Dynamics Class Reference

Dynamics processor supporting compression, limiting, expansion, and gating. More...

#include <gin_dynamics.h>

Public Types
enum	Type {   compressor , limiter , expander , gate ,   compexp }
	Dynamics processing types. More...

Public Member Functions
	Dynamics ()=default
	~Dynamics ()=default
void	setSampleRate (double sampleRate)
void	setNumChannels (int ch)
void	setMode (Type t)
void	setLinked (bool l)
void	setParams (float attackS, float holdS, float releaseS, float thresh, float ratio, float kneeWidth)
void	setInputGain (float g)
void	setOutputGain (float g)
void	setDetectorMode (EnvelopeDetector::Mode m)
	How the detector measures level: peak, mean square or RMS.
void	setRMSWindow (float seconds)
	Averaging window for the ms and rms detector modes, in seconds.
void	setAnalogTC (bool analog)
	Analog rather than digital time constants.
void	setLookahead (float seconds)
	Delays the audio so the detector sees each transient before the gain does, in seconds.
int	getLatencySamples () const
	Lookahead delay in samples, for AudioProcessor::setLatencySamples().
void	setSidechainFilter (float highpassHz, float lowpassHz)
	Filters the detector's view of the signal without touching what is heard.
void	setDetectorPeak (float freqHz, float q, float gainDb)
	A peak in the detector path, again heard only through what it makes the gain do.
void	setAutoMakeupGain (bool enabled)
	Enables automatic makeup gain for compressor/limiter modes.
void	reset ()
void	process (juce::AudioSampleBuffer &buffer, juce::AudioSampleBuffer *envelopeOut=nullptr)
	Processes a buffer, with the detector following that same buffer.
void	process (juce::AudioSampleBuffer &buffer, const juce::AudioSampleBuffer &sidechain, juce::AudioSampleBuffer *envelopeOut=nullptr)
	Processes a buffer, with the detector following a separate side chain.
const LevelTracker &	getInputTracker ()
const LevelTracker &	getOutputTracker ()
const LevelTracker &	getReductionTracker ()
float	calcCurve (float detectorValue)

Detailed Description

Dynamics processor supporting compression, limiting, expansion, and gating.

Dynamics provides a versatile dynamics processing engine that can function as a compressor, limiter, expander, or gate. Based on algorithms from "Designing Audio Effect Plug-ins in C++" by Will Pirkle, it features attack/hold/release envelope detection, adjustable ratio and threshold, soft knee, and optional channel linking for stereo operation.

The compexp mode combines downward compression (above threshold) with upward compression (below the inverted threshold), using the same ratio for both.

Key Features:

• Five processing types: compressor, limiter, expander, gate, compexp
• Configurable attack, hold, release times
• Adjustable threshold and ratio
• Soft knee for smooth compression
• Independent input/output gain
• Stereo linking (process based on max of both channels)
• Side chain input for ducking, de-essing and source keyed gating
• Optional envelope output for visualization
• Level tracking for input/output metering

Usage:

Dynamics comp;
comp.setSampleRate(44100.0);
comp.setNumChannels(2);
comp.setMode(Dynamics::compressor);
comp.setLinked(true); // Link stereo channels
 
// Set compression parameters
comp.setParams(0.01f,   // 10ms attack
               0.0f,     // no hold
               0.1f,     // 100ms release
               -20.0f,   // -20dB threshold
               4.0f,     // 4:1 ratio
               6.0f);    // 6dB soft knee
 
comp.setInputGain(0.0f);   // 0dB input
comp.setOutputGain(3.0f);  // +3dB makeup gain
 
comp.process(audioBuffer);
 
// keyed off something else: duck the music under the voiceover
comp.process(musicBuffer, voiceBuffer);
 
// Get metering
float inputLevel = comp.getInputTracker().getLevel();
float outputLevel = comp.getOutputTracker().getLevel();

See also

EnvelopeDetector, LevelTracker

Member Enumeration Documentation

Type

enum Dynamics::Type

Dynamics processing types.

Determines the type of dynamics processing applied:

• compressor: Reduces gain above threshold (ratio > 1:1)
• limiter: Hard limiting (very high ratio, fast attack)
• expander: Increases dynamic range (reduces gain below threshold)
• gate: Attenuates signals below threshold (extreme expansion)
• compexp: Combined compressor + expander (expander threshold is inverse of compressor threshold)

Enumerator
compressor	Compression (reduce loud signals)
limiter	Limiting (prevent peaks above threshold)
expander	Expansion (increase dynamic range)
gate	Gate (cut quiet signals)
compexp	Combined compressor and expander (expander uses inverse threshold)

Constructor & Destructor Documentation

Dynamics()

Dynamics::Dynamics ( )

default

~Dynamics()

Dynamics::~Dynamics ( )

default

Member Function Documentation

setSampleRate()

void Dynamics::setSampleRate

(

double

sampleRate

)

setNumChannels()

void Dynamics::setNumChannels

(

int

ch

)

setMode()

void Dynamics::setMode

(

Type

t

)

setLinked()

void Dynamics::setLinked

(

bool

l

)

setParams()

void Dynamics::setParams	(	float	attackS,
		float	holdS,
		float	releaseS,
		float	thresh,
		float	ratio,
		float	kneeWidth
	)

setInputGain()

void Dynamics::setInputGain

(

float

g

)

setOutputGain()

void Dynamics::setOutputGain

(

float

g

)

setDetectorMode()

void Dynamics::setDetectorMode

(

EnvelopeDetector::Mode

m

)

How the detector measures level: peak, mean square or RMS.

See also

EnvelopeDetector::Mode, setRMSWindow

setRMSWindow()

void Dynamics::setRMSWindow

(

float

seconds

)

Averaging window for the ms and rms detector modes, in seconds.

See also

EnvelopeDetector::setRMSWindow

setAnalogTC()

void Dynamics::setAnalogTC

(

bool

analog

)

Analog rather than digital time constants.

Digital constants reach 99% of the target in the time you asked for, analog ones 63% - the same curve a real capacitor charging through a resistor follows. Analog is slower to arrive and sounds gentler on the attack; digital does what the number says.

setLookahead()

void Dynamics::setLookahead

(

float

seconds

)

Delays the audio so the detector sees each transient before the gain does, in seconds.

Without it, an attack fast enough to catch a peak has already let the front of that peak through.

Costs latency, which the host has to be told about - see getLatencySamples(). 5 ms is plenty for most material.

getLatencySamples()

int Dynamics::getLatencySamples

(

)

const

Lookahead delay in samples, for AudioProcessor::setLatencySamples().

setSidechainFilter()

void Dynamics::setSidechainFilter	(	float	highpassHz,
		float	lowpassHz
	)

Filters the detector's view of the signal without touching what is heard.

A high pass stops kick drums pumping the whole mix; a low pass stops cymbals doing the same.

Parameters

highpassHz	detector high pass, 0 for none
lowpassHz	detector low pass, 0 for none

setDetectorPeak()

void Dynamics::setDetectorPeak	(	float	freqHz,
		float	q,
		float	gainDb
	)

A peak in the detector path, again heard only through what it makes the gain do.

Boost around 6 kHz and the compressor turns into a de-esser; boost the low mids and it leans on boxiness.

Parameters

freqHz	centre frequency
q	width
gainDb	boost or cut, 0 disables the filter

setAutoMakeupGain()

void Dynamics::setAutoMakeupGain

(

bool

enabled

)

Enables automatic makeup gain for compressor/limiter modes.

When enabled, gain is automatically applied to compensate for the gain reduction, bringing peaks back to 0dB.

Parameters

enabled

True to enable auto makeup gain

reset()

void Dynamics::reset

(

)

process() [1/2]

void Dynamics::process	(	juce::AudioSampleBuffer &	buffer,
		juce::AudioSampleBuffer *	envelopeOut = nullptr
	)

Processes a buffer, with the detector following that same buffer.

Parameters

buffer	audio to process, in place
envelopeOut	optional detector envelope, as a linear level. Written to channel 0 when linked and to every channel otherwise, so it needs as many channels as the processor has.

process() [2/2]

void Dynamics::process	(	juce::AudioSampleBuffer &	buffer,
		const juce::AudioSampleBuffer &	sidechain,
		juce::AudioSampleBuffer *	envelopeOut = nullptr
	)

Processes a buffer, with the detector following a separate side chain.

The gain is worked out from sidechain and applied to buffer: ducking, de-essing, and gating something on a source other than itself. A gated reverb is the case that shows why it matters - keyed off its own output, a gate can only chase an envelope the reverb has already smeared, so it opens softly instead of on the transient and its timing shifts whenever the decay does.

Parameters

buffer	audio to process, in place
sidechain	what the detector follows. Needs at least as many samples as buffer; fewer channels is fine, the last one is reused, so a mono key drives them all. Input gain scales what the detector sees here just as it does without a side chain.
envelopeOut	optional detector envelope, as above

getInputTracker()

const LevelTracker & Dynamics::getInputTracker

(

)

getOutputTracker()

const LevelTracker & Dynamics::getOutputTracker

(

)

getReductionTracker()

const LevelTracker & Dynamics::getReductionTracker

(

)

calcCurve()

float Dynamics::calcCurve

(

float

detectorValue

)

---

The documentation for this class was generated from the following file:

• gin_dynamics.h