measured audio, after filtering out the lowest and highest values using Ignore lowest and Ignore highest, the minimum level,
the average value, the medium value (at position Minimum Median Position) and the variance are determined.
Minimum Multiplier, AVG Multiplier, Link error '2872' and Sigma (variance multiplier) are used to determine how much audio will be
removed. These values may be changed after the measurement.
Removing audio begins immediately after Collect Data is disabled. Use the Difference to check that (almost) no real audio gets
removed.

The average value and sigma (variance) are used to determine how much more audio should be removed than the minimum. If this is
increased too much, artifacts will become more audible.

General panel
Main PNR settings.

Enabled
Enables PNR Noise & Hum removal.

After Declipper (Declip first)

Perform PNR Noise & Hum after the Declipper.
If your material is clipped first and then noise/hum is added, for example if you play clipped CD's on a system that has a hum, the hum
needs to be removed first, and declipper should be done afterwards. In that case, leave this setting off.
But, if for some reason the clipping happened after the noise/hum was added, then you can enable this setting to declip first, and remove
the noise and hum afterwards.

Minimum Multiplier
Multiplier for the minimum amount of audio measured.
For completely constant tones, using 100% here will completely remove the sound, lower values will always reduce less. Normally 100%
is a good value here.

AVG Multiplier
Multiplier for the average level that was measured.
 Sigma (variance multiplier)
Multiplier for the sigma (variance) level.

Sigma Steepness
Steepness when going from Minimum to Sigma filtering behavior.

ANALYSIS STEP panel

The settings here are [b]only[/b] used during analysis!

Ignore lowest
Percentage of lowest measurements to ignore.

Ignore highest
Percentage of highest measurements to ignore.

Collect Data
Enables collection of PNR data.

Enable this on moments of silence (with the disturbing sounds present), so the filter can learn what to remove. You might need to click
RESET ANALYSIS DATA to remove data that was collected earlier. This step needs to be performed at the correct Sample rate, Latency,
Quality (CPU load) and Input gain setting - when any of these settings are changed, the learned information becomes unusable, and the
filter is disabled until either the settings are restored or a new learning stage has been performed.

RESET ANALYSIS DATA

Clears all the learned information.

Always press this when a new Collect Data step will be performed with changed audio (different disturbing sounds).
Declipper section
Repairs clipped audio: Removes distortion and restores dynamics.

Perfect Declipper improves the audio quality of too loud recordings. This includes most modern CD's and MP3's (see 'Why should I use it'
on the right). It does this by calculating the missing (clipped) information from the data that is still available.
Declipping consists of 3 steps:
Input Tilt detection
Clipping detection
Restoration
How well declipping works is determined by the combination of these 3 steps.
Why should I use it?
In the 1980's, CD's were expensive and only available in high-end systems - and hence most CD's were recorded at the best possible
quality. In the last two decades, due to what is called the loudness war, music has been released at continuously increasing volume levels.
This has come at a cost: Reduction of dynamics and clipping. Clipping means that the loudest spikes in the music are cut off, which causes
digital clipping distortion. In the last few years, it has gotten so bad that in some cases you can even clearly hear the distortion on laptop
speakers.
Perfect Declipper can restore the clipped parts of the audio, in many cases the result is indistinguishable from the original, not clipped,
recording. In the process, it also restores part of the dynamics.
Do I need high-end audio equipment to hear the difference?
No! In fact, I've received many emails and messages from people who stated that their cheap equipment suddenly sounded like a high-end
system to them. Even on the cheapest headphones the difference in quality is amazing. Just check the samples on this site.
Is declipping useful if we are going to clip the audio again anyway at the end of the processing?
Yes! Stereo Tool's Advanced Clipper detects whether distortion is noticeable and usually does not cause audible distortion. And if the
declipper output is fed into another processor, especially older processors that were not built to handle clipped audio, the increased
dynamics help a lot to generate a much more constant and much better quality sound.

General panel
Overall declipper controls.

Enabled
Enables the declipper.

Input section
Description of the type of input signal.
Many things in the input path can affect clipped samples. On a CD, it's usually (but not always!) easy to see: Samples at the absolute top or
bottom are clipped, and other samples are not.
But in many cases the audio from the CD does not reach Stereo Tool directly. There can be an analog studio (which can causing tilting of
the audio), recordings can be compressed with a lossy compressor such as an MP3 encoder (try to avoid that if possible! It makes perfect
declipping impossible, although it's still possible to make a big improvement.)
Input Type panel
How the audio reaches Stereo Tool.

Analog input
Check this if the input is analog, or if it has been resampled.

If the input is analog or resampled, the position in the audio where samples are taken 'moves', which makes it more difficult to
distinguish between clipped and not clipped samples.

If this checkbox is checked, if a sample is detected as clipped, the 2 surrounding samples will also be marked as clipped. This does
mean that the set of 'good' samples is reduced, which makes restoration a bit harder. But the opposite, not detecting a sample as
clipped while it actually is clipped, is much worse for a good reconstruction.

Input MP3 panel

Settings to declip MP3 audio.
After MP3 encoding, audio gets spread in time, and if a sample is clipped the samples close to it will also get a part of the distortion. So,
samples close to clipped samples which have a relatively high value all need to be treated as clipped.

Compressed (MP3 etc.) input

The audio is (or can be) MP3 encoded.

MP3 dirty area

The number of samples around a clipped sample that might be affected by MP3 encoding.

MP3 ignore samples near clipping if above

Selects which sanples to mark as damaged by MP3 encoding.

Samples in the MP3 dirty area above this percentage are marked as clipped. Note again that marking more samples as clipped makes
restoration harder, but not marking a clipped sample as clipped is worse.

Input Tilt panel

Settings to detect input tilt.
Input tilt can be caused by an analog path to Stereo Tool, but there are also many CD's which contain clipped and then tilted audio.
See also Input tilt - this is where you should fix a constant tilt in the input to Stereo Tool, which also improves the audio quality. The detection
in this panel only detects the tilt but does not restore it.

Tilt detection
Enables automatic tilt detection.

Stop declipping if tilt not detectable

If tilt cannot be detected, do not declip.

If there's clipping, then there are usually a lot of samples on a line on the top or bottom of the waveform. And that means that the tilt can
be detected. If nothing like this is present, it makes sense to assume that the audio has not been clipped, and to turn the declipper off.

Default tilt
Default tilt. Normally 0 degrees.

If there's a constant tilt present, you can set it here. However, see MP3 dirty area for a better solution that also improves the audio quality.

Lowest tilt
The lowest possible tilt value.

Highest tilt
The highest possible tilt value.

Max tilt adjustment per measurement

The maximum speed at which tilt correction can change per tilt measurement.

Immediately fixing anything that is detected may sometimes cause big errors, so instead the tilt level follows the measurements slowly.

Restoration section
Settings that control the declipping after detection.
These settings control things like CPU load, but also how much of clipping can be restored.

CPU usage panel

Settings that control qualiy and CPU load.

Quality / Precision (increases CPU load)

Conttols how long the declipper keeps calculating to improve the results.

This affects both the CPU load and quality. Higher values remove more distortion (but the effect is small because it's only the last bit of
remaining distortion that's affected).
 Step size reduction (higher = better hiss restoration; increases CPU load)
Affects reconstruction of hiss-like sounds on top of loud bass.
Loud bass in distorted audio can remove long blocks of audio. If this audio is hard to predict (noise-like sounds), it takes more effort to
properly restore it. This slider increases the quality of such highs, but at the cost of a big CPU load increment. From 99% to 98%, the
CPU load drops by half. The step to 96% is another drop by half, the next step to 92% is another drop by a factor two. If your CPU can
handle it and you play tracks with this type of distortion, you might want to set it a bit higher than the default value.

Max. input distortion (increases CPU load)

Determines how much distortion can be restored.

A higher value also causes a slightly higher CPU load.

Peak restoration panel

Settings that affect the clipping restoration algorithm.

Force restored samples to be higher than original

Does not allow restored samples to be closer to 0 than the input signal.

This sounds logical, but in some cases the reconstruction algorithm finds a solution that has lower values than the input. If you force
these to be higher, it can happen that restoration does not sound very good. On the other hand this setting protects longer blocks of high
frequency noise with a bass sound which might not even be clipped, but are incorrectly detected as clipped. In some cases the declipper
can introduce some intermodulation distortion - although other measures have been taken to avoid this. With this setting enabled that
won't happen.

Maximum restored peak level

The maximum increment of a restored sample.

Certain waveforms, for example block wave forms, would cause a nearly-infinite peak level. This slider forces the level to stay below a
certain level, relative to the original level of the sample.

Seting a low value may cause distortion to remain present.

Detection section
Settings that control clipping detection.
A lot of different detection mechanisms are used to determine if a sample is clipped or not. Incorrectly detecting a sample as clipped is bad,
because it reduces the number of 'reliable' samples that we can use to restore the audio. But, incorrectly detecting a clipped sample as not
clipped is worse, because the reconstruction will use bad data.
Because of this, multiple methods are used to detect clipping, which are all designed to detect too many samnples as clipped and not to
miss any. Then, the result of all these methods is combined, and only if all methods agree that a sample is clipped, it will be marked as
such.

Clipping detection panel

Settings that control clipping detection.

Stop declipping if peaks did not increase above

Ignore the declipping result if it did not increase the peak level enough.

Declipping should cause the peak level to increase a bit, otherwise the samples that were detected as being clipped probably were not
clipped in the first place. So in that case, the result of the restoration should be ignored and the original sample values should be used.

Stop declipping if texture did not increase above

Ignore the declipping result if the output is still a flat line.

Declipping should increase the texture of what was a (clipped) straight line. If not, the samples that were detected as being clipped
probably were not clipped in the first place. So in that case, the result of the restoration should be ignored and the original sample values
should be used.

Sample definitely clipped if above

Always treat a sample if it is clipped above this level.

This level is relative to the maximum peak level over a period of time, and can be different at the top and the bottom of the waveform.

For CD input with clipping that has not been tilted, a value of 99% should be used here.

Sample probably clipped if above

A sample is probably clipped if it is above this level.
This level is relative to the maximum peak level over a period of time, and can be different at the top and the bottom of the waveform.

For CD input with clipping that has not been tilted, a value of 99% should be used here - there is no such thing as 'probably clipped' in
that case.

Dynamically increase probably clipped level

Sample probably clipped if above is dynamically increased based on a histogram of sample values.

Sample possibly clipped if above

A sample may be clipped if it is above this level, but not below it.
 This level is relative to the maximum peak level over a period of time, and can be different at the top and the bottom of the waveform.

For CD input with clipping that has not been tilted, a value of 99% should be used here - there is no such thing as 'maybe clipped' in that
case.

Dynamically increase possibly clipped level

Sample possibly clipped if above is dynamically increased based on a histogram of sample values.

Sample not clipped if below

Sample values below this (absolute) level are never clipped.

Make sure that Input gain is set correctly, otherwies this may cause the declipper to stop functioning at very low input levels.

Maximum deviation from straight line

If clipped samples do not seem to form a line, detect them as not clipped.

This slider controls how far the samples may deviate from a straight line.

Sample not clipped if surrounding higher peaks

If a sample is surrounded by higher peaks, it cannot be clipped.

Maximum surrounding higher peaks size

The maximum amount surrounding peaks may be higher than this one.
If this peak is a lot lower and Maximum surrounding higher peaks size is enabled, it will be marked as not clipped.

Peak level measurement: Long term: Ignore loudest

Ignore the loudest few sample values for peak level measurement.

Due to several causes (MP3 encoding, analog input), sometimes a few samples can peak above the clipping level. So it's safer to ignore
the loudest few samples. This slider controls how many of such loud samples should be ignored.
This is a long-term measurement, so the percentage of values to be ignored should be equal to or lower than that for Peak level
measurement: Short term: Ignore loudest.

Peak level measurement: Short term: Ignore loudest

Ignore the loudest few sample values for peak level measurement.
Due to several causes (MP3 encoding, analog input), sometimes a few samples can peak above the clipping level. So it's safer to ignore
the loudest few samples. This slider controls how many of such loud samples should be ignored.

This is a short-term measurement, so the percentage of values to be ignored should be equal to or higher than that for Peak level
measurement: Long term: Ignore loudest.

Declip if
Determines how the maximum level is measured - short term, long term or combined.
Either means that it uses the minimum of the two, both mean that it uses the maximum of the two - in which case the Long term vs short
term margin is used to adjust the short term level.

Long term vs short term margin

Adjustment of the short-term level measurement w.r.t. the long term.

Only detect clipping if noise above and below line is symmetrical

Does not declip if the samples above and below the clipping level look different.

If a sound is clipped and then noise is added (analog audio path, tilt, MP3 encoding), then when looking at the histogram it should
normally look roughly symmetrical above and below the clipping level. If that's not the case, it may be an indicaton that clipping was
detected incorrectly.

Veil section
An extra mechanism to prevent incorrect clipping detection.
A veil is placed on both the top and the bottom of the waveform. It rests on the peaks of the waveform, and drops down between those
peaks. Only samples that are close to the veils can be clipped.

Veil panel
The veil settings.

Veil base elasticity

How fast the veil drops after a peak.

If the material of the veil is very elasical, it can fall down very rapidly, otherwise it moves down slowly.

Veil spike elasticity

How fast the veil drops directly around a spike.
Around a spike, you want the veil to drop faster to not miss any clipped samples.

Veil spike to normal time

Size of the area around the peaks with a different veil elasticity.
Clipped if distance to veil is below
Samples can only be clipped if they are this close to the veil.
Dequantizer section
Bit depth increaser.
The Dequantizer can takes files of a certain bit depth (CD's are 16 bit, but you might also have recordings that use less), and calculate a few
extra bits. It can clean up quantization noise in the process, if the recording was created without dithering.
This process does generate MP3-like artifacts, however the level of those is much lower than that of the noise that is removed.

Enabled
Enables the Dequantizer.

Input
Tells the Dequantizer how many bits the input signal has.

Setting this level too high creates more MP3-like artifacts, setting it too low won't fix all the quantization effects.

Reduction strength panel

Maximum amount of quantization noise removal.
For dithered audio, this level should be set to 2, for non-dithered audio, 1 is sufficient.
Setting this level too high creates more MP3-like artifacts.

Maximum reduction
Maximum amount of reduction per frequency.

Block sizes
Determines whether only big blocks of audio are processed, or also smaller chunks.
Smaller chunks can lead to more accurate removal and a cleaner sound, but also more short-term MP3-like 'chirping' artifacts. It also
uses more processing power.
Delossifier section
The DeLossifier attempts to clean up the effects of lossy compressed (MP3 etc.) input.
Lossy compression (such as MP3 compression) reduces the amount of data needed to store or stream audio by 'throwing away' things that
it deems 'nearly inaudible'. Especially at lower bitrates, this causes very clearly noticeable artifacts that are typical for lossy compression.
Beside the fact that these artifacts are often already noticeable, the processing that is done by Stereo Tool (or any other processor)
invalidates the assumptions that the compressor has made on which effects are noticeable and which are not.
The DeLossifier attempts to detect these artifacts, and to clean them up.

Pre-ringing killer (experimental - big latency!) panel

Attempts to remove pre-ringing from lossy compressed audio files.
One of the typical artifacts in MP3 and other lossy compressed files is pre-ringing. This means that a sound that should start suddenly, has
a sort of 'pre-echo': Instead of starting suddenly you first hear it come up softly, and then there's a sudden jump when the sound was really
supposed to start. This gives the typical 'smeared-out', non-punchy sound on lower bitrate MP3s.
This filter attempts to detect pre-ringing, and reduce it.
Important: Because to be able to determine whether a sound is caused by pre-ringing or it is a sound that should be there, this filter needs
to look ahead a lot, which adds a lot of extra latency.

Enabled
Enables the Pre-Ringing Killer

Difference
Plays only the sounds that are determined to be pre-ringing.
If Left channel only (testing) is enabled, the pre-ringing sound is played on the left channel, and the audio on the right channel is
unaffected. This makes it easy to hear if only pre-ringing is removed (the sounds stop at the moment when a kick is played), or more
sounds are removed.

Maximum relative volume reduction

Determines how strong the pre-ringing killer works.

At 100%, the audio that is determined to be pre-ringing is removed completely. At higher levels, more audio is removed. This may cause
'gaps' before punchy sounds, but it also improves the dynamics further.

Left channel only (testing)

For testing: Pre-ringing is only removed for the left channel.

This also affects Difference.

Noise removal section
Removes background noise, including 16-bit quantization noise.
The noise gate removes background noise, which is mainly useful for older recordings. But even for recent CD's it helps to let it work very
gently, because quantization noise (the number of bits on a CD is not infinite) can be increased a lot by processing which makes music
sound harsh. Gentle Total noise level and Noise gate level per band solve this.

Noise gate panel

General Noise Gate settings

Enabled
Turns the Noise Gate on.

Total noise level panel

Global control of how much noise is removed.

Total noise level

The maximum amount of noise to be removed.
This can be fine-tuned for different frequency ranges in Noise gate level per band.

Noise gate level per band panel

Fine-tuning of the amount of noise to be removed per frequency range.

Band {} noise level

The maximum amount of noise to be removed per frequency range, relative to Total noise level.

FM Hiss (removes FM hiss from INPUT signal!) panel

Removes FM stereo hiss on the input (!) signal.
If you need to rebroadcast audio from another radio station, and reception is not optimal, this filter can remove the FM stereo hiss without
touching the rest of the sound. The sound stays stereo. Even quite severe stereo hiss can be nearly completely removed.

Settings panel
Settings that control the FM Stereo Hiss filter's operation.

FM median pos: Ordered histogram position that determines noise level

Controls hiss vs. tone detection.
A lower value means that sounds are more easily detected to be noise - a too low value will reduce not only stereo hiss but also other
stereo sounds.
A higher value means that even hiss is not detected properly anymore, and noise removal will not work at all anymore.

Hint: Use the Difference checkbox to hear if what is removed is really only the stereo hiss.

Multiply median at most with

Controls hiss vs. tone detection in combination with Multiply median at most with.

Higher values mean that sounds are more easily detected as noise.

Reduction strength of estimated FM hiss level

Determines how much FM hiss can be removed, compared to the measured hiss level.
This normally needs to be set a bit above 100%, because otherwise the hiss is not always completely removed. If it is set too high, other
sounds might be damaged as well.

History size (Response speed)

The amount of time that hiss detection uses.
If this time is short, FM stereo hiss will be removed nearly immediately when it becomes present. However, the detection of noise vs.
tones and the level detection is less accurate. With longer times, it takes longer for the filter to kick in but it has less effect on the sound
and removes the FM hiss a bit more accurately (or actually, the Reduction strength of estimated FM hiss level setting can be reduced a
bit).
Rough/smooth: Median position in ordered history histogram
Controls how to combine the measurements in History size (Response speed).
The histogram of measured noise levels over History size (Response speed) is ordered, and then the value at this position is used as
the noise level.
If the FM stereo hiss is very constant, use a low value. This reduces incorrect FM stereo hiss detection. If for some reason the stereo hiss
is not constant and changes rapidly, you might need a higher value.

Set this value as low as possible until you notice that it starts to miss hiss.
 Pre amplifier
Amplifies the signal before most of the processing occurs.
Declipper, Link error '10182' and AGC if Gating based on volume before Pre Amp is enabled ignore this setting. If you want to correct a too
low sound card input level, use Input gain instead.

Post amplifier
Amplifies the output signal.
This slider amplifies the output signal with a value after all the processing has finished. This is in effect a volume slider. If the Simple
Clipper, Advanced Clipper and/or Hard Limit output is enabled, this value should not be set higher than 0.00 dB (x1.0). Limit Post Amplifier
at 0 dB can be set to enforce this. Otherwise, to avoid distortion, make sure that the output bar display is never completely filled: Too loud
output means that the sound will be distorted.

Limit Post Amplifier at 0 dB

Limits Post amplifier at or below 0 dB.
Normally, there's no good reason to set Post amplifier higher than 0 dB, except maybe when you're testing certain things.

Levels panel
Input and output levels.

Quick sound adjustments panel

A few settings that allow quickly making changes in the sound.

RDS panel
RDS output overview.
Natural Dynamics section
Increases the dynamics for music that lacks dynamics.

Beside clipping (see Declipper), moderm music also often lacks dynamics. For an audio processor, it's much easier to work with dynamic
music that has never been comprssed than to work with already compressed music - it is much harder to make such music sound good.
Natural Dynamics boosts punchy sounds in music, while attempting to avoid boosting other sounds or to boost punch in already very
dynamic music.

Main panel
The main Natural Dynamics settings.

Enabled
Enables Natural Dynamics.

Effect strength
Increases or decreases the effect of Natural Dynamics.

The Effect strength sliders of all the Natural Dynamics bands are multiplied by this value.

Bands panel
Controls the number of multiband compressor bands.

Bands
Controls the content of each band.

Band coupling panel

Controls coupling between adjacent multiband bands.
To avoid very extreme effects from the multiband compressor if certain frequency ranges are nearly absent or very loud in the incoming
signal, the bands can be tied together to stop a single band to move very far away from the adjacent bands.

Band coupling
Coupling between all bands.

This number defines how strongly bands are coupled if they are exactly one octave apart (Frequency doubles between bands). Bands
are coupled stronger if they are closer together and weaker if the distance between them is bigger.

Low Freedom
Ignore coupling for the lowest band.
The lowest band is somewhat special: If you don't allow it to move freely, absense of bass or presence of very strong bass cannot be
handled properly. On the other hand, if you want the output to stay true to the original, that's actually a good thing.
With this slider you can determine how much of band coupling is ignored for the lowest band. Note that since bands are still coupled in
 both directions, the changed value of this lowest band will also have some effect on adjacent bands.

High Freedom
Ignore coupling for the highest band.

The highest band is somewhat special: If you don't allow it to move freely, absense of highs or presence of very strong highs cannot be
handled properly. On the other hand, if you want the output to stay true to the original, that's actually a good thing.
With this slider you can determine how much of band coupling is ignored for the highest band. Note that since bands are still coupled in
both directions, the changed value of this lowest band will also have some effect on adjacent bands.

Flat frequency response panel

This slider helps to keep the frequency response of both sweeps and pink noise flat.
If you play a sweep through a multiband compressor, it happens frequently that the output is louder in some places than in other. Usually, it
is louder around the crossover frequency between bands, although this also depends on the amount of compression.
A good value for Flat Frequency Response can only be found by trial and error, the value that gives the flattest response on sweeps should
be used. It is generally also a good idea to test the response for pink noise; this slider has very little effect on pink noise but it should be flat
as well, except for intentional non-flatness.
Update: With properly setup band frequencies, this slider is not needed. The default frequencies in Stereo Tool have not yet been adjusted
for this. But they will be in the future, making this slider useless for most users.

Flat band tops

Changes the shape of multiband processing bands.

This enables a different band splitting mode with flatter top areas of the different bands, and a different mechanism to keep the frequency
response flat.
The advantage of this is that bands have less impact on each other, which can be used to generate a more stable sound image.

See also Flat tops.

Flat frequency response

The flatness value.

0 does nothing, 100% moves the measurement strength at crossover frequencies from -6 dB to 0 dB. See the thin lines in the Bands
display.

Monitor
Plays only the output of this band.

Strength section
Controls how strongly Natural Dynamics works.

Strength panel
Controls how strongly Natural Dynamics works for each band.

Effect strength
How strongly does the filter work.

Dynamics Detection
Reduces the dynamics boost for already dynamic sound.

The effect is reduced if the signal is already very dynamic. This slider offsets the detection of how dynamic the sound already is. Result is
displayed in the bars on the right (gray dotted part that comes in from the right: The bigger this is, the more dynamic ND thinks the sound
is, and the effect of the filter is reduced as indicated by this gray dotted area).

Initial boost section

The settings for the initial required volume boost estimation.
All the sounds that are above a certain level are increased in level. Punch filter is used afterwards to remove unwanted boosts.

Initial detection - Boosts sounds above average level panel

The Initial Boost settings.

Follow speed
Determines how fast the average audio level is adjusted.

Only sounds above the average level (plus a theshold) can be boosted.

Multiply
Calculated level above the long-term level is multiplied by this.

Stereo Tool calculates how loud the current sound is compared to the 'average' level. 0 = equally loud, 1 = twice as loud etc. This
calculated level is multiplied by Multiply above. Basically, bigger value means more effect.
 Subtract
This is subtracted from the result of Multiply.
Bigger subtract value means that the level where the filter becomes active shifts upwards (sounds must be louder for the filter to do
anything).

Maximum boost
Volume boost may never be more than this.

Punch filter section

Only boost punch sounds.
After Initial boost, all the sounds that are louder than a specific level are boosted. But it's better to boost only sounds that are 'punchy', in
other words, if the volume goes up rapidly. This filter removes boosts found in the first step if the sounds are not punchy.

Punch filter - Removes non-punch boosts from Initial boost panel

Boost only punch sounds.

Filter Punch
Turn this filter on.

You should actually never turn this off, except when you're tweaking the settings of the initial detection.

Multiply
At each location the amount of punch is calculated, the result is multplied by this.

Can be used to shift the response behavior.

Subtract
Subtracts a threshold from the result of Multiply.

A higher value means that far less audio gets a boost.

Rise Time
Over how much time difference do we measure volume differences.

Should be bigger for lower frequencies because a waveform takes more time to go up there.

Spread Time
The area in which we allow the volume to go up.
(see Initial boost) is related to the punch rise time, but may be made a bit bigger.

Reduce for jump above

If the level suddenly increases more than this, reduce the amount of boost.

Smoothing section
Settings to avoid artifacts.
Controls how fast the volume may go up (should be fast!) and down (should not be fast, to avoid bad effects).

Filter smoothing panel

The smoothing settings.

Rise speed
How fast the volume may go up.

Should be slow enough to avoid distortion, but fast enough to keep the punch.

Fadeout speed
How fast the volume may go down.

If the volume goes down too fast, ugly vibrating sounds can emerge. These times should be long, but fast enough to avoid a noticeable
boost of other sounds after a punchy sound.

Bands section
Frequencies and slope steepness of all the frequency bands.

Bands panel

Frequency
The center frequency of each band.
 Slope to {}
Steepness of the left slope of the band.

Less steepness generally gives a more natural, but sometimes harder to control sound.

Flat tops
The level at which the top of this band must be cut off.

If no compression/limiting occurs, or if all the bands are compressed/limited by the same amount, the end result is guarranteed to be flat
in frequency response.

Slope from {}
Slope of the right side of the band.
Less steepness in general gives a more natural, but harder to control sound.

Detection section
Controls the detection of the input level.

Level detection panel

The level detection settings.

Channel separation
Process channels separately, combined, or in between.

At 0%, the two channels will always behave the same. At 100%, they move completely separate of each other.

Detection type
Chooses between RMS or Peak level measurement.

Peak mode can cause quite large reactions to a single small spike in the sound. RMS mode responds more like human hearing does,
but low frequencies seem to be counted a lot stronger than in peak mode, which easily causes pumping.

Look-ahead time
Lets the compresor respond to the sound a bit in the future.
This means that the initial spike of a loud sound gets reduced better, which can give a more natural sound.

The attack of the limiter is already protected, and if you don't use very short attack times for the compressor this probably has little effect.

Non-standard tweaks section

Settings that control compressor/limiter envelope detection.
In a compressor we have 2 things: An 'envelope', basically a line that follows the audio level, and the compressor behavior itself. If the level
drops a lot, release is faster - and this is based on the envelope. Now, if the envelope just follows sample levels, then there will be a lot of
near-0 values (just when a waveform crosses 0) which would cause infinitely fast release behavior. The envelope line needs to be made
such that this doesn't happen.
So, around a peak in the waveform, for the surrounding samples we should not allow the envelope to reach much lower values than the
value of that peak.
That works fine for high frequencies. But if you take a bass, the sample values are dropping slowly and in the valleys the level will still
approach 0. Which still causes issues with release behavior. Because of that, there's some code that measures DC offset and increases
the Base smoothing to something close to infinity when there's more DC offset present. Base smoothing controls how big the area is that's
considered for DC measurement (lower frequency = bigger area). What we are actually measuring here is DC offset in a specific direction
divided by total (absolute) power.

Non-standard tweaks panel

RMS block size

The size of the area around the current sample used to calculate the RMS level.

Bigger values means less precise timing of attack/release behavior, but also less effect from low frequencies (less pumping). Generally,
the RMS block size should be set just high enough to not cause distortion when using the limiters (Threshold level) a lot.

Base smoothing
Controls envelope smoothing around peaks in the waveform.

Lower values may cause distortion, but too high values reduce the precision of the limiters and (to a much lesser extent) the compressor
release behavior.

Bass detection
Controls upto which frequency bass should be detected for Smooth Bass power.

Smooth Bass power

If bass is present (Bass detection), increases the Base smoothing temporarily.

This slider controls how strongly the bass affects the release behavior. Setting it higher means more release slowdown when we see
bass.
If you use this for a single multiband band, then you need less of this because there are less other frequencies that hinder bass
detection.
Phase rotation section
Phase Rotation makes peak levels of asymmetrical sounds lower, to protect the compressors, limiters and clippers.
This protects the compressors, limiters and clipper against certain types of sounds that can easily distort. Examples are voices (especially
female voices) and trumpets. They are very asymmetrical, with big spikes in one direction, and if there's also a bass sound present they
tend to suffer from intermodulation distortion.

General panel
The most important Phase Rotation settings.

Enabled
Enables Phase Rotation.

At start of processing
Perform Phase Rotation before all other filters.
By default, Phase Rotation is performed just before the clipper. If this setting is enabled, it is moved to just before Natural Dynamics and
the AGC.
At low latencies (1024 sampples and below), Phase Rotation is always done at the start of the processing chain.

Behavior panel
Setting that control how Phase Rotation works.
This is difficult to set up, and requires a lot of tweaking. The red spiral display shows the effect of the settings, except for that you'll have to
play a lot of difficult sounds through it and check how they look and how well the clipper handles it.

Start frequency
Phase Rotation works on sounds above this frequency.

Initial speed
Controls how abruptly Phase Rotation starts to work above Start frequency.
Lower values can cause a huge 'bump' in the phase response, which can be noticeable (some people even seem to like it).

Rotations
The amount of "moving" sounds back and forward.

Smooth
Smooths abrupt changes in the graph, and hence in the sound.

Low latency panel

Settings that affect phase rotator artifacts and delay in low latency modes.
These settings effect latencies at and below 1024 samples.

Low Latency Look back

Amont of samples to look back to.

A higher value increases the CPU load, but reduces artifacts.

Low Latency Look ahead (increases latency)

Number of samples to look ahead.
The number of samples to look ahead. This increases both CPU load and latency, but reduces artifacts.

Phase Delay section

Purposely introduce non-phase linear behavior.
This is thought to be useful to create certain bass sounds, but reports on the usefulness of this vary.

Phase Delay Equalizer panel

Configures the Phase Equalizer.
Enabled
Enables the Phsae Equalizer.

Lookahead (better quality, more latency)

Amount of lookahead.
Higher values give a better sound, but also more latency.
AGC section
The AGC filter attempts to make the audio level constant, without causing noticeable changes in the sound.
If does so by - if possible very gradually - increasing and decreasing the audio level. If sudden spikes are present, it can remove those
spikes without lowering the output level too much, to avoid pumping and similar annoying effects.
The AGC never increases the incoming audio level, it only decreases it if it gets too loud. Use Pre amplifier to increase the input level.

Before AGC AGC output

Piece of audio before and after AGC. The output signal (bright green) is much more constant in volume than the input signal.

Main panel
Main AGC settings.

Enabled
Enables the AGC.

AGC Target output level

Determines the target maximum output level.

This is the input level for the next processing step, usually the Multiband Compressor.

Matrix panel
Matrix mode lets the AGC work on Left+Right and Left-Right instead of Left and Right channels. (Experimental)
This makes the stereo signal very constant - and probably way too strong for most purposes. The output contains nearly the same amount
of audio in L+R and L-R, which means that there's an extreme amount of stereo present, so it usually needs to be tamed down a bit
afterwards. This mode is experimental and may have unwanted side effects.
You should probably disable Stereo Boost when you use this.
For FM, Matrix AGC without taming down the stereo tends to increase multipath distortion problems, if you're going to try this out, enable the
Stokkemask filter to compensate for that. And tame the stereo down a bit, otherwise it's really too much.

Matrix mode (L+R / L-R)

Enables the experimental Matrix mode.

Reduce stereo
If set below 100%, reduces the effect of Matrix.

L+R to L-R channel mix

Add a portion of L+R to the L-R channel for the volume measurements.
 This reduces the amount of stereo increment if there's a big difference between L+R and L-R.

Max instead of Mix

Changes the behavior of L+R to L-R channel mix.

Instead of adding a portion of L+R to L-R, use the maximum of a portion of L+R and L-R. This reduces the maximum stereo increment.

Current levels panel

Shows AGC activity.

General section
General AGC settings.

Gate panel
Stops the volume from rising when there is (near) silence.
If the input level is very low (noise, silence), raising the output level might cause annoying effects (increasing noise levels during silence,
followed by a sudden drop when the sound starts again). When using Gating, if the input level is below the configured Gating level, the gain
rise is reduced or stopped (Band 1+2 upspeed and Band 3 upspeed are dynamically reduced).

Gating based on volume before Pre Amp

Determines whether Gating is based on the actual input level, or the input level after Pre amplifier.

Controls the behavior of Gating level.

If this is enabled, the gating level is based on the input level, ignoring the setting of Pre amplifier. In other words, increasing Pre
amplifier does not require changing the Gating level.
If this is disabled, gating just responds to the actual input volume of the AGC - if Pre amplifier is increased, the Gating level also needs
to be increased to get the same behavior as before.

Gating level
Stops the volume from rising when there is (near) silence.
If the input level is very low (noise, silence), raising the output level might cause annoying effects (increasing noise levels during silence,
followed by a sudden drop when the sound starts again). This slider determines that if the input level is below the configured Gating
level, the gain rise is reduced or stopped (Band 1+2 upspeed and Band 3 upspeed are dynamically reduced).
See also Gating based on volume before Pre Amp.

Spikes panel
Handling of sudden burst of loud audio.
The AGC slowly adjusts the level to keep the average level constant. This section overrides the standard AGC behavior to immediately lower
the level very rapidly if the volume suddenly increases a lot. Without this, the AGC output would in some cases still contain very loud audio.

Sudden volume jump protection: Threshold

If the volume suddenly increases by more than this amount, sudden burst protection is activated.

If you set this level too low, sudden burst protection will be activated too soon and there will be a loss of dynamics. If it's set too high, loud
bursts will be let through to the next processing steps. You can see that loud burst protection is active in the output meters - when burst
protection is used a piece of the meters gets a different color.

Sudden volume jump protection: Slope size

Slope from Sudden volume jump protection: Threshold to full burst protection.

The slope adjusts the size of the area in which more and more protection occurs.

Behavior panel
Some specific settings to make the AGC work better.

Allow louder female voices

Avoids AGC level drops caused by loud female voices in songs.

In some tracks with really loud female voices (for example, many Celene Dion songs), the AGC will lower the volume for loud notes,
causing pumping. This setting detects such sounds and lowers them before adjusting the level.

This has been superseded by the much more powerful Side chain.

ITU-BS.1770 panel
ITU-BS.1770 loudness metering compliant mode.
ITU-BS.1770 is the base upon which the R128 loudness level metering has been built. In ITU-BS.1770 mode, the AGC adjusts the level
based on how human hearing works, instead of the actual power of the audio. Bass is counted less strong, and higher frequencies are
counted stronger.

ITU-BS.1770 Head correction

 AGC responds more to high frequencies because they sound louder to humans.

ITU-BS.1770 Bass correction

AGC responds less strong to bass because to human ears they seem to sound less loud.

BS412 panel
Prepares the AGC for the BS412 levelling which is required for FM stations in some European countries.
Among others, stereo (L-R) sounds are treated as less important than mono (L+R) sounds, because in the BS412 level measurement
that's also the case.

Prepare for BS412 if BS412 limiter enabled

Turns BS412 preparation on.

Misc panel
Some settings that normal users don't need.

Startup input level

Sets the AGC level upon program startup.

For normal processing, it doesn't really matter where the AGC starts, but if you use Stereo Tool to master music for example, and you use
Stereo Tool as a VST plugin, then you might want to be able to configure the start level of the AGC to avoid unwanted volume effects in the
first few seconds after start of processing.

Up push strength near minimum reduction

Lets the AGC move towards 0 dB (no action) faster when it gets close to it.

Bands section
Select 1, 2 or 3 bands AGC.
[b]While the AGC offers upto 3 bands, since the Side chain was added, the best result can be obtained using only 1 band comgined with
Side chain. The rest of the text is kept here to explain what the other settings do, but we strongly advise to use 1 band with Side chain.
1 band gives the best approach of the total RMS volume. However, loud bass sounds will cause other frequencies to be dropped (which
makes sense, as they count as part of the RMS volume).
2 bands sounds more constant. Band 1 contains all the sound (hence behaves identical to the 1 band AGC), band 2 contains frequencies
above 200 Hz. There are 2 issues when using 2 bands:
The volume of the two bands may move apart, causing the audio to sound different.
In the 2nd band, because very low frequencies are ignored, loud higher frequencies such as loud voices in music may cause the volume
of band 2 to drop.

To solve these issues, the band 1 volume is not dropped below the band 2 volume unless the bass level is really loud, and the band 2
volume is not dropped below the band 1 volume to protect against volume drops on loud voice sounds. To configure this behavior, see the
sliders Raise band 2 output level above band 1 if its volume drops below, But never raise band 2 more than this above band 1 and Keep
band 1 at band 2 level if it stays less than this above band 2.
3 bands is identical to 2 bands (see the previous paragraph), except that very loud highs are reduced. This time, also the level of the 3rd
band is never increased above that of band 2. Reduce band 3 further if its volume gets above is used to set the target maximum highs
level.

Bands panel

Bands
Controls the content of each band.

Band 1+2 upspeed

Determines how fast the volume level is increased when the output level is below the AGC Target output level.

When the output volume has been lowered due to too loud sounds, this slider determines how fast the output volume can be increased
again. A higher value means that the average output level gets closer to the target level, but may also cause pumping. A low value may
cause source material with big volume changes to come out too soft on average - and the quieter parts will stay very quiet.

Band 1+2 downspeed

Determines how fast the volume level is decreased when the output level is below the AGC Target output level.
When the output volume would be louder than the set maximum AGC Target output level, this slider determines how fast the output
volume is reduced. A higher value means that the average output level gets closer to the target level, but may also cause sudden volume
drops when very short loud spikes occur. A low value may cause spikes to remain when the sound suddenly increases a lot. See Spikes
(and also the no longer advised Remove remaining peaks above) for a solution for that.

Remove remaining peaks above

Removes short volume spikes that remain at the end of the AGC processing.

Superseded by Spikes, use that instead for better results.

The AGC responds slowly to volume changes, to keep the effects on the audio as small as possible. This does mean that if the volume
suddenly increases a lot, a loud 'spike' of sound can remain. This slider determines how much 'spike' is allowed above the configured
Band 1+2 upspeed; anything louder than that is reduced.
 If this slider is set too high, loud spikes remain; if it is set too low, too much spikes are removed, which takes out 'kicks' from the audio,
making it sound too 'flat'.

If this setting causes peaks to be removed, black bars are displayed in the output bars at the bottom of the window. Ideally, these should
only occur when they are needed (sudden volume jumps), not during 'normal' music (like every bass kick).

Compatibility & behavior panel

Some recent improvements to the AGC behavior can be turned on or off here for compatibility.
Older presets that were made before these improvements were added still work as intended if these settings are kept off. For newer
presets, it's a good idea to turn them on.

Improved Release
Enabled improved Release behavior.

Originally, if the input volume dropped a lot there release could get incredably fast, especially if the AGC was at a very low level. This
means that different input levels cause different AGC behavior, which is bad. Improved Release fixes this. The only reason that this
setting is available is that older presets were made without this feature, and they still need to work properly.

Block size
Adjust the RMS measurement block size.
With a smaller block size, staccato-sounds are better adjusted in level - the volume is lowered more (the silence is more or less ignored)
which better corresponds to how human hearing works.

Band links section

Configures linking between bands and channels.

Stereo panel
Configures linking between left and right channel.

Channel separation
Determines how independent the two channel volumes can move.
If both AGC channels behave completely independent of each other, a loud tone on one channel may cause strange stereo effects
because other tones are reduced on one channel, but not on the other. On top of that, the total audio content changes if this happens.

If both AGC channels do exactly the same, a loud tone on one channel causes volume drops on the other channel, which can also be
unwanted.

This slider allows choosing an intermediate setting.

Combine remove remaining peaks channels

Lowers both channels if one needs to be lowered by Remove remaining peaks above.

Without this, a drop on a single channel can sounds really bad, especially when listening to headphones.

Band links panel

Links between bands.

Raise band 2 output level above band 1 if its volume drops below
Configure band 2 protection against volume drops due to loud mid or high frequencies.

See Bands. This slider is used to tell the AGC how loud band 2 is expected to be compared to band 1. Band 2 is processed with a lower
Target RMS level, based on this setting. Normally this should lead to roughly identical dynamic amplification levels for band 1 and 2. If -
due to loud mid or high frequencies - band 2 is much louder than the configured level, its output volume is not dropped below the output
volume of band 1. This means that, in cases where relatively loud mid or high frequencies are present, the 2 band AGC starts behaving
more like a 1 band AGC, which gives better protection against unwanted volume drops and rises.

But never raise band 2 more than this above band 1

Configure band 2 protection against very loud highs relative to the lows.

See Bands. If there are only bass sounds present, the band 2 output level could rise indefinitely, while band 1 would be kept very low.
This greatly increases noise levels. (For example, if the bass in the input is reduced by a factor 40, and the highs are not reduced at all, in
total the highs are 40 times louder than the lows). This slider configures how much the band 2 output level can rise above the band 1
output level.

Keep band 1 at band 2 level if it stays less than this above band 2
Configures how much extra bass is needed to drop band 1 output level below band 2 output level.
See Bands. If the bass is just a bit too loud (the band 1 output level would drop slightly below the band 2 output level), keeping the band 1
output level equal to the band 2 output level gives much better results, because it better preserves the original audio content. But if the
bass gets very loud, it does need to be dropped. This slider configures how much louder band 1 may get before its output level is
reduced.

Reduce band 3 further if its volume gets above

Configure the maximum high frequency RMS level.

See Bands. If 3 bands are used, this slider configures the maximum high frequency RMS level, relative to the AGC Target output level.
 Note that band 3 will never be louder than band 2, so setting this to 100% makes the 3 band AGC equal to the 2 band AGC. Setting it to
0% completely removes the highs.

Side chain section

Lets certain frequencies be counted more than others by the AGC.

Side chain panel

PEQ Sidechain
Lets certain frequencies be counted more than others by the AGC.

See for example this image:

As you can see, low bass frequencies are ignored. This means that in a track, if suddenly the bass kicks in, the level is hardly adjusted.
Similarly, frequencies between 600 and 2000 Hz are counted a lot less strongly, which helps to keep the level constant for loud female
voices. This avoids pumping.

Power Bass section

Adds a lot of low end bass to (mostly older) tracks that lack such bass.

This results in a much deeper and warmer sound which is now present for both old and new music. On tracks that already have a lot of
deep bass it has little effect.

Power Bass panel

The Power Bass settings.

Power Bass
Enables Power Bass.

Maximum extra bass

The maximum amount the low bass may be increased.

Bass frequency
The frequency upto which Power Bass functions most strongly.
 Slope (steepness)
How fast the Power Bass response drops after the Bass frequency.

Boost release
How fast the bass AGC can come back up after being kicked down.

This value is relative to the AGC Band 1+2 upspeed, so 10.0 means that this release can go 10 times as fast as that setting.

Ignore side chain

Ignores any side chain for the Power Bass behavior.

In some cases, side chains have so much impact on the bass that Power Bass will always boost the bass, even when it's not needed at
all. This setting should probably always be on (and will probably be removed in a future version).

Power Highs panel

Adds a lot of high end to (mostly older) tracks that lack high end.
This results in a much brighter and more constant sound for both older and newer tracks.
It works the same way as Power Bass, but for high frequencies.

Maximum extra highs

The maximum amount the highs may be increased.

Highs frequency
The frequency above which Power Highs functions most strongly.

Bass AGC (Old) section

Should no longer be needed.
This was used to protect the clipper against too much bass, but with all the improvements to the multiband compressor, limiter and clipper
and the introduction of Power Bass since Bass AGC was created, it should probably not be used anymore.

Compressor section
Singleband compressor locked to the AGC output.
Reduces the dynamic range of the audio and limits it.
Volume compression (A.K.A. audio level compression) reduces the dynamic range of a sound. This means that loud sounds become
softer, and soft sounds become louder.
Limiting limits the maximum audio level below a certain threshold.
For a lengthy discussion about compression, see Wikipedia: Audio level compression.
The compressors and limiters in Stereo Tool are protected against causing distortion. So very aggressive settings and large amounts of
limiting can safely be used.

General panel
General compressor/limiter settings.

Enabled
Turns the compressor/limiter on.

Compressor type
Analog or Digital compressor type.

The Analog compressor type is intended to replace the Digital one. It's behavior is generally more natural, so if you are starting on a new
preset, it's probably a good idea to use Analog mode. On top of the better end result, it also uses far less processing power.

Quick adjust panel

Achieve several effects with a single slider.
Most of these sliders impact the value of several other sliders that are described below, to achieve a certain effect.

Aggressiveness (hot)
Makes attack and decay faster or slower. Sound is more squashed.

This slider adjusts both attack and decay to have more aggressive compression.

Levels panel
Attack, release and ratio.
These are standard settings that almost every compressor has.
 Threshold level relative to AGC
The input level above which the compressor becomes active.
This level is relative to Target output level.

Knee
Makes the transition around the threshold more smooth.

At the threshold the response to slightly different input levels changes abruptly. Knee smooths the transition.

Attack panel
Attack settings.

Attack
The time a 86% volume reduction due to a higher input level takes.

If the input level increases a bit, the volume goes down more slowly than if it increases a lot. This means that it's not possible to give a
value in dB/ms.

Attack Flatness
Lets the compressor respond faster to small differences and slower to bigger ones.

Small differences in level are thus quickly compensated, with helps to reach the target level much faster. And the compressor attack
responds less aggressively to big volume changes.

Limit panel
Limiter settings.

Limit level
The maximum output level of the limiter.

Limit speed
Determines how fast the limiter goes up and down.

Setting this faster reduces intermodulation distortion and pumping, but increases harmonics distortion. Also, for faster levels the output
is louder.

Limiter distortion
Allows the limiter attack to distort.

Some people like this effect, especially on low frequency audio - bass kicks get a special type of 'edge'.

Limiter max release

Controls the release behavior of the limiter.

The limiter attack is always as fast as possible without causing distortion. The same is true for release, but in some cases the release
behavior can be too prudent. This slider overrides the standard limiter release behavior: If the release behavior that would be used based
on the adaptive algorithm is slower than this, the configured release time is used instead. This does mean that very fast release times
can cause some distortion.

Limit before compress

Protects the compressor against big spikes which are limited anyway.

If this is enabled, if a sound will be limited, the compressor will act as if the signal is limited before entering the compressor. As a result,
it will go down less fast on sudden loud sounds.

Release
The time it takes for the output level to climb by 10 dB if the input level falls silent.

Release hold time

Time for the 'brake' on the release to fade out.

When attack has been active, release is not immediately activated to avoid excessive movement. Instead, the release is held back for a
while. This slider determines how long.

Release Flatness
Lets the compressor respond faster to small differences and slower to large ones.
Small differences in level are quickly compensated, with helps to reach the target level much faster as long as differences in level are small.
This gives a much more sparkling, 'alive', sound. But... Big differences are less quickly compensated. See Release Inertia for a solution for
that.
Another explanation to further clarify things: In the compressors, if there's a volume change, it takes quite long for the level to 'stabilize'.
That's because the closer the actual level gets to the 'target' level, the slower it moves (the shape is asymptotic). Something similar
happens in release. This seems to be a good thing, and traditionally this is what compressors do.
What Flatness does is:
If the difference in level is 6 dB, nothing changes
If the difference in level is less than 6 dB, for Flatness values > 1 the change speed is increased.
If the difference in level is greater than 6 dB, for Flatness values > 1 the change speed is decreased.
More technical: The Flatness'th root of the difference in level is used - so for 2 that's the square root etc.
What this means: The higher the Flatness value is, the more the movement to the new level will look like a straight line instead of an
asymptote.
 Release Inertia
Adjusts release behavior to match human hearing for more natural results.
Without Inertia and Release Flatness, after a very big volume spike the speed at which the audio returned was always the same - but
determined by how much it had to move up. So, if the volume dropped by 6 dB and after 100 ms the volume went up 3 dB, then for a volume
drop of 12 dB that would be 6 dB. Sounds perfect.
But it's not. Say you have a huge drop, for example after a very loud 'S' in the high frequency band, where normal volume differences are at
most a few dB and this S suddenly sticks out 20 dB. For a difference of 4 dB, after 100 ms the difference in level is 1 dB - 75% of the
difference is reduced. Now, this last 1 dB is really nearly unnoticeable, so for your ears the release kinda stops after 100 ms. But, for a
difference of 20 dB, after 100 ms the difference is still 5 dB! And you need more than another 100 ms before you reach this 1 dB point.
So, after a loud sound you hear a gap at settings that sound good for small volume differences.
Release Flatness helps a lot for the final part of release: Small differences get compensated faster. But at the same time, bigger differences
take longer to recover, which causes the same effect for really big differences as before.
Inertia fixes this. With inertia combined with Release Flatness you can make the release happen in a nearly constant time, without the
slowdown at the end that you would have without Release Flatness, but also without the slower recovery for very big volume differences.
Basically, the release happens in a nearly straight line, but the slope of the release depends on how much level must be compensated.
With high Inertia values, release can even be faster for very big differences than for smaller ones, which can be good to quickly fill up the gap
after a loud sound.
For bigger Gamma values you need bigger Inertia values.
In case things are not yet clear now, here's another explanation: For release, especially large differences must be compensated very fast -
for 2 reasons:
Big differences mean very dynamic input, and for more dynamic input it's good that more compression occurs.
If you have a loud sound, and it takes multiple seconds for the level to get back, that sounds really bad.
Example:
Sound drops by 4 dB. When 3 dB has been restored, you really won't hear much difference anymore in level.
Sound drops by 40 dB. Now, when 39 dB has been restored you really don't hear much difference anymore.
So in one case when 75% restoration is there we're good, in the other we need 97.5%. And since - without Release Flatness - the behavior
is asymptotic, reaching 97.5% takes multiple times as long as reaching 75%. Higher Release Flatness values only make things worse.
Why is this bad? Well, it makes it nearly impossible to find a good Release (time to raise 10 dB), what works well for small differences will
be far too slow for big differences, and what works well for big differences will sound very aggressive on small differences.
So, the time it takes for the level to be restored to a level where human hearing stops to notice a difference - say 1 dB below the target level -
must be nearly constant.
Inertia ('heavyness') makes sure that once release is moving up, the speed won't slow down until the target is reached. For big drops the
effect is much bigger than for small drops, which is exactly what is needed.
Release Inertia and Release Flatness must be configured to work properly together. The best way to do this is to record a sample with
different level tones (Loud - soft, loud - less soft, loud - just a little less loud), and check if all take approximate the same time to reach a level
slightly below the target level.
Analogy
If you have to drive 10 meters, you just barely hit the gass and drive very slowly.
If you have to drive 1 km, you hit the gass and speed up (Release hold time), then release the gass and let the car roll slowing down towards
the end.
With Inertia, you would not release the gas until you're very close to the end and then hit the brakes to stop.

Continuous Release
The size of the area around the current sample used to calculate the RMS level.

Bigger values means less precise timing of attack/release behavior, but also less effect from low frequencies (less pumping). Generally,
the RMS block size should be set just high enough to not cause distortion when using the limiters (Threshold level) a lot.

Gate level
If the input level is lower than this, release is slowed down.

Detection panel
Compressor detection settings.

Detection type
Chooses between RMS or Peak level measurement.

Peak mode can cause quite large reactions to a single small spike in the sound. RMS mode responds more like human hearing does,
but low frequencies seem to be counted a lot stronger than in peak mode, which easily causes pumping.

ITU-BS.1770 Bass
Respond less strong to bass because to human ears it seem to sound less loud.

ITU-BS.1770 Head
Respond more to high frequencies because they sound louder to humans.

Feedback
Chooses between feed forward and feedback mode.
In feed forward mode, the input is used directly for the measurement. In feedback mode, the output level is measured instead of the input
level.

Feedback mode is known to sound more natural, but the level control is far less accurate. For example, say the input level is 6 dB too
 loud and the ratio is 1:1000. Then in feed forward mode, the level will be reduced by about 6 dB. But in feedback mode, once the level is
reduced by about 3 dB, the compressor will 'see' that it needs about 3 dB of reduction and not reduce the level further.

Ratio
Determines how strongly the compressor responds to changing input levels.
Say, at one moment a sound comes in at the threshold level, so nothing happens to it. If another sound comes in at 6 dB above the
theshold level, the input should be reduced by half. The ratio indicates how much of the increase in input level is not removed. At a the
lowest ratio (1:1), the compressor is basically disabled. At the maximum ratio, 1000:1, 1/1000th of the increase is kept.

Channel separation
Process channels separately, combined, or in between.

At 0%, the two channels will always behave the same. At 100%, they move completely separate of each other.

Look-ahead time
Lets the compresor respond to the sound a bit in the future.

This means that the initial spike of a loud sound gets reduced better, which can give a more natural sound.

The attack of the limiter is already protected, and if you don't use very short attack times for the compressor this probably has little effect.

Non-standard attack panel

Protection against spikes for slow compressor settings.
Some presets use very slow attack and release times. This can sound great, but the level control for sudden volume increases is less
good.
This section contains the settings for an extra compressor that takes over in such cases. It does not affect normal audio.

Loud burst protection

Enables loud burst protection.

Level difference
Increases the level of the 2nd compressor with faster attack.

Because the attack is so much faster, the audio level of the 2nd compressor is generally a bit lower. If we would take the minimum of the
two, we would always look at the 2nd compressor, but that should only happen in extreme cases. By increasing the output level and then
taking the maximum of the two, the 2nd compressor only has an effect on the sound if its output level is quite a bit lower. For example, if
this value is set to 2.00, the 2nd compressor will not kick in if the level difference is less than 6 dB.

Minimum drop
Disables the 2nd compressor if the attenuation didn't suddenly drop a lot.

The 2nd compressor should only be active if there's a huge difference between the volume when using a normal and very fast attack, but
that's not all - if you play very dynamic music it should not kill the punch. This slider controls how much the attenuation must have
suddenly dropped (in the fast attack 2nd compressor) for it to be taken into account.

Fast Attack
The fast attack time.
To be useful, this must be a lot smaller than Attack - typical values are around 1-5 ms.

Release speedup
Controls how much faster the 2nd compressor release is.

Beside a faster attack, the release for the 2nd compressor can also be made faster. This helps to prevent long-term volume drops after a
short loud spike in the sound. This value controls how much faster the release is than Release (time to raise 10 dB).

Non-standard release panel

Experimental settings. Should probably not be used.

Dynamic release
Dynamically increase the release speed if the volume drops more.

If this is set to 0 the release always runs at exactly the same speed. A similar effect can be reached with Release Flatness.

Dynamic release to 0 dB
Release acts as if the input level is always at 0 dB.
So the release speed depends only on how deep the level has dropped. See also Continuous Release.

Non-standard tweaks panel

Settings that control compressor/limiter envelope detection.
In a compressor we have 2 things: An 'envelope', basically a line that follows the audio level, and the compressor behavior itself. If the level
drops a lot, release is faster - and this is based on the envelope. Now, if the envelope just follows sample levels, then there will be a lot of
near-0 values (just when a waveform crosses 0) which would cause infinitely fast release behavior. The envelope line needs to be made
such that this doesn't happen.
So, around a peak in the waveform, for the surrounding samples we should not allow the envelope to reach much lower values than the
value of that peak.
That works fine for high frequencies. But if you take a bass, the sample values are dropping slowly and in the valleys the level will still
approach 0. Which still causes issues with release behavior. Because of that, there's some code that measures DC offset and increases
the Base smoothing to something close to infinity when there's more DC offset present. Base smoothing controls how big the area is that's
considered for DC measurement (lower frequency = bigger area). What we are actually measuring here is DC offset in a specific direction
divided by total (absolute) power.

RMS block size

The size of the area around the current sample used to calculate the RMS level.
Bigger values means less precise timing of attack/release behavior, but also less effect from low frequencies (less pumping). Generally,
the RMS block size should be set just high enough to not cause distortion when using the limiters (Threshold level) a lot.

Peak smoothing
Controls envelope smoothing around peaks in the waveform.
Lower values may cause distortion, but too high values reduce the precision of the limiters and (to a much lesser extent) the compressor
release behavior.

Bass detection
Controls upto which frequency bass should be detected for Smooth Bass power.

Smooth Bass power

If bass is present (Bass detection), increases the Base smoothing temporarily.
This slider controls how strongly the bass affects the release behavior. Setting it higher means more release slowdown when we see
bass.
If you use this for a single multiband band, then you need less of this because there are less other frequencies that hinder bass
detection.

Side chain section

Lets the compressor respond more or less to certain frequencies.
For example, you can boost the bass in the side chain to increase pumping caused by bass - or you can reduce it to reduce bass pumping.
For the compressor that comes after the AGC, it's probably a good idea to have a side chain that looks like the AGC Side chain.

Side chain panel

Stereo section
Several filters that repair and change stereo separation.

AZIMUTH section
Reparation of AZIMUTH (phasing) errors.
AZIMUTH errors are often present in tape recordings, and also on some cheap CDs. Phasing problems causes playing a recording in
mono or through a surround system to result in very ugly artifacts. But even normal stereo playback may sometimes sound a bit unpleasant.
The phasing offset is automatically detected and removed by this filter.
This filter only works properly if the sounds at the left and right channel are similar. If this is not the case for a longer period, the azimuth
correction will slowly be reduced.

AZIMUTH panel

Azimuth correction enabled

Enables AZIMUTH correction.

AZIMUTH limit
Maximum tape head displacement (assuming cassette tapes) to be detected and resolved.
0 disables this filter. Suggested value: 40 m.

AZIMUTH change speed

Maximum speed at which the filter follows detected phasing errors.
This slider controls how fast the azimuth correction can change if the measured displacement is differs from the correction that is taking
place.

Use a very low value ( 0.1 m) to correct for a constant azimuth correction.

Use a higher value (~ 0.4 m) to also correct for rapid changes.

To avoid getting too much effects from measurement noise, use the lowest possible value where the blue line (the actually performed
correction) can keep up with the red dots (the measurements).

Stereo Boost section

Stereo Boost is the preferred filter to add stereo separation to music.
It basically just uses a multiband compressor to increase the level of the L-R channel. It can be configured to never reduce stereo
separation (in tracks with completely separate channels) and it does also not create too much (anti-phase) boost of the L-R channel.

Stereo Boost panel

Stereo Boost enabled

Enables the Stereo Boost filter.

Protect against excessive reverb

Removes sounds (usually reverb) made louder by a strong boost of the L-R channel.
Reverb often gets increased a lot when using strong Stereo Boost settings. Some people like this - but if you want to keep the audio
closer to the original, turn this on.
In extreme cases, this can cause artifacts.

Never reduce below original level

Never allow the stereo separation to get reduced by Stereo Boost.

Stereo Boost strength

Configures how strongly Stereo Boost works.
Higher values sound 'wider', but also increase certain sounds (see Protect against excessive reverb) and may cause multipath
problems with FM reception (see Stereo multiplier, Multipath clipper and Stokkemask FM).
 Stereo Boost maximum amplification
The maximum increase of the L-R channel.

Stereo multiplier
Multiplies the L-R channel with this value.
This can be used to reduce extreme stereo separation. For example, if Stereo Boost strength and Stereo Boost maximum
amplification are increased and this slider is decreased, the stereo separation for most tracks should not change much, but extremes
will be removed.

Values below 1.00 can be beneficial when using Multipath clipper and Stokkemask FM.

Stereo Image section

Stereo Image gives independent control over phase differences and instrument placement. But it can easily create artifacts.
The most interesting feature is to convert stereo to mono without cancellation (loss) of sounds. This creates a sound that is as full as the
stereo image, but it is in fact mono. This can be done by setting both Image phase amplifier and Image width amplifier to 0.00. If you are
not located between the speakers, when you press the MONO switch on your audio system you often still hear a big difference, because a
lot of the sound disappears. After setting these two sliders to 0, that doesn't happen anymore. Example uses are radio stations (FM, AM,
streaming) that broadcast in mono, and people who are deaf in one ear who want to listen with headphones.
Several sliders of the Stereo Image manipulator cause severe artifacts and should only be used as described. They are marked as 'partially
deprecated'. Please read this section carefully if you enable the Stereo Image filter.

Stereo Image panel

Stereo Image enabled

Enables the Stereo Image filter.

Center bass
Prepares the bass to be played on a system with a single subwoofer.

If there is a phase difference in the bass between the left and right channel, and the sound is played using a single bass speaker, the
bass will get deformed and lowered in volume. If this filter is turned on, phase differences for bass sounds are removed completely,
which solves this problem. This occurs only very rarely. Note however that when the Phase shift slider is set to a high value, this will
occur much more frequently. When listening with headphones, this somewhat reduces the stereo effect.

In some cases this filter can cause severe artifacts.

Phase shift
Adds a constant phase offset between the two channels.

This slider can be used to add a phase offset between the channels. Both -180 and +180 cause the channels to be the opposite of each
other, 0 is normal output.

Image phase amplifier is performed first!

Image phase amplifier

Increases or decreases phase differences between the left and right channel, without touching instrument placement.

Moves between 0 (no phase differences between the channels), 1 (no change) up to 8 (8 times as much phase difference as in the
original signal). 0 is VERY useful for converting to MONO, the resulting sound can be downmixed to mono without any distortion or loss of
sounds, which occur in normal stereo to mono conversion. This creates a much fuller and undistorted mono sound. Note that "0" does
not mean that the output signal is mono, because the instrument locations are not affected by the phase slider. To get mono sound, also
set Image width amplifier to 0.

This filter creates artifacts, mainly for values above 1.00. When playing compressed audio, especially lower (< 192 kbit/s) bandwidth MP3
files, setting phase to a high value will also very strongly amplify the already present MP3 encoding artifacts, which results in a very poor
sound quality.

Image width amplifier

Changes the placement of sounds without changing their phase differences.

Moves between 0 (all sounds in the center, 1 (no change) up to 8 (the sounds are moved 8 times further away from the center than in the
original signal, if possible of course). Note that "0" does not mean that the output signal is mono, because the phase differences are not
affected by the width slider. With this setting, for someone who hears both speakers it still sounds like stereo, but if someone hears only
one speaker all the sounds are present.

Setting width to a very high value will almost always introduce artifacts, so it should be used with care - or not at all.

Image phase amplifier maximum separation strength

Sets the maximum allowed level separation per frequency.

Could be useful for example for FM radio stations, this can ensure that the maximum phase difference stays below a certain level, which
reduces signal loss when a receiver switches to mono.

This setting currently introduces artifacts. It should not be used unless it is really necessary. Changes may be made to it in later versions.

Image phase amplifier maximum angle

Maximum phase difference between the left and right channel.
 When the total phase difference between the two channels gets above this level, the Image phase amplifier is reduced temporarily to
keep it below the level that is set here.

Mono or stereo only

-100 %: Play ONLY the mono sounds.

0 %: Don't do anything
+100 %: Play ONLY the stereo sounds.
If an instrument is only present on one channel, -100% will completely remove it. If an instrument is present at the center, +100% will
completely remove it.
Note: Image width amplifier is performed first!
Use 'stereo only' with care: High values can cause annoying artifacts.

ACR Stereo section

ACR Stereo is to date the most effective stereo enhancer in Stereo Tool.
ACR Stereo adds a copy of the L-R (stereo, difference between channels) information to the original stereo signal. This creates a sparkling
stereo effect. In cases when the stereo effect in the input signal is already very strong, or almost not present, this filter switches off
automatically. On top of these things, the stereo effect is stronger for transients.
Several users have reported that the bass also sounds much better with this filter enabled.
This type of stereo widening has very little impact on reverb (which often gets boosted by other types of stereo widening). When used for FM,
it also has very little impact on multipath distortion (see also Multipath clipper).

ACR Stereo panel

ACR Stereo enabled

Enabled ACR Stereo.

L-R Level Boost

Maximum boost of the L-R signal that's added to the original stereo signal.

This is only the maximum, the effect can be reduced by Limit below, Stop all action if input is wider than and if there's almost no L-R
signal present, Or smaller than.

L-R Delay
Amount of delay of the L-R signal that is added to the original stereo signal.

When using more delay, the sound appears to 'open up', however it might start to sound like an echo, especially for high frequency
transients.

When using 0, there is no delay and the existing stereo information is boosted.

Limit below
Do not add more than this amount of L-R stereo.

This level is relative to the mono (L+R) audio level.

Stop all action if input is wider than

Turns the widener off completely if the input is already very wide.

If the input signal already contains a lot of stereo content, the ACR Stereo is turned off.

This is particularly important when playing audio with 100% channel separation such as old Beatles tracks where voices are often
recorded on one channel, and the instruments on the other. If a delayed (L-R Delay) version of the signal would be added, it would sound
like an echo on the complete signal. And if there's no delay, it would cause anti-phase signals to appear in the audio.

Or smaller than
Switches stereo enhancement off when the signal is nearly mono.

If the left and right channel levels are not perfectly equal, sounds such as a DJ speaking through a microphone would constantly be
moved off-center even more - and if a delay (L-R Delay) is used, an echo would be audible all the time. To protect against this, the filter
switches off is the signal is nearly mono.

Input, Extra Stereo & Limiter panel

Shows the input (top bar) and added (bottom bar) amount of stereo signal.
The bottom bar also shows if the effect is reduced (gray bar) by Limit below, Stop all action if input is wider than or Or smaller than.

Bandpass panel
Controls the frequency range on which the stereo widener works.
Widening deep bass sounds is useless and may actually reduce the reception quality. The human hearing cannot distinguish where bass
comes from (unless you're using headphones).
Widening high frequencies in combination with a delay may cause an echo-like effect. Some people like this, others don't.
 Highpass from
Frequencies below this frequency are not widened at all.

Highpass to
Frequencies above this frequency are widened by the maximum amount.

Lowpass from
Frequencies below this frequency are widened by the maximum amount.

Lowpass to
Frequencies above this frequency are widened by End level.

End level
The amount of stereo widening for frequencies above Lowpass to.

Punch detection panel

Controls if transients are widened more than constant sounds.
Widening transients more helps to protect against boosting reverb. It also tends to keeps voices in music more audible.

Volume effect
Controls how much effect punch (transient) detection has.
At 0%, everything is boosted maximally regardless of detection of punch. At 100%, the amount of widening for constant tones appoaches
none at all.
Setting this too high may reveal some unwanted effects and cause the total effect of the stereo widener to be less than desirable.

Fadeout
Controls how fast the volume drops down to Volume effect after a transient.

L+R Contrast
Increases effect of detected transients (punch).

Punch detection often doesn't reach 100%; this setting boosts the effect to reach maximum widening. The detected punch will never
exceed 100%.
For punch detection, the minimum of L+R and L-R Contrast is used.

L-R Contrast
Increases effect of detected transients (punch) in the stereo (L-R) signal.
See also L+R Contrast. To protect against unwanted effects (a mono-only sound which is not present in the L-R stereo signal but would
still affect the stereo separation), the minimum of L+R Contrast and L-R Contrast is used to detect punch.

Attack
Attack speed for punch detection.

Release
Release speed for punch detection.

Relative volume release

Controls the difference in release speeds between 2 compressors to determine punch.
Punch is detected by looking at the difference in amount of attack between two compressors. This slider determines how much slower
the release works for the 2nd compressor.

Response speeds panel

Lookahead time
Lookahead time for the total behavior.

Fake Stereo section

Adds a fake stereo effect. Intended to make mono recordings sound more like stereo.

Fake Stereo panel

Fake Stereo enabled

Enables the Fake Stereo effect.

Channel delay (left)

Delay between the left and right channel.

This introduces a (very rudimentary) stereo effect. Note that this also changes recordings that are already in stereo, that it also makes
sounds like voices stereo (which is generally considered bad), and that the result will sound very bad when played back in mono.
Equalizer section
Increase or reduce the presense of frequencies.
The freely drawable equalizer makes it very easy to make frequency ranges louder of softer.

Equalizer panel

Enabled
Enables the equalizer.

PEQ
Graph that controls the increment or reduction of the level per frequency.
True Bass section
Lost harmonics generator.
True Bass attempts to generate harmonics that either weren't recorded properly (due to for example a highpass filter, a bad microphone) or
appear to be missing.
It can generate bass at a specific frequency, which matches the rest of the signal. True Bass was designed to only generate bass that
sounds like it should always have been there.
Some examples: Input: Square wave, 50 Hz, with the base frequency (50 Hz) filtered out - only 150, 250, 350, 450 Hz etc. remain. True Bass
will recreate a 50 Hz tone in this situation:
Input:

Output:

However, if instead of a square wave, we feed it a sine wave at 150 Hz, it will [b]not[/b] generate a subharmonic:
Input:
Output:

True Bass panel

Enabled
Enables True Bass.

Place in chain panel

Determines where True Bass is located in the processing chain.

Before multibands
Sets the place in the processing chain where True Bass is performed.
For consistency, it's best to place it before the Multiband Compressor, or if you use 2 Multiband compressors, at least before Multiband
2.

The effect on the audio is bigger if you place it after the multiband compressors, but you risk getting too loud bass in some cases, and
the clipper can easily be overdriven if that happens.

Band 1 panel
First True Bass filter.

level
The amount of effect that True Bass has.

Using level 100% matches the 'natural' effect of the filter. Using more can give unnatural effects.

Controls panel
True Bass audio controls.

Peak frequency
The frequency around which this True Bass filter works.
The filter drops off pretty steeply to higher frequencies, but also to lower ones.

Look at frequencies upto

 Input frequencies to look at to generate harmonics.
If this frequency is set very high, asymmetrical sounds such as voices can cause rumbling effects. If it is set too low, if only one harmonic
is seen the filter will not do anything (see the images of a square wave and sine wave above), a too steep filter will make the filter "think"
that a square wave which' ground frequency is gone is actually a sine wave.

Upper slope
Steepness of the lowpass filters.

Drop steepness (relative to 300%%/oct)

Rolloff steepness of the effect of the filter below Peak frequency.

Pre-ringing filter level

Amount of pre-ringing allowed.
Since the subharmonic frequency is lower, it can sometimes start before the original bass sound started. This sounds very strange. This
setting determines how much pre-ringing is allowed. Setting it too low will also remove reconstructed bass harmonics if they are a lot
louder than the original.

Assume voices are above

Assume that voices start above this frequency.
Voices (and other asymmetrical sounds) can cause rumbling bass in the output. If there's very little content below this frequency, the
effect of True Bass is reduced. This avoids having weird bass harmonics during news broadcasts, for example.

Allowed voice rumble

Amount of voice rumble that is allowed through.

Band 2 panel
Second True Bass filter.
This can be used to boost bass at multiple frequencies.
Multiband Compressor section
Reduces the dynamic range of the audio and limits it, using a configurable number of bands.
Volume compression (A.K.A. audio level compression) reduces the dynamic range of a sound. This means that loud sounds become
softer, and soft sounds become louder.
Limiting limits the maximum audio level below a certain threshold.
For a lengthy discussion about compression, see Wikipedia: Audio level compression.
The compressors and limiters in Stereo Tool are protected against causing distortion. So very aggressive settings and large amounts of
limiting can safely be used.

Main panel
General compressor/limiter settings.

Enabled
Turns the compressor/limiter on.

Compressor type
Chooses between Analog and Digital compressor mode.

Analog mode was added later, and is the preferred mode. It sounds slightly better and it uses far less processing power.
For the first Multiband Compressor, you can choose whether you want to use Digital or Analog mode. Multiband 2 only supports Analog
mode.

To keep this manual readable, the manual for the first Multiband Compressor (this one) will only describe the Digital parameters. For
Analog mode, also for the first Multiband Compressor, see Multiband 2.

Quick adjust panel

Achieve several effects with a single slider.
Most of these sliders impact the value of several other sliders that are described below, to achieve a certain effect.

Density
Makes attack and decay faster or slower. Sound is more squashed.
This slider adjusts both attack and decay to have more aggressive compression.

Works on all bands simultaneously.

Aggressiveness (hot)
Makes attack and decay faster or slower. Sound is more squashed.
This slider adjusts both attack and decay to have more aggressive compression.

Works on all bands simultaneously.

Main levels panel

Controls at which input level the compressor and limiter become active.

Drive
Amplification of the input before the compressor/limiter.

Output level
Amplification of the output level after the compressor and limiter.
It is generally a good idea to make sure that the output level of each filter is set such that disabling the filter does not change the level.
This makes it much easier to compare what each filter does (it can be turned on and off without having to adjust other settings).

Attack Times multiplier

Value with which to multiply all Attack Time.

Release Times multiplier

Value with which to multiple all Release Time.
 Release hold Times multiplier
Value with which to multiple all Release hold time times.

Median display panel

Displays the median attenuation value of compressor bands.
If the median amount of attenuation of each band of the multiband compressor is roughly equal, then soft and loud sounds are handled
roughly the same.
On some radio stations, you can clearly hear that when a song ends and fades out, the frequency content shifts. This happens because at
some point some frequencies are still being compressed while others are not, or no frequencies are compressed anymore and the sound
is identical to the input.
That by itself is fine, but if the multiband compressor changes the sound of the input, not by using Band mix but by modifying compressor
settings such as Threshold level and Ratio, then this type of problems start to occur.
If you configure the multiband compressor such that on average, after playing a lot of differnt program content through it, all bands have the
same amount of compressor action, then this will not or hardly happen (only for songs that sound very different from the average).
The median display shows dotted bars in the multiband compressor attenuation meters, so you can run a few hours of programming
through it and then check if they look fine - and adjust some levels if needed.

Reset median calculation

Throws away all the historic median data and starts to measure anew.

Hide median display

Removes the median displays.

Also slightly reduces the CPU load.

Bands panel
Controls the number of multiband compressor bands.

Bands
The number of bands.

If you change the number of bands, all the Frequency and RMS block size sliders will get new default values. A popup will ask you if you
want to update the sliders to these new default values.

Flat frequency response panel

This slider helps to keep the frequency response of both sweeps and pink noise flat.
If you play a sweep through a multiband compressor, it happens frequently that the output is louder in some places than in other. Usually, it
is louder around the crossover frequency between bands, although this also depends on the amount of compression.
A good value for Flat Frequency Response can only be found by trial and error, the value that gives the flattest response on sweeps should
be used. It is generally also a good idea to test the response for pink noise; this slider has very little effect on pink noise but it should be flat
as well, except for intentional non-flatness.
Update: With properly setup band frequencies, this slider is not needed. The default frequencies in Stereo Tool have not yet been adjusted
for this. But they will be in the future, making this slider useless for most users.

Flat frequency response

The flatness value.
0 does nothing, 100% moves the measurement strength at crossover frequencies from -6 dB to 0 dB. See the thin lines in the Bands
display.

Flat band tops

Changes the shape of multiband processing bands.
This enables a different band splitting mode with flatter top areas of the different bands, and a different mechanism to keep the frequency
response flat.
The advantage of this is that bands have less impact on each other, which can be used to generate a more stable sound image.

See also Flat tops.

Compatibility mode (bad)

Compatibility option for older presets.

There was a bug in the implementation of Flat Frequency Response, older presets might depend on it. Don't use this for new presets!

Band coupling panel

Controls coupling between adjacent multiband bands.
To avoid very extreme effects from the multiband compressor if certain frequency ranges are nearly absent or very loud in the incoming
signal, the bands can be tied together to stop a single band to move very far away from the adjacent bands.

Band coupling
Coupling between all bands.
 This number defines how strongly bands are coupled if they are exactly one octave apart (Frequency doubles between bands). Bands
are coupled stronger if they are closer together and weaker if the distance between them is bigger.

Only pull down

Pull bands down if adjacent bands are lower, but never pull bands up.

This gives better level control.

Low Freedom
Ignore coupling for the lowest band.

The lowest band is somewhat special: If you don't allow it to move freely, absense of bass or presence of very strong bass cannot be
handled properly. On the other hand, if you want the output to stay true to the original, that's actually a good thing.

With this slider you can determine how much of band coupling is ignored for the lowest band. Note that since bands are still coupled in
both directions, the changed value of this lowest band will also have some effect on adjacent bands.

High Freedom
Ignore coupling for the highest band.

The highest band is somewhat special: If you don't allow it to move freely, absense of highs or presence of very strong highs cannot be
handled properly. On the other hand, if you want the output to stay true to the original, that's actually a good thing.

With this slider you can determine how much of band coupling is ignored for the highest band. Note that since bands are still coupled in
both directions, the changed value of this lowest band will also have some effect on adjacent bands.

Band linking panel

Controls band linking between adjacent multiband bands.
The idea is similar as Band coupling, but the type of coupling is very different. One of the things that can easily be done with Band Linking is
to ensure that the lowest band doesn't get attenuated less than band 2, which helps to avoid 'thunder bass' during voices, and to ensure that
the highest band doesn't get attenuated less than the band before that, which avoids too much of the highest highs, which can sound
unnatural.

Non-linear

Link 2->1

Link 3->2

Link N-2->N-1

Link N-1->N

Monitor
Plays only the output of this band.

Speeds section
Attack, release and ratio.
These are standard settings that almost every compressor has.

Speeds panel

Frequency
The center frequency of each band.

Ratio
Determines how strongly the compressor responds to changing input levels.
Say, at one moment a sound comes in at the threshold level, so nothing happens to it. If another sound comes in at 6 dB above the
theshold level, the input should be reduced by half. The ratio indicates how much of the increase in input level is not removed. At a the
lowest ratio (1:1), the compressor is basically disabled. At the maximum ratio, 1000:1, 1/1000th of the increase is kept.

Attack Time
The time a 86% volume reduction due to a higher input level takes.

If the input level increases a bit, the volume goes down more slowly than if it increases a lot. This means that it's not possible to give a
value in dB/ms.

Release Time
The time it takes for the output level to climb by 10 dB if the input level falls silent.

Release hold time

Time for the 'brake' on the release to fade out.

When attack has been active, release is not immediately activated to avoid excessive movement. Instead, the release is held back for a
while. This slider determines how long.
Levels section
Controls at which input level the compressor and limiter become active.

Levels panel

Threshold level
Amplification of the input before the compressor/limiter.

Knee
Makes the transition around the threshold more smooth.

At the threshold the response to slightly different input levels changes abruptly. Knee smooths the transition.

Gate slowdown
If the input level is lower than this, release is slowed down.

Limit
The maximum output level of the limiter.

Limit before compress

Protects the compressor against big spikes which are limited anyway.

If this is enabled, if a sound will be limited, the compressor will act as if the signal is limited before entering the compressor. As a result,
it will go down less fast on sudden loud sounds.

Limiter max release

Controls the release behavior of the limiter.

The limiter attack is always as fast as possible without causing distortion. The same is true for release, but in some cases the release
behavior can be too prudent. This slider overrides the standard limiter release behavior: If the release behavior that would be used based
on the adaptive algorithm is slower than this, the configured release time is used instead. This does mean that very fast release times
can cause some distortion.

Band mix
Output level of this band.
Use the Band Mix settings to increase or decrease the presence of frequency bands.

Behavior section
Settings that change the standard compressor behavior.

Behavior panel

Attack Flatness
Lets the compressor respond faster to small differences and slower to bigger ones.

Small differences in level are thus quickly compensated, with helps to reach the target level much faster. And the compressor attack
responds less aggressively to big volume changes.

Release Flatness
Lets the compressor respond faster to small differences and slower to large ones.
Small differences in level are quickly compensated, with helps to reach the target level much faster as long as differences in level are
small. This gives a much more sparkling, 'alive', sound. But... Big differences are less quickly compensated. See Release Inertia for a
solution for that.
Another explanation to further clarify things: In the compressors, if there's a volume change, it takes quite long for the level to 'stabilize'.
That's because the closer the actual level gets to the 'target' level, the slower it moves (the shape is asymptotic). Something similar
happens in release. This seems to be a good thing, and traditionally this is what compressors do.
What Flatness does is:
If the difference in level is 6 dB, nothing changes
If the difference in level is less than 6 dB, for Flatness values > 1 the change speed is increased.
If the difference in level is greater than 6 dB, for Flatness values > 1 the change speed is decreased.
More technical: The Flatness'th root of the difference in level is used - so for 2 that's the square root etc.
What this means: The higher the Flatness value is, the more the movement to the new level will look like a straight line instead of an
asymptote.

Release Inertia
Adjusts release behavior to match human hearing for more natural results.
Without Inertia and Release Flatness, after a very big volume spike the speed at which the audio returned was always the same - but
determined by how much it had to move up. So, if the volume dropped by 6 dB and after 100 ms the volume went up 3 dB, then for a
volume drop of 12 dB that would be 6 dB. Sounds perfect.
But it's not. Say you have a huge drop, for example after a very loud 'S' in the high frequency band, where normal volume differences are at
most a few dB and this S suddenly sticks out 20 dB. For a difference of 4 dB, after 100 ms the difference in level is 1 dB - 75% of the
difference is reduced. Now, this last 1 dB is really nearly unnoticeable, so for your ears the release kinda stops after 100 ms. But, for a
 difference of 20 dB, after 100 ms the difference is still 5 dB! And you need more than another 100 ms before you reach this 1 dB point.
So, after a loud sound you hear a gap at settings that sound good for small volume differences.
Release Flatness helps a lot for the final part of release: Small differences get compensated faster. But at the same time, bigger
differences take longer to recover, which causes the same effect for really big differences as before.
Inertia fixes this. With inertia combined with Release Flatness you can make the release happen in a nearly constant time, without the
slowdown at the end that you would have without Release Flatness, but also without the slower recovery for very big volume differences.
Basically, the release happens in a nearly straight line, but the slope of the release depends on how much level must be compensated.
With high Inertia values, release can even be faster for very big differences than for smaller ones, which can be good to quickly fill up the
gap after a loud sound.
For bigger Gamma values you need bigger Inertia values.
In case things are not yet clear now, here's another explanation: For release, especially large differences must be compensated very fast
- for 2 reasons:
Big differences mean very dynamic input, and for more dynamic input it's good that more compression occurs.
If you have a loud sound, and it takes multiple seconds for the level to get back, that sounds really bad.
Example:
Sound drops by 4 dB. When 3 dB has been restored, you really won't hear much difference anymore in level.
Sound drops by 40 dB. Now, when 39 dB has been restored you really don't hear much difference anymore.
So in one case when 75% restoration is there we're good, in the other we need 97.5%. And since - without Release Flatness - the
behavior is asymptotic, reaching 97.5% takes multiple times as long as reaching 75%. Higher Release Flatness values only make things
worse.
Why is this bad? Well, it makes it nearly impossible to find a good Release (time to raise 10 dB), what works well for small differences
will be far too slow for big differences, and what works well for big differences will sound very aggressive on small differences.
So, the time it takes for the level to be restored to a level where human hearing stops to notice a difference - say 1 dB below the target
level - must be nearly constant.
Inertia ('heavyness') makes sure that once release is moving up, the speed won't slow down until the target is reached. For big drops the
effect is much bigger than for small drops, which is exactly what is needed.
Release Inertia and Release Flatness must be configured to work properly together. The best way to do this is to record a sample with
different level tones (Loud - soft, loud - less soft, loud - just a little less loud), and check if all take approximate the same time to reach a
level slightly below the target level.
Analogy
If you have to drive 10 meters, you just barely hit the gass and drive very slowly.
If you have to drive 1 km, you hit the gass and speed up (Release hold time), then release the gass and let the car roll slowing down
towards the end.
With Inertia, you would not release the gas until you're very close to the end and then hit the brakes to stop.

Continuous Release
The size of the area around the current sample used to calculate the RMS level.

Bigger values means less precise timing of attack/release behavior, but also less effect from low frequencies (less pumping). Generally,
the RMS block size should be set just high enough to not cause distortion when using the limiters (Threshold level) a lot.

Burst protection section

Protection against spikes for slow compressor settings.
Some presets use very slow attack and release times. This can sound great, but the level control for sudden volume increases is less
good.
This section contains the settings for an extra compressor that takes over in such cases. It does not affect normal audio.

Burst protection panel

Loud burst protection

Enables loud burst protection.

Loud burst level difference

Increases the level of the 2nd compressor with faster attack.

Because the attack is so much faster, the audio level of the 2nd compressor is generally a bit lower. If we would take the minimum of the
two, we would always look at the 2nd compressor, but that should only happen in extreme cases. By increasing the output level and then
taking the maximum of the two, the 2nd compressor only has an effect on the sound if its output level is quite a bit lower. For example, if
this value is set to 2.00, the 2nd compressor will not kick in if the level difference is less than 6 dB.

Fast Attack
The fast attack time.
To be useful, this must be a lot smaller than Attack - typical values are around 1-5 ms.

Release speedup
Controls how much faster the 2nd compressor release is.

Beside a faster attack, the release for the 2nd compressor can also be made faster. This helps to prevent long-term volume drops after a
short loud spike in the sound. This value controls how much faster the release is than Release (time to raise 10 dB).
 Minimum drop
Disables the 2nd compressor if the attenuation didn't suddenly drop a lot.

The 2nd compressor should only be active if there's a huge difference between the volume when using a normal and very fast attack, but
that's not all - if you play very dynamic music it should not kill the punch. This slider controls how much the attenuation must have
suddenly dropped (in the fast attack 2nd compressor) for it to be taken into account.

Density
Adjusts both Drive and Band mix to have more compression but the same average output level.

Bands section
Frequencies and slope steepness of all the frequency bands.

Bands panel

Slope to {}
Steepness of the left slope of the band.

Less steepness generally gives a more natural, but sometimes harder to control sound.

Flat tops
The level at which the top of this band must be cut off.

If no compression/limiting occurs, or if all the bands are compressed/limited by the same amount, the end result is guarranteed to be flat
in frequency response.

Slope from {}
Slope of the right side of the band.
Less steepness in general gives a more natural, but harder to control sound.

Band coupling section

Band coupling matrix panel

Coupl

Auto

Norm

Detection section
Controls the detection of the input level.

Level detection panel

The level detection settings.

Channel separation
Process channels separately, combined, or in between.

At 0%, the two channels will always behave the same. At 100%, they move completely separate of each other.

Detection type
Chooses between RMS or Peak level measurement.

Peak mode can cause quite large reactions to a single small spike in the sound. RMS mode responds more like human hearing does,
but low frequencies seem to be counted a lot stronger than in peak mode, which easily causes pumping.

Feedback
Chooses between feed forward and feedback mode.

In feed forward mode, the input is used directly for the measurement. In feedback mode, the output level is measured instead of the input
level.

Feedback mode is known to sound more natural, but the level control is far less accurate. For example, say the input level is 6 dB too
loud and the ratio is 1:1000. Then in feed forward mode, the level will be reduced by about 6 dB. But in feedback mode, once the level is
reduced by about 3 dB, the compressor will 'see' that it needs about 3 dB of reduction and not reduce the level further.

Look-ahead time
Lets the compresor respond to the sound a bit in the future.
This means that the initial spike of a loud sound gets reduced better, which can give a more natural sound.
 The attack of the limiter is already protected, and if you don't use very short attack times for the compressor this probably has little effect.

Limiter distortion
Allows the limiter attack to distort.

Some people like this effect, especially on low frequency audio - bass kicks get a special type of 'edge'.

Non-standard tweaks section

Settings that control compressor/limiter envelope detection.
In a compressor we have 2 things: An 'envelope', basically a line that follows the audio level, and the compressor behavior itself. If the level
drops a lot, release is faster - and this is based on the envelope. Now, if the envelope just follows sample levels, then there will be a lot of
near-0 values (just when a waveform crosses 0) which would cause infinitely fast release behavior. The envelope line needs to be made
such that this doesn't happen.
So, around a peak in the waveform, for the surrounding samples we should not allow the envelope to reach much lower values than the
value of that peak.
That works fine for high frequencies. But if you take a bass, the sample values are dropping slowly and in the valleys the level will still
approach 0. Which still causes issues with release behavior. Because of that, there's some code that measures DC offset and increases
the Base smoothing to something close to infinity when there's more DC offset present. Base smoothing controls how big the area is that's
considered for DC measurement (lower frequency = bigger area). What we are actually measuring here is DC offset in a specific direction
divided by total (absolute) power.

Non-standard tweaks panel

RMS block size

The size of the area around the current sample used to calculate the RMS level.
Bigger values means less precise timing of attack/release behavior, but also less effect from low frequencies (less pumping). Generally,
the RMS block size should be set just high enough to not cause distortion when using the limiters (Threshold level) a lot.

Peak smoothing
Controls envelope smoothing around peaks in the waveform.
Lower values may cause distortion, but too high values reduce the precision of the limiters and (to a much lesser extent) the compressor
release behavior.

Bass detection
Controls upto which frequency bass should be detected for Smooth Bass power.

Smooth Bass power

If bass is present (Bass detection), increases the Base smoothing temporarily.

This slider controls how strongly the bass affects the release behavior. Setting it higher means more release slowdown when we see
bass.

If you use this for a single multiband band, then you need less of this because there are less other frequencies that hinder bass
detection.

Classic Multiband Compressor section

The deprecated Classic Multiband Compressor is only available for compatibility reasons.
Do not use this for new presets. Use the new Multiband Compressor instead.
Volume compression (A.K.A. audio level compression) reduces the dynamic range of a sound. This means that loud sounds become
softer, and soft sounds become louder.
Multiband volume compression means that the audio is separated into seperate frequency bands, and each frequency band is compressed
separately. This means that for example bass and hi-hats are treated separately, and a loud peak of one of the two does not affect the other.
This greatly reduces pumping, and it makes different recordings sound more equal. (If one recording has a loud bass and the next has a
loud hi-hat, after using the multiband compressor the loudness of the bass and the hi-hat is lowered.) See Wikipedi: Audio level
compression for a lengthy description.
The Multiband compressor in Stereo Tool splits the audio into 10 frequency bands, ranging from extremely low bass sounds (40 Hz) upto
very high highs (16 kHz). Then it compresses or limits each of those bands separately.

Limiting means that audio below a certain volume is left untouched. If the volume gets above this level (A in the graph below), the output
volume is lowered such that the resulting output stays at the set maximum volume A. Regardless how much the input level is increased, the
output level stays the same.
Compression affects all audio: Low volume sounds are amplified, high volume sounds are de-amplified. If the input volume is increased
further, it has less and less effect on the audio - but a bit of the increase is always maintained.
In general, when the goal is to make the volume as constant as possible, use limiting. If the goal is sound quality, maintaining the dynamics
of the recording, use compression.
For better compression results, read Achieving good compression.

Enabled
Enables the old Multiband compressor.

If Multiband Compressor is enabled, this old version is automatically disabled. This can be used to adjust the new Multiband Compressor
to sound similar to the Classic Multiband Compressor.

Equalizer panel
Configures the 10-bands equalizer.

Equalizer
Enables or disables the 10-bands equalizer.

Equalize after multiband

Switches between equalizing before performing multiband compression (hence boosting or reducing the compression of bands), or
afterwards (behaving like a normal equalizer).

This setting is only available for convenience. Disabling it gives the same result as enabling it and multiplying the Band {} soft limit
levels by the equalizer levels.

Clip bands
Enabled or disables the multiband clipper.

When the volume of a band gets higher than the value set in Band {} soft limit, clipping can be used to cut off the sound that is too loud.

Using clipping improves the sound quality because very short very loud spikes that are left over after compressing or limiting are
removed. This makes the sound far less "jumpy", and the output volume more constant.

Limit <--> Compress

Determines whether the audio is compressed or limited.

Limiting means that audio below a certain volume is left untouched. If the volume gets above this level, the output volume is lowered such
that the resulting output stays at the set maximum volume (Band {} soft limit).

Compression affects all audio: Low volume sounds are amplified, high volume sounds are de-amplified.

In general, when the goal is to make the volume as constant as possible, use limiting. If the goal is sound quality, maintaining dynamics,
use compression.

Flat frequency response

Tone comes out at roughtly the same level regardless of whether they are at the center or between frequency bands.
Without this mode, if a single tone is in between two bands it will come out about twice as loud as when it's in the center of a band. This is
most noticeable when playing a tone sweep, but normal music can also be affected by this. Especially bass frequencies can often not be
controlled accurately without Flat frequency response mode.
Flat frequency response mode enables Post filter (cleanup distortion). If that filter was not already enabled, the CPU load is increased.

Lock band 1+2

Don't allow the lowest two bands to move differently.

This avoids certain bass artifacts.

Steepness protection

Max steepness
Artifact filtering.

Lower differences between bands means less artifacts, but also that the multiband compressor is less effective.

Post filter (cleanup distortion)

Enables high quality mode.

This filters out compression artifacts. Doing so increases the CPU load, and it's only useful when very aggressive compression or limiting
is used (high Band {} upspeed or Band {} downspeed levels) - otherwise there are no compression artifacts to filter.

Link upspeeds
If enabled, when one up speed slider setting is changed, they all move to the same location.

Link downspeeds
If enabled, when one down speed slider setting is changed, they all move to the same location.

Bands

Bands panel
The settings for the Classic Multiband Compressor bands
Band {} frequency
The center frequency of each band.

Band {} equalizer
Sets the equalizer level for a frequency band.
The output volume of each band is multiplied by this value. This can be done before or after Multiband compression (see Equalize after
multiband).

Band {} soft limit

The target maximum output level for this frequency band.
For each band, how much of the output volume range may be used before the amplification must be turned down.
This is the value A in the graph above.

Band {} upspeed
Determines how fast the volume goes up when the output volume is below the Band {} soft limit.

For each band, when the output volume is lowered due to too loud sounds, this slider determines how fast the output volume can be
increased again. A higher value means that the amplification can increase faster. Too high values can cause a cracking sound.

Band {} downspeed
Determines how fast the volume for this frequency band is lowered when the output level gets above Band {} soft limit.
For each band, when a sound that is louder than the set maximum occurs, this value determines how fast the output volume is lowered.
Too high values make the sound very flat, too low values may cause the amplification to be lowered too slowly, causing (probably
unwanted) loud sounds when an instrument starts playing suddenly.

When this value is 1, the output volume never gets above the level set by Band {} soft limit. Lower values mean that sudden peaks can
cause the volume to be shortly higher. Clipping can be used to remove such peaks (see Clip bands and Band {} clip).

Band {} clip
Sets the clipping level for a frequency band.

When the volume of a band gets higher than the value set in Band {} soft limit, clipping can be used to cut off the sound that is too loud.
The value of each clipping slider determines at which volume clipping starts, for example when the clipping slider is set to 1.50, clipping
starts when the volume gets above 1.50 times the Band {} soft limit value.

Using clipping improves the sound quality because very short very loud spikes that are left over after compressing or limiting are
removed. This makes the sound far less "jumpy", and the output volume more constant. Clipping too much (at too low levels) however
makes the sound dull and lifeless.
Multiband 2 section
Second Multiband compressor.
The 2nd Multiband compressor can be used in many ways, for example to fine-tune the output of the first, or to add some density.
Because the first Multiband Compressor can be used in both Analog and Digital mode and this one only supports Analog mode, to keep
the descriptions clear, for the first Multiband Compressor the Digital mode parameters are described, and for this one the Analog
parameters are explained.

Main panel
General compressor/limiter settings.

Enabled
Turns the compressor/limiter on.

Quick adjust panel

Achieve several effects with a single slider.
Most of these sliders impact the value of several other sliders that are described below, to achieve a certain effect.

Density
Makes attack and decay faster or slower. Sound is more squashed.

This slider adjusts both attack and decay to have more aggressive compression.
Works on all bands simultaneously.

Aggressiveness (hot)
Makes attack and decay faster or slower. Sound is more squashed.
This slider adjusts both attack and decay to have more aggressive compression.

Works on all bands simultaneously.

Main levels panel

Controls at which input level the compressor and limiter become active.

Drive
Amplification of the input before the compressor/limiter.

Output level
Amplification of the output level after the compressor and limiter.

It is generally a good idea to make sure that the output level of each filter is set such that disabling the filter does not change the level.
This makes it much easier to compare what each filter does (it can be turned on and off without having to adjust other settings).

Attack Times multiplier

Value with which to multiply all Attack Time.

Release Times multiplier

Value with which to multiple all Release Time.

Median display panel

Displays the median attenuation value of compressor bands.
If the median amount of attenuation of each band of the multiband compressor is roughly equal, then soft and loud sounds are handled
roughly the same.
On some radio stations, you can clearly hear that when a song ends and fades out, the frequency content shifts. This happens because at
some point some frequencies are still being compressed while others are not, or no frequencies are compressed anymore and the sound
is identical to the input.
That by itself is fine, but if the multiband compressor changes the sound of the input, not by using Band mix but by modifying compressor
settings such as Threshold level and Ratio, then this type of problems start to occur.
If you configure the multiband compressor such that on average, after playing a lot of differnt program content through it, all bands have the
same amount of compressor action, then this will not or hardly happen (only for songs that sound very different from the average).
The median display shows dotted bars in the multiband compressor attenuation meters, so you can run a few hours of programming
through it and then check if they look fine - and adjust some levels if needed.

Reset median calculation

Throws away all the historic median data and starts to measure anew.

Hide median display

Removes the median displays.

Also slightly reduces the CPU load.

Bands panel
Controls the number of multiband compressor bands.

Bands
The number of bands.
If you change the number of bands, all the Frequency and RMS block size sliders will get new default values. A popup will ask you if you
want to update the sliders to these new default values.

Flat frequency response panel

This slider helps to keep the frequency response of both sweeps and pink noise flat.
If you play a sweep through a multiband compressor, it happens frequently that the output is louder in some places than in other. Usually, it
is louder around the crossover frequency between bands, although this also depends on the amount of compression.
A good value for Flat Frequency Response can only be found by trial and error, the value that gives the flattest response on sweeps should
be used. It is generally also a good idea to test the response for pink noise; this slider has very little effect on pink noise but it should be flat
as well, except for intentional non-flatness.
Update: With properly setup band frequencies, this slider is not needed. The default frequencies in Stereo Tool have not yet been adjusted
for this. But they will be in the future, making this slider useless for most users.

Flat frequency response

The flatness value.

0 does nothing, 100% moves the measurement strength at crossover frequencies from -6 dB to 0 dB. See the thin lines in the Bands
display.

Flat band tops

Changes the shape of multiband processing bands.

This enables a different band splitting mode with flatter top areas of the different bands, and a different mechanism to keep the frequency
response flat.

The advantage of this is that bands have less impact on each other, which can be used to generate a more stable sound image.

See also Flat tops.

Compatibility mode (bad)

Compatibility option for older presets.

There was a bug in the implementation of Flat Frequency Response, older presets might depend on it. Don't use this for new presets!

Wideband Gate panel

Gate that effects all bands.
Similar to Release Gate, but based on the total input, and it affects all bands simultaneously.

Band coupling panel

Controls coupling between adjacent multiband bands.
To avoid very extreme effects from the multiband compressor if certain frequency ranges are nearly absent or very loud in the incoming
signal, the bands can be tied together to stop a single band to move very far away from the adjacent bands.

Band coupling
Coupling between all bands.

This number defines how strongly bands are coupled if they are exactly one octave apart (Frequency doubles between bands). Bands
are coupled stronger if they are closer together and weaker if the distance between them is bigger.

Only pull down

Pull bands down if adjacent bands are lower, but never pull bands up.
This gives better level control.

Low Freedom
Ignore coupling for the lowest band.
 The lowest band is somewhat special: If you don't allow it to move freely, absense of bass or presence of very strong bass cannot be
handled properly. On the other hand, if you want the output to stay true to the original, that's actually a good thing.
With this slider you can determine how much of band coupling is ignored for the lowest band. Note that since bands are still coupled in
both directions, the changed value of this lowest band will also have some effect on adjacent bands.

High Freedom
Ignore coupling for the highest band.
The highest band is somewhat special: If you don't allow it to move freely, absense of highs or presence of very strong highs cannot be
handled properly. On the other hand, if you want the output to stay true to the original, that's actually a good thing.
With this slider you can determine how much of band coupling is ignored for the highest band. Note that since bands are still coupled in
both directions, the changed value of this lowest band will also have some effect on adjacent bands.

Band linking panel

Controls band linking between adjacent multiband bands.
The idea is similar as Band coupling, but the type of coupling is very different. One of the things that can easily be done with Band Linking is
to ensure that the lowest band doesn't get attenuated less than band 2, which helps to avoid 'thunder bass' during voices, and to ensure that
the highest band doesn't get attenuated less than the band before that, which avoids too much of the highest highs, which can sound
unnatural.

Non-linear

Link 2->1

Link 3->2

Link N-2->N-1

Link N-1->N

Monitor
Plays only the output of this band.

Attack section
Attack settings.

Attack panel

Frequency
The center frequency of each band.

Attack Time
The time a 86% volume reduction due to a higher input level takes.

If the input level increases a bit, the volume goes down more slowly than if it increases a lot. This means that it's not possible to give a
value in dB/ms.

Max Attack Speed

The maximum attack speed.
This is for the "automatic/hydraulic door" behavior: Push harder against it and at some point (this speed) it won't move faster anymore.

Auto Attack Shape

Automatically determines the optimal value for Attack Shape.

Attack Shape
Determines the attack behavior.
Attack and Release Shape must be balanced to get good results. Normally, they are filled in automatically, but you can still change them
to get a different behavior.
Standard behavior (input and output):
 Too low Attack Shape will cause the attack to start later and go down too deep:

Too high Attack Shape will cause the attack to jump down instantly, but using a more asymptotical behavior - if you increase the Attack
time to compensate, it will take very long before the target level is reached:

See also Release Shape.

Exceed Max Attack Speed

Allows exceeding Max Attack Speed if the input is very loud.

This is the automatic door behavior again: If you push very hard against a hydraulically controlled door, the oil in the hydraulic system
starts to heat up, and as a consequence, the oil becomes thinner and the door will start moving faster. This setting controls how much
faster the door can move if the pressure is high for a while (see also Exceed Attack Rise).

Exceed Attack Rise

Determines how fast Exceed Max Attack Speed responds to a very loud input level.

In the hydraulic door example, this determines how fast the oil heats up.

Release section
Release settings.

Release panel

Release Time
The time it takes for the output level to climb by 10 dB if the input level falls silent.

Max Release Speed

The maximum release speed.

Similar to Max Attack Speed, but for release.

Auto Release Shape

Automatically determines the optimal value for Release Shape.

Release Shape
Determines the release behavior.
Attack Shape and release shape must be balanced to get good results. Normally, they are filled in automatically, but you can still change
them to get a different behavior.
Standard behavior (input and output):

Too low Attack Shape will cause the attack to start later and go down too deep:

If you lower it even further the effect is even more extreme:

Too high Release Shape will cause the release have a more asymptotical behavior, it will take very long before the target level is
reached:

See also Attack Shape.

Exceed Max Release Speed

Allows exceeding Max Release Speed if the input level drops a lot.

Hydraulic door behavior: If you release the door it will close at a certain speed. In this case, if you were to put pressure on it to close
faster, the oil in the hydraulic system warms up and becomes more fluid, which makes the door close a bit faster. This control
determines how much faster the door can move if the pressure if high for a while (see also Exceed Release Rise).

Exceed Release Rise

Determines how fast Exceed Max Release Speed responds to a very low input level.

In the hydraulic door example, this determines how fast the oil heats up.

Exponential Release
Release faster if the input level drops more.
Release Gate section
Gating settings for release.
If the current output is much softer than the target level, this stops release from acting. This helps for example to avoid releasing when
there's only noise present, which would boost the noise. Also, for example during speech, if there's a moment of silence, this stops the
compressor from releasing completely, which would cause the attack to get a very big spike that it needs to handle.

Gate speed
Determines how fast the gate closes.

This value is relative to Release Time.

Gate freeze
Level below which all release action stops.
If the output audio is below this level, release stops completely. Be careful with this, if it is set too high, it can cause the compressor to not
come back after a big spike!
See also Gate slowdown.

Gate slowdown
If the input level is lower than this, release is slowed down.

Levels section
Levels panel

Density
Adjusts both Drive and Band mix to have more compression but the same average output level.

Threshold level
Amplification of the input before the compressor/limiter.

Knee
Makes the transition around the threshold more smooth.

At the threshold the response to slightly different input levels changes abruptly. Knee smooths the transition.

Band mix
Output level of this band.

Use the Band Mix settings to increase or decrease the presence of frequency bands.

Limiters section
Limiter settings.

Limiters panel

Limit
The maximum output level of the limiter.

Limit speed

Limit before compress

Protects the compressor against big spikes which are limited anyway.

If this is enabled, if a sound will be limited, the compressor will act as if the signal is limited before entering the compressor. As a result,
it will go down less fast on sudden loud sounds.

Sound section
Sound panel

Feedback
Chooses between feed forward and feedback mode.

In feed forward mode, the input is used directly for the measurement. In feedback mode, the output level is measured instead of the input
level.

Feedback mode is known to sound more natural, but the level control is far less accurate. For example, say the input level is 6 dB too
loud and the ratio is 1:1000. Then in feed forward mode, the level will be reduced by about 6 dB. But in feedback mode, once the level is
reduced by about 3 dB, the compressor will 'see' that it needs about 3 dB of reduction and not reduce the level further.
 Ratio
Determines how strongly the compressor responds to changing input levels.

Say, at one moment a sound comes in at the threshold level, so nothing happens to it. If another sound comes in at 6 dB above the
theshold level, the input should be reduced by half. The ratio indicates how much of the increase in input level is not removed. At a the
lowest ratio (1:1), the compressor is basically disabled. At the maximum ratio, 1000:1, 1/1000th of the increase is kept.

Channel separation
Process channels separately, combined, or in between.

At 0%, the two channels will always behave the same. At 100%, they move completely separate of each other.

Detection type
Chooses between RMS or Peak level measurement.

Peak mode can cause quite large reactions to a single small spike in the sound. RMS mode responds more like human hearing does,
but low frequencies seem to be counted a lot stronger than in peak mode, which easily causes pumping.

Peak mode

ITU-BS.1770

Lookahead
Lookahead time.
This determines how much the compressor looks ahead. For fast attack speeds (under 10 ms), this helps a lot to remove the short
spikes that remain after compressing.

Lowpass below
Lowpasses the compressor control signal.

By lowpassing the control signal from the compressor, distortion from fast compressor action is removed. It also helps smoothe out
remaining spikes when the Attack is a bit too slow, since the filtering works symetrically (not only forward in time).

Bands section
Controls the number of multiband compressor bands.

Bands panel
Controls the number of multiband compressor bands.

Slope to {}
Steepness of the left slope of the band.
Less steepness generally gives a more natural, but sometimes harder to control sound.

Flat tops
The level at which the top of this band must be cut off.
If no compression/limiting occurs, or if all the bands are compressed/limited by the same amount, the end result is guarranteed to be flat
in frequency response.

Slope from {}
Slope of the right side of the band.

Less steepness in general gives a more natural, but harder to control sound.

Spectral Balance section

Spectral Balance Compensation panel

Spectral Balance Compensation

Threshold

Slope

Power

Smooth slope from 100%%

Band coupling section

Band coupling matrix panel
Coupl

Auto

Norm
Bandpass section
Configures filters to remove very low (bass) or very high frequencies.

General panel
General Bandpass filter settings.

Enabled
Turns bandpass filtering on or off.

Highpass panel
Highpass (removes bass) settings.

Highpass frequency
Remove very deep bass sounds.
Controls the highpass frequency. Tones below this frequency are removed.

The behavior of the highpass filter is affected by the setting of Phase linear highpass filter.

Phase linear highpass filter

Determines whether highpass filtering is done phase linear (with artifacts), or non-phase linear (without artifacts).
Phase linear means that the shape of the bass is not changed (except that low frequencies are removed). This causes some artifacts
(chopper like sounds), especially at lower latency or quality settings.
Non-phase linear means that the shape of the bass is allowed to change. It will still sound the same, but the locations of peaks in the
signal will move.
Possible settings:
Automatic selection
Phase linear at latencies 2048 and 4096 samples, non-phase linear at latencies 512 and 1024 samples.
Always phase linear
Never phase linear
An additional advantage of non-phase linear filtering is that it moves different bass frequencies around in time. This is not or hardly
noticeable, but it does offer an advantage to the Advanced Clipper filter for short bass spikes: Because the bass is spread over time,
Loudness can boost it further, resulting in louder bass levels without affecting the rest of the sound.

Highpass filter order (non-phase linear)

The order of the highpass filter.
A higher value means steeper filtering. This can in some cases lead to unnatural sounding bass.

Post filter to avoid DC offset

Extra filter later in the chain to remove any newly added DC offset.

Reduce post filter artifacts

Filters the post filter output to remove some artifacts caused by filtering.

Lowpass panel
Settings that control removal of high frequencies.

Lowpass frequency
Remove high frequencies.
Controls the lowpass frequency. Tones above this frequency are removed.
This filter is very steep. The volume starts to drop a few hundred Hz below the configured frequency, and no frequencies above the set
frequency should be coming through. The lowpass filter is always phase linear.
Some special values are:
15000
This is the lowpass frequency that should officially be used for FM stations. Since the filter is very steep, slightly higher frequencies
should also work, and result in a better output quality.
4500 Hz
The lowpass frequency for AM stations in Europe.

Filter before processing (never below 16 kHz)

Filters out some high frequencies to protect processing against audio problems.

Some tracks from around 2010-2013 have a high pitched tone in them, which causes compressors (not just in Stereo Tool but in nearly
all audio processors) to go haywire. If this setting is enabled, such frequencies are filtered out before processing starts, which solves the
problem.
Bass Boost section
The bass boost filter deforms the bass sounds to make them sound louder, without causing very big volume spikes as a normal equalizer
would do.
If you have enough bass on newer music but not on older music, you should look at Power Bass instead.

General panel
General Bass Boost settings

Enabled
Enables the Bass Boost filter.

Strength panel
Controls the amount of bass boost.

Drive (higher -> lower output)

Strength
Determines the amount of bass boost that takes place.

0% is no change, 100% is maximum boost.

Frequencies panel
Determines the frequencies at which Bass Boost works, and how it sounds.

Maximum boost from 0 upto

Determines the frequency upto which the maximum amount of bass boost takes place.

Set this slider higher to have more bass boost - but the bass will sound distorted if it is set too high.

Then dropping to no boost at

Determines the frequency above which no bass boost occurs anymore.

Frequencies above this frequency are completely ignored in the bass boost filter.

Set this slider higher to have more bass boost - but the bass will sound distorted if it is set too high.

Allow harmonics from 0 upto

Bass boost artifacts below this frequency are not cleaned up at all.

Set this slider higher to have more bass boost - but the bass will sound distorted if it is set too high.

Then drop to no harmonics at

Bass boost artifacts above this frequency are completely removed.
Set this slider higher to have more bass boost - but the bass will sound distorted if it is set too high.

Behavior panel
Some other settings that control Bass Boost's behavior.

Relative to automatically detected peak level

Scales Maximum bass peak level according to the predicted bass level coming out of the AGC and Multiband filters.

If this is disabled, the maximum bass peak level does not change if any other settings are changed. If it is enabled, the filter will make a
prediction of the maximum bass level coming out of the AGC and Multiband filters, and assume that 100% is this maximum level (short
bass spikes in Multiband are ignored in this calculation).

Maximum bass peak level

Determines the maximum bass output level.

Preserve louder peaks if needed to preserve bass volume

Stops the bass level from being reduced if the input bass level is louder than the configured Maximum bass peak level.
If the bass that comes in is louder than the level set in Maximum bass peak level, the louder peaks are preserved if this is enabled.
Singleband Compressor section
Reduces the dynamic range of the audio and limits it.
Volume compression (A.K.A. audio level compression) reduces the dynamic range of a sound. This means that loud sounds become
softer, and soft sounds become louder.
Limiting limits the maximum audio level below a certain threshold.
For a lengthy discussion about compression, see Wikipedia: Audio level compression.
The compressors and limiters in Stereo Tool are protected against causing distortion. So very aggressive settings and large amounts of
limiting can safely be used.

General panel
General compressor/limiter settings.

Enabled
Turns the compressor/limiter on.

Compressor type
Analog or Digital compressor type.

The Analog compressor type is intended to replace the Digital one. It's behavior is generally more natural, so if you are starting on a new
preset, it's probably a good idea to use Analog mode. On top of the better end result, it also uses far less processing power.

Quick adjust panel

Achieve several effects with a single slider.
Most of these sliders impact the value of several other sliders that are described below, to achieve a certain effect.

Density
Use more compression.

Changes the input and output levels of the compressor such that the total output level is not affected but compressor starts to work at
much lower input levels.

Aggressiveness (hot)
Makes attack and decay faster or slower. Sound is more squashed.
This slider adjusts both attack and decay to have more aggressive compression.

Levels panel

Singleband Drive
Amplification of the input before the compressor/limiter.

Output Level
Amplification of the output level after the compressor and limiter.

It is generally a good idea to make sure that the output level of each filter is set such that disabling the filter does not change the level.
This makes it much easier to compare what each filter does (it can be turned on and off without having to adjust other settings).

Threshold level
The input level above which the compressor becomes active.

Knee
Makes the transition around the threshold more smooth.

At the threshold the response to slightly different input levels changes abruptly. Knee smooths the transition.

Attack panel
Attack settings.

Attack
The time a 86% volume reduction due to a higher input level takes.
 If the input level increases a bit, the volume goes down more slowly than if it increases a lot. This means that it's not possible to give a
value in dB/ms.

Max Attack Speed

Auto Attack Shape

Attack Shape

Exceed Max Attack Speed

Exceed Attack Rise

Attack Flatness
Lets the compressor respond faster to small differences and slower to bigger ones.

Small differences in level are thus quickly compensated, with helps to reach the target level much faster. And the compressor attack
responds less aggressively to big volume changes.

Limit panel

Limit level
The maximum output level of the limiter.

Limit speed

Limiter distortion
Allows the limiter attack to distort.

Some people like this effect, especially on low frequency audio - bass kicks get a special type of 'edge'.

Limiter max release

Controls the release behavior of the limiter.
The limiter attack is always as fast as possible without causing distortion. The same is true for release, but in some cases the release
behavior can be too prudent. This slider overrides the standard limiter release behavior: If the release behavior that would be used based
on the adaptive algorithm is slower than this, the configured release time is used instead. This does mean that very fast release times
can cause some distortion.

Limit before compress

Release panel

Release (time to raise 10 dB)

The time it takes for the output level to climb by 10 dB if the input level falls silent.

Max Release Speed

Auto Release Shape

Release Shape

Exceed Max Release Speed

Exceed Release Rise

Exponential Release

Release hold time

Time for the 'brake' on the release to fade out.
When attack has been active, release is not immediately activated to avoid excessive movement. Instead, the release is held back for a
while. This slider determines how long.

Release Flatness
Lets the compressor respond faster to small differences and slower to large ones.
Small differences in level are quickly compensated, with helps to reach the target level much faster as long as differences in level are
small. This gives a much more sparkling, 'alive', sound. But... Big differences are less quickly compensated. See Release Inertia for a
solution for that.
Another explanation to further clarify things: In the compressors, if there's a volume change, it takes quite long for the level to 'stabilize'.
That's because the closer the actual level gets to the 'target' level, the slower it moves (the shape is asymptotic). Something similar
happens in release. This seems to be a good thing, and traditionally this is what compressors do.
What Flatness does is:
If the difference in level is 6 dB, nothing changes
If the difference in level is less than 6 dB, for Flatness values > 1 the change speed is increased.
If the difference in level is greater than 6 dB, for Flatness values > 1 the change speed is decreased.
More technical: The Flatness'th root of the difference in level is used - so for 2 that's the square root etc.
What this means: The higher the Flatness value is, the more the movement to the new level will look like a straight line instead of an
asymptote.
 Release Inertia
Adjusts release behavior to match human hearing for more natural results.
Without Inertia and Release Flatness, after a very big volume spike the speed at which the audio returned was always the same - but
determined by how much it had to move up. So, if the volume dropped by 6 dB and after 100 ms the volume went up 3 dB, then for a
volume drop of 12 dB that would be 6 dB. Sounds perfect.
But it's not. Say you have a huge drop, for example after a very loud 'S' in the high frequency band, where normal volume differences are at
most a few dB and this S suddenly sticks out 20 dB. For a difference of 4 dB, after 100 ms the difference in level is 1 dB - 75% of the
difference is reduced. Now, this last 1 dB is really nearly unnoticeable, so for your ears the release kinda stops after 100 ms. But, for a
difference of 20 dB, after 100 ms the difference is still 5 dB! And you need more than another 100 ms before you reach this 1 dB point.
So, after a loud sound you hear a gap at settings that sound good for small volume differences.
Release Flatness helps a lot for the final part of release: Small differences get compensated faster. But at the same time, bigger
differences take longer to recover, which causes the same effect for really big differences as before.
Inertia fixes this. With inertia combined with Release Flatness you can make the release happen in a nearly constant time, without the
slowdown at the end that you would have without Release Flatness, but also without the slower recovery for very big volume differences.
Basically, the release happens in a nearly straight line, but the slope of the release depends on how much level must be compensated.
With high Inertia values, release can even be faster for very big differences than for smaller ones, which can be good to quickly fill up the
gap after a loud sound.
For bigger Gamma values you need bigger Inertia values.
In case things are not yet clear now, here's another explanation: For release, especially large differences must be compensated very fast
- for 2 reasons:
Big differences mean very dynamic input, and for more dynamic input it's good that more compression occurs.
If you have a loud sound, and it takes multiple seconds for the level to get back, that sounds really bad.
Example:
Sound drops by 4 dB. When 3 dB has been restored, you really won't hear much difference anymore in level.
Sound drops by 40 dB. Now, when 39 dB has been restored you really don't hear much difference anymore.
So in one case when 75% restoration is there we're good, in the other we need 97.5%. And since - without Release Flatness - the
behavior is asymptotic, reaching 97.5% takes multiple times as long as reaching 75%. Higher Release Flatness values only make things
worse.
Why is this bad? Well, it makes it nearly impossible to find a good Release (time to raise 10 dB), what works well for small differences
will be far too slow for big differences, and what works well for big differences will sound very aggressive on small differences.
So, the time it takes for the level to be restored to a level where human hearing stops to notice a difference - say 1 dB below the target
level - must be nearly constant.
Inertia ('heavyness') makes sure that once release is moving up, the speed won't slow down until the target is reached. For big drops the
effect is much bigger than for small drops, which is exactly what is needed.
Release Inertia and Release Flatness must be configured to work properly together. The best way to do this is to record a sample with
different level tones (Loud - soft, loud - less soft, loud - just a little less loud), and check if all take approximate the same time to reach a
level slightly below the target level.
Analogy
If you have to drive 10 meters, you just barely hit the gass and drive very slowly.
If you have to drive 1 km, you hit the gass and speed up (Release hold time), then release the gass and let the car roll slowing down
towards the end.
With Inertia, you would not release the gas until you're very close to the end and then hit the brakes to stop.

Continuous Release
Increases release speed if the level is reduced more.

This makes the release behavior non-linear.

Release Gate panel

Gate speed

Gate freeze

Gate slowdown
If the input level is lower than this, release is slowed down.

Detection panel

Detection type
Chooses between RMS or Peak level measurement.

Peak mode can cause quite large reactions to a single small spike in the sound. RMS mode responds more like human hearing does,
but low frequencies seem to be counted a lot stronger than in peak mode, which easily causes pumping.

Peak mode

ITU-BS.1770 Bass
Respond less strong to bass because to human ears it seem to sound less loud.

ITU-BS.1770 Head
 Respond more to high frequencies because they sound louder to humans.

Feedback
Chooses between feed forward and feedback mode.

In feed forward mode, the input is used directly for the measurement. In feedback mode, the output level is measured instead of the input
level.

Feedback mode is known to sound more natural, but the level control is far less accurate. For example, say the input level is 6 dB too
loud and the ratio is 1:1000. Then in feed forward mode, the level will be reduced by about 6 dB. But in feedback mode, once the level is
reduced by about 3 dB, the compressor will 'see' that it needs about 3 dB of reduction and not reduce the level further.

Ratio
Determines how strongly the compressor responds to changing input levels.

Say, at one moment a sound comes in at the threshold level, so nothing happens to it. If another sound comes in at 6 dB above the
theshold level, the input should be reduced by half. The ratio indicates how much of the increase in input level is not removed. At a the
lowest ratio (1:1), the compressor is basically disabled. At the maximum ratio, 1000:1, 1/1000th of the increase is kept.

Channel separation
Process channels separately, combined, or in between.

At 0%, the two channels will always behave the same. At 100%, they move completely separate of each other.

Lowpass below

Lookahead

Look-ahead time
Lets the compresor respond to the sound a bit in the future.

This means that the initial spike of a loud sound gets reduced better, which can give a more natural sound.

The attack of the limiter is already protected, and if you don't use very short attack times for the compressor this probably has little effect.

Non-standard attack panel

Protection against spikes for slow compressor settings.
Some presets use very slow attack and release times. This can sound great, but the level control for sudden volume increases is less
good.
This section contains the settings for an extra compressor that takes over in such cases. It does not affect normal audio.

Loud burst protection

Enables loud burst protection.

Level difference
Increases the level of the 2nd compressor with faster attack.
Because the attack is so much faster, the audio level of the 2nd compressor is generally a bit lower. If we would take the minimum of the
two, we would always look at the 2nd compressor, but that should only happen in extreme cases. By increasing the output level and then
taking the maximum of the two, the 2nd compressor only has an effect on the sound if its output level is quite a bit lower. For example, if
this value is set to 2.00, the 2nd compressor will not kick in if the level difference is less than 6 dB.

Minimum drop
Disables the 2nd compressor if the attenuation didn't suddenly drop a lot.
The 2nd compressor should only be active if there's a huge difference between the volume when using a normal and very fast attack, but
that's not all - if you play very dynamic music it should not kill the punch. This slider controls how much the attenuation must have
suddenly dropped (in the fast attack 2nd compressor) for it to be taken into account.

Fast Attack
The fast attack time.
To be useful, this must be a lot smaller than Attack - typical values are around 1-5 ms.

Release speedup
Controls how much faster the 2nd compressor release is.

Beside a faster attack, the release for the 2nd compressor can also be made faster. This helps to prevent long-term volume drops after a
short loud spike in the sound. This value controls how much faster the release is than Release (time to raise 10 dB).

Non-standard release panel

Experimental settings. Should probably not be used.

Dynamic release
Dynamically increase the release speed if the volume drops more.

If this is set to 0 the release always runs at exactly the same speed. A similar effect can be reached with Release Flatness.

Dynamic release to 0 dB
Release acts as if the input level is always at 0 dB.
 So the release speed depends only on how deep the level has dropped. See also Continuous Release.

Non-standard tweaks panel

Settings that control compressor/limiter envelope detection.
In a compressor we have 2 things: An 'envelope', basically a line that follows the audio level, and the compressor behavior itself. If the level
drops a lot, release is faster - and this is based on the envelope. Now, if the envelope just follows sample levels, then there will be a lot of
near-0 values (just when a waveform crosses 0) which would cause infinitely fast release behavior. The envelope line needs to be made
such that this doesn't happen.
So, around a peak in the waveform, for the surrounding samples we should not allow the envelope to reach much lower values than the
value of that peak.
That works fine for high frequencies. But if you take a bass, the sample values are dropping slowly and in the valleys the level will still
approach 0. Which still causes issues with release behavior. Because of that, there's some code that measures DC offset and increases
the Base smoothing to something close to infinity when there's more DC offset present. Base smoothing controls how big the area is that's
considered for DC measurement (lower frequency = bigger area). What we are actually measuring here is DC offset in a specific direction
divided by total (absolute) power.

RMS block size

The size of the area around the current sample used to calculate the RMS level.

Bigger values means less precise timing of attack/release behavior, but also less effect from low frequencies (less pumping). Generally,
the RMS block size should be set just high enough to not cause distortion when using the limiters (Threshold level) a lot.

Peak smoothing
Controls envelope smoothing around peaks in the waveform.
Lower values may cause distortion, but too high values reduce the precision of the limiters and (to a much lesser extent) the compressor
release behavior.

Bass detection
Controls upto which frequency bass should be detected for Smooth Bass power.

Smooth Bass power

If bass is present (Bass detection), increases the Base smoothing temporarily.
This slider controls how strongly the bass affects the release behavior. Setting it higher means more release slowdown when we see
bass.

If you use this for a single multiband band, then you need less of this because there are less other frequencies that hinder bass
detection.

Side chain section

Lets the compressor respond more or less to certain frequencies.
For example, you can boost the bass in the side chain to increase pumping caused by bass - or you can reduce it to reduce bass pumping.

Side chain panel

Classic Singleband section

The deprecated Classic Singleband Compressor is only available for compatibility reasons.
Limiting & Clipping section
The final clippers and limiters.
Stereo Tool has 2 clippers built in:
Simple Clipper
This clipper is indeed very simple. It does not handle distortion too well, so you should only use it to remove some occasional spikes.
Advanced Clipper
The Advanced Clipper uses a lot of techniques to minimize or even completely remove digital clipping and intermodulation distortion. It
can produce an extremely loud, yet still dynamic and distortion-free sound. The Advanced Clipper requires registration.
These clippers may still leave some very small spikes through. To remove those, Hard Limit output can be used.

Final peak remover panel

Removes small peaks above 0 dB in the audio.

Hard Limit output

Removes any small remaining peaks after the rest of the processing.
This is the final step of the processing chain. If there are any peaks left above 0 dB, this removes them by quickly lowering the volume
around the peak. This filter was built to avoid creating harmonics, not to sound good - so it should not be used to remove big spikes.

If the Post Amp slider is set close to 1.00x (0 dB), HARD LIMIT should always be enabled to protect against clipping. For FM output it
should also always be enabled, to protect against overshoots.

Prepare for lossy compression (MP3/AAC/OGG/...) panel

Protects the output against overshoots when using lossy compression (MP3 etc.) on it.
If you compress the tightly clipped output of Stereo Tool with a lossy codec such as MP3, AAC, Ogg, you will notice that you get spikes again.
This is caused by the fact that almost any change in the tightly clipped signal will cause it to be less tightly clipped. To avoid getting
distortion, you need to lower the output level.
But because differences in the signal occur much more in higher frequency ranges, it helps to boost the highs a bit before clipping, and
lowering them again afterwards.

This graph shows the effect of different settings on a 128 kbit/s MP3 file.
Lossy Compression affects Advanced Clipper and Hard Limit output.

Pre-emphasized clipping/limiting (MP3/AAC/OGG/...)

Turns pre-emphasized limiting and clipping on.
 Pre-emphasis time
Pre-emphasis.

The amount of pre-emphasis to use before, and de-emphasis after, clipping.

Protect lossy FM stream to transmitter panel

Extra FM pre/de-emphasis (bad!)

Simple Clipper section

DEPRECATED! Clips the audio below 0 dB with little protection against distortion.
This filter is good at removing short spikes, but causes distortion if it needs to clip too much.
If the Advanced Clipper is enabled, this filter is in bypass mode, except that it amplifies the audio with the value set in the slider.

Enabled
Turns the Simple Clipper on.

Adjust volume
Amplifies the input level by this amount.

Pre-limiter enabled (avoids distortion)

Distortion protection.

Quickly lowers the volume if the input level exceeds this level. Use with care, may cause distortion by itself if the cause of the high input level
is loud bass.

Pre-limiter volume
The level at which the Pre Limiter starts to work.

Advanced Clipper section

Clips the output, while removing distortion caused by clipping.
Boosts the output level without increasing peak level. The Advanced Clipper can make the sound up to 12 dB louder, without causing higher
peaks (hence clipping) in the output.
The clipper in Stereo Tool detects and actively removes (almost) any distortion that is audible by humans - but nothing else! This results in a
clean, open, dynamic and very bright sound.
This filter is intended for FM, AM or internet radio stations that want to sound loud. It can also be used to create a 'denser' sound.
Loudness leaves some small spikes in the signal. Therefore, if you use this filter, and the Post amplifier slider is set at (or just below) 0 dB,
make sure that Hard Limit output is enabled to avoid distortion caused by clipping. If you use Stereo Tool to prepare an FM signal, always
turn Hard Limit output on when using the Advanced Clipper, regardless of the Post Amp slider setting.

Volume boost panel

Volume settings that control the clipper behavior.

Advanced Clipper
Enables the Advanced Clipper.
When the Advanced Clipper is enabled, the Simple Clipper is automatically disabled, but the advanced clipper level is still adjusted for
the simple clipper's Amplification.

Clipper drive (Loudness)

Sets the amplification before going into the clipper.
Clipping always occurs at 0 dB, this slider adjusts the input level to match.

Relative non-FM loudness

Oversample limiters and clippers panel

Forces the clipper to use in 4 times oversampled mode.
This removes any spikes over 0 dB that are created in the sound card during playback. These spikes can cause distortion based on the
sound card design, but they don't have to.
This means that the clipping is stronger, hence may cause slightly more artifacts for high frequencies. For FM processing, this is
automatically enabled regardless of the setting.

Oversampled limiting and clipping

Turns oversampled limiting and clipping on.

Composite clipping (FM) panel

Composite clipping greatly improves the quality of loud FM transmissions.
The composite clipper can create upto 160% left and right channel modulation while keeping the total modulation at 100%. It also allows the
usage of Single Sideband stereo coding and several reception improvements.
If the audio is clipped normally, and stereo encoding and RDS encoding are added afterwards, the output that is generated will only rarely
peak to the maximum modulation. Multiple dB's of signal headroom are wasted by clipping the audio without looking at the stereo pilot, RDS
signal and how the L+R and L-R modulated at 38 kHz contribute to the peak level.
To give a simple example: If there's a high frequency peak in the audio, and both the stereo pilot and RDS signals are peaking in the
opposite direction, then the peak in the audio could be clipped at a much higher level. It is possible to 'hide' audio peaks inside the holes
created by the pilot and RDS signals. The effect of the 38 kHz modulation of the L-R signal is even bigger but also much harder to explain.
When using normal audio clipping and then adding the stereo and RDS signals, the left and right channel audio level after demodulation on
an FM receiver will always be a bit below 100%. FM composite limiter overdrive can get it closer to 100%, but it will never exceed 100%.
When using composite clipping, values over 160% can be reached. The audio level coming out of the receiver still won't exceed 100%
though, because only high frequencies can reach levels above 100%, and they are reduced with de-emphasis in the receiver.
The composite clipper can not be used if you are using an external stereo coder, and works slightly less good with an external RDS encoder.
For a description of all the settings, see Composite Clipping.

Pre-limiter section
Pre-Limiter and phase-optimizer panel

Monitor Output

Bands

Reference clipping level (0 dB to match clipper level)

Never limit bands below (relative)

Smothering (0 for normal, 1 for top half only)

Tones only (0 for any sound)

Tones threshold (100 for any sound)

After ABDP

Pre-Limiter effect panel

Bass section
Settings that affect the sound of bass.

Bass shape panel

Deforms the bass to achieve louder levels and sound better on some speakers.
Deforms low bass sounds if they are being clipped by making the tops 'flatter', to make them sound louder at the same peak level. This
adds harmonics, so if too much is used it can sound bad, especially with electronic music which contains very clean bass sounds. Another
effect is that bass is more audible on speakers that are bad at reproducing bass.

Bass shape: Strength

Controls how much deep bass sounds are made louder.

Bass shape: Max harmonic frequency

Controls upto which frequency the bass may be deformed to sound louder.

Bass sounds upto this frequency can be deformed to make them sound louder without getting higher peak levels. Setting this value
higher gives more bass, at the cost of the sound quality of the bass. Values upto about 160 Hz should be fine.

Bass shape: Badness (may cause knocking sound)

DC offset panel
Settings that control DC offsets in the clipper output.

Forcibly remove DC caused by Advanced Clipper (reduces quality)

Don't allow Loudness to add a DC offset.

In some cases, clipping only occurs in one direction (for example because there are spikes in one direction, but not in the opposite).
Normally, Loudness will move the DC offset level to reduce the needed amount of clipping. This may however cause a DC offset in the
output for a longer period of time, which is not always accurately reproduced by a sound card. So if you are feeding an FM transmitter with
your sound card, and it's not phase linear, it's usually a good idea to enable this setting to avoid getting too big spikes in your FM
modulation.
Bass protection (Deprecated) panel
Was needed in the past to protect against intermodulation distortion. Also boosts deep bass.

Highs section
Settings that protect the clipper against excessive highs.
Excessive highs can punch holes in the rest of the audio, which is generally perceived as very annoying in music (it's ok with speech). This
is usually only a problem for FM output, where the highs are boosted by Pre-emphasis and a lot of clipping is used to reach high output
levels.

Highs de-esser panel

Reduces excessive highs caused by Pre-emphasis.
It tries to do this with as little impact on the sound as possible. This filter only responds if the total level would be too high, so if there are no
other sounds present it will not reduce the highs a lot, regardless of how loud they are. This means that it normally has no impact on
speech, only on music. For FM output, the de-esser is far less needed when using Composite Clipping than when using traditional clipping.

De-esser limit
Sets the value above which the de-esser needs to work.

De-esser slope
Ignore de-esser for small drops.

If the highs level is only slightly too high, this slider reduces the effect of the de-esser, leaving a brighter sound. If the highs get really loud,
the de-essing behavior is not changed.

Small filter (latency 1024) panel

Extra filtering at latency 1024

Highs hole punch protection panel

Protect against volume drops caused by loud highs.
Protects the sound against volume drops caused by extremely loud (often pre-emphasized) highs. The only reason why this is configurable
is that it takes a lot of CPU power.

Leif's highs gap protection

Bass priority

Highs/Bass Threshold

Highs gap protection: Always

Always turns Highs hole punch protection on.

Highs gap protection: Always HQ

Always run Highs hole punch protection in high quality mode.

High quality mode uses twice as much CPU power as normal quality mode, but it's more effective at removing volume drops.

Highs gap protection: Preemphasized

Turns on Highs hole punch protection when Pre-emphasis or Prepare for lossy compression (MP3/AAC/OGG/...) is used.

Highs gap protection: Preemphasized HQ

Run Highs hole punch protection in high quality mode for pre-emphasized audio.

High quality mode uses twice as much CPU power as normal quality mode. For pre-emphasized audio it's wise to turn it on, because the
high frequencies can be extremely loud.

Smoothe highs early (potentially more distortion)

CPU section
Settings that control the CPU usage (vs. quality) of the Advanced Clipper.

CPU panel

Strictness (CPU)
Controls how strictly the clipper clips.
Lower values leave more small spikes in the output signal, which (if you don't want spikes) need to be removed by HARD LIMIT. And
HARD LIMIT removes the spikes by lowering the output level. Strictness is configurable because higher Strictness levels cost much more
CPU processing power.
 Take some shortcuts (CPU, reduces quality)
Some calculations are done less precisely. May slightly reduce audio quality.

Take oversampling shortcuts

Some processing steps are done at a lower than optimal sample rate.

Stronger clipping
Forces the clipper to more aggressively remove spikes.
This allows a lower Strictness (CPU) setting to reach the same level of remaining (small) spikes which can be cleaned up with Hard
Limit output. The audio quality may get slightly reduced though.

Medium stop (CPU)

Distortion protection section

Settings that protect the Advanced Clipper against intermodulation distortion.
Intermodulation distortion is the type of distorting that is caused by loud bass sounds, which create 'holes' in other frequency ranges.
(Intermodulation distortion is actually any kind of distortion created by modulation effects between different frequencies, but the type caused
by low bass frequencies is harder to clean up).

Protect panel
Intermodulation distortion protection settings.

Advanced Bass Distortion Protection

Enables bass intermodulation distortion protection.

Difference

Sloppy (louder, more open, more distortion)

Reduce IMD voice vibrations

Delay (punchy bass) panel

Delay clipping
Create a very punchy bass sound.

Delay bass clipping at the start of a new bass sound for this amount of time. The onset of the bass is not reduced, leading to a very
punchy bass sound. But also potentially to some distortion - although this type of very brief distortion is almost not noticeable.

Clipping level during delay

The clipping level used during the delay.

This level should be at or above Always clip deep bass below.

Relative sensitivity
Reduces the bass sensitivity effect during delayed clipping.

Despite what was said in Delay clipping, in some cases delayed clipping can cause noticeable distortion. Because of that, the clipping
level will still be reduced a bit if Reduce bass for mids or Reduce bass for highs indicate that it should be reduced. This slider controls
how strongly it responds during the onset of a bass.
If Always clip deep bass below is 90%, Dynamically reduce deep bass to is 70%, Clipping level during delay is 100%, then the
difference between the maximum (90%) and calculated (say 80%) level is used and multplied by this value. So say it's set to 0.5, then the
delayed clipping level that is used will be 100% - 0.5 * (90% - 80%) = 95%.

Response speed
Determines how quickly a new delayed clip can occur.

A punchy sound means that the volume goes up suddenly. To determine whether the volume goes up, we need to keep track of the
average level - this value tells the detection algorithm what the lowest frequency is that it can expect in the input, which helps to determine
how fast it should respond to lower sample values.

Bass panel

Reduce asymmetric bass

Clips bass harder if it's asymmetrical.

Asymmetrical bass (just like asymmetrical mids, which are handled by Phase rotation) are as bad for the sound as symmetrical bass
sounds at half the frequency - meaning that they can easily cause very noticeable intermodulation distortion. If this setting is checked, the
detection of asymmetrical bass causes the bass clipping level to be lowered, which protects the other sounds.

Smooth slope

Bass clipping strictness

Determines how strictly the bass is clipped.

Higher is better, but also causes a higher CPU load.

 Very low bass - Clipping from 0 Hz upto
Bass clipping works fully on bass frequencies below this one.

Very low bass - Drop to no clipping at

Bass clipping completely ignores frequencies above this one.

Always clip deep bass below

Whatever the input is, the bass may never exceed this level.

So even if there's only bass present, if this level is set low, it will not reach full modulation.

Dynamically reduce deep bass to

Reduce bass further if it causes problems.

If Reduce bass for mids or Reduce bass for highs indicates that the bass is damaging other sounds, by potentially causing
intermodulation distortion, the bass clipping level is reduced further, at most this level.
If this level is much lower than Always clip deep bass below, the effect on the bass level may become noticeable and annoying.

Mids IMD protection panel

Mids clipping

Reduce bass for mids

Reduces the bass clipping level if loud mid frequencies are present.

Bass combined with loud mid frequencies can easily cause intermodulation distortion. Setting this sensitivity higher causes the bass
clipping level to be dropped further. This also means that the bass level goes down.

Constant tone distortion protection: Smoothe mid frequencies

The number of bins around a spike that are protected against distortion.

Constant tones such as voices are especially susceptable to intermodulation distortion from bass sounds. Or actually, there is not more
intermodulation but it is far more noticeable.
This slider controls, when a frequency that sticks out is detected, how big the area around it is that needs to be protected against bass
intermodulation distortion.
Setting this higher gives more protection, but also slightly lowers the output level.

Constant tone distortion protection: Peak detection steepness

Controls the detection of single frequency peaks for Link error '3398'.

See Constant tone distortion protection: Smoothe mid frequencies. Increasing this value allows more peaks to be detected and
protected. But if too many peaks are protected, there will be a slight volume loss.

Highs protection panel

Highs clipping

Reduce bass for highs

Reduces the bass clipping level if loud high frequencies are present.

Bass combined with loud high frequencies can easily cause intermodulation distortion. Setting this sensitivity higher causes the bass
clipping level to be dropped further. This also means that the bass level goes down.

Allowed highs distortion

Determines how strongly the high frequencies are protected against bass.

Without protection, a combination of loud bass and loud highs leads to quite horrible intermodulation distortion from the bass in the
highs, as can be heard on my FM radio stations. Stereo Tool contains a unique filter that detects when this is noticeable, and fixes it.

This slider controls how strongly the protection works. Opposed to what you may expect, protecting it too much can by itself cause
something that sounds like intermodulation distortion, and the protection slightly lowers the level of the highs frequencies.

A value of a few percent seems to work best. This also depends a bit on the Allowed highs distortion clipping strictness slider.

Allowed highs distortion clipping strictness

How strictly highs intermodulation distortion is repaired.
Affects both the quality and the CPU load.

Highs priority
Selects between clean highs and constant volume levels.

High frequencies can get extremely loud due to FM Pre-emphasis. With this slider, you can choose if you want to keep the volume of all
other sounds constant (low values), or if you want loud highs to be able to push other frequencies down. That allows the highs to sound
brighter, but pumping caused by loud highs is very annoying in music.

If you are using Composite Clipping, higher values are less problematic, because highs are less problematic.

Upsampled highs clipping (useless)

Sound section
Settings that control the sound.

Open sound panel

By allowing a bit of extra distortion in some situations, the highs sound far more open.
The tight noise filtering can cause the highs to sound somewhat 'squeezed', while in many cases allowing a bit of extra noise in the high
frequencies can make the highs sound both louder and less 'restrained'/squeezed.

Airy Highs short

Allows extra highs distortion for loud short-term noise sounds.

Airy Highs long

Allows extra distortion for longer-term (~100 ms) loud high noise sounds.

Airy Highs tone sensitivity

Reduces Airy Highs short and Airy Highs long if the highs contain tones.

Time spread Short

Slightly changes the sound when Airy Highs short is used.

Time spread Medium

Slightly changes the sound when Airy Highs long is used.

Artistic effects panel

Add some distortion, always or to give more headroom to loud sounds.
Originally, this section was intended to add distortion as an 'artistic effect', but for that purpose Controlled Distortion was added later which
gives a much more precise control.
It can however also be used to allow a bit more distortion when it's most likely not noticeable, and the sound is really loud. This makes the
sound less 'squashed'.

Dirty Bass
Allows more distortion in bass frequencies, caused by bass frequencies.

Dirty Mids
Allows more distortion in the mid frequencies.

Dirty Highs
Allows more distortion in the high frequencies.

Protect tones
Reduces dirty mids and highs when tones are present.

Only for loud highs

Turns all the Dirty settings off if there are no loud highs present.

Sparkling Highs panel

Sparkling Highs

Strength

Soft focus

Punch (distorts tones)

Highs Brilliance

Controlled Distortion panel

Adds a controlled amount of distortion, to replicate the sound of certain processors.
Most compressors don't have a clipper that delivers a signal that's as clean and free of distortion as Stereo Tool. Some people (not many...)
seem to like this distortion. With Controlled Distortion you can replicate the distortion from those other processors, while still avoiding some
really nasty types of distortion.

Enable
Enables Controlled Distortion.

Base level
Extra distortion allowed for all frequencies.
It is likely (but still needs to be checked) that with a very small amount of allowed distortion, the clipper might sounds more open. And the
amount of extra distortion is so low that it's hardly noticeable.

IMD sensitivity
 Sensitivity for the introduction of intermodulation distortion.
With the added distortion, also certain really nasty types of distortion, such as intermodulation distortion in female voices combined with
bass sounds can occur. This slider controls the detection of this type of distortion.

IMD threshold
Controls the response to Limiter distortion.

Stokkemask FM section
ITU-R SM.1268 compliance settings.
The Stokkemask (ITU-R SM.1268) is a mask on the RF spectrum (see RF spectrum analyzer). Enabling this mask makes the RF spectrum
less wide, which improves multipath distortion and reception strength of your station, and reduces disturbances caused to weaker stations
at nearby frequencies.
Using Stokkemask is mandatory in the Netherlands, but it improves reception everywhere. For stations that don't need to be compliant but
do want the improved reception, Multipath clipper can be used - it has less impact on the audio and the CPU load but it should have the
same advantages.
The Stokkemask filter can be used without using Composite Clipping, but the sound quality will suffer a lot, and the effects of the RDS and
stereo pilot cannot be taken into account, which means that compliance cannot be completely guarranteed.

Stokkemask (same as under FM Transmitter) panel

Enables Stokkemask clipping mode.

ITU-R SM.1268 (Stokkemask, improves reception)

Enables Stokkemask clipping.

Force Stokkemask even if not using Composite Clipping (bad for audio)

Stokkemask output level (if no MPX output)

Tells the non-composite Stokkemask clipper how loud the audio will be broadcast.
Basically, this is used to tell the Stokkemask clipper how much the audio level is reduced (below 100% modulation) because stereo and
RDS are added.

Allow small overshoots (improves audio)

Don't stay within the mask, allow occasional spikes for better audio quality.

Stokkemask CPU usage panel

Settings that affect the Stokkemask clipping CPU usage and audio quality.

Skip smoothing (less strict)

Turns the smoothing off completely.

See Skip smoothing (less strict). The mentioned smoothing is turned off completely. The RF spectrum will show overshoots.

Sloppy smoothing (lower audio quality)

Skip some steps in the Stokkemask filtering.
When spikes outside the Stokkemask are detected in the RF spectrum, the signal around those spikes needs to be clipped more
strongly. Around means in both directions. If this checkbox is checked, this is only done in forward direction, because moving backwards
through the data (at 16x oversampling) is very heavy, and doing this in one direction has almost the same effect as doing it in both
directions.
You might want to keep this turned off if you really need to be 100% compliant.

Strictness (CPU)
Decreases the effect of the Stokkemask clipper by using more CPU power.
If this is set higher, multiple RF measurements are done to check if the Stokkemask clipper manages to get the signal inside the
Stokkemask - once it does, the amount of Stokkemask clipping can be reduced, which also reduces the effect of the Stokkemask clipper
on the audio quality and stereo separation.

Use area (CPU)