Initial implementation of sample-exact models and controllers

Also featuring a very efficient buffer-based system for transporting sample-exact control data
Also interpolation for automations
The native Amplifier is a reference implementation for taking advantage of sample-exact data and is currently
the only one that does so, it can be used to test things out, and as documentation/example for implementing the
same elsewhere

plugins/Amplifier/Amplifier.cpp

@@ -74,29 +74,62 @@ bool AmplifierEffect::processAudioBuffer( sampleFrame* buf, const fpp_t frames )
 	double outSum = 0.0;
 	const float d = dryLevel();
 	const float w = wetLevel();
+	
+	ValueBuffer * volBuf = m_ampControls.m_volumeModel.hasSampleExactData() 
+		? m_ampControls.m_volumeModel.valueBuffer() 
+		: NULL;
+		
+	ValueBuffer * panBuf = m_ampControls.m_panModel.hasSampleExactData()
+		? m_ampControls.m_panModel.valueBuffer() 
+		: NULL;
+		
+	ValueBuffer * leftBuf = m_ampControls.m_leftModel.hasSampleExactData()
+		? m_ampControls.m_leftModel.valueBuffer() 
+		: NULL;
+
+	ValueBuffer * rightBuf = m_ampControls.m_rightModel.hasSampleExactData()
+		? m_ampControls.m_rightModel.valueBuffer() 
+		: NULL;
+
 	for( fpp_t f = 0; f < frames; ++f )
 	{
 //		qDebug( "offset %d, value %f", f, m_ampControls.m_volumeModel.value( f ) );
 		
 		sample_t s[2] = { buf[f][0], buf[f][1] };
 
-		// convert vol/pan values to left/right values
-		const float left1 = m_ampControls.m_volumeModel.value( f ) *
-			( m_ampControls.m_panModel.value( f ) <= 0
-			? 1.0
-			: 1.0 - m_ampControls.m_panModel.value( f ) / 100.0 );
-		const float right1 = m_ampControls.m_volumeModel.value( f ) *
-			( m_ampControls.m_panModel.value( f ) >= 0
-			? 1.0
-			: 1.0 + m_ampControls.m_panModel.value( f ) / 100.0 );
+		// vol knob
+		if( volBuf )
+		{
+			s[0] *= volBuf->values()[ f ] * 0.01f;
+			s[1] *= volBuf->values()[ f ] * 0.01f;
+		}
+		else
+		{
+			s[0] *= m_ampControls.m_volumeModel.value() * 0.01f;
+			s[1] *= m_ampControls.m_volumeModel.value() * 0.01f;
+		}
 
-		// first stage amplification
-		s[0] *= ( left1 / 100.0 );
-		s[1] *= ( right1 / 100.0 );
+		// convert pan values to left/right values
+		const float pan = panBuf 
+			? panBuf->values()[ f ] 
+			: m_ampControls.m_panModel.value();
+		const float left1 = pan <= 0
+			? 1.0
+			: 1.0 - m_ampControls.m_panModel.value( f ) * 0.01f;
+		const float right1 = pan >= 0
+			? 1.0
+			: 1.0 + m_ampControls.m_panModel.value( ) * 0.01f;
 
 		// second stage amplification
-		s[0] *= ( m_ampControls.m_leftModel.value( f ) / 100.0 );
-		s[1] *= ( m_ampControls.m_rightModel.value( f ) / 100.0 );
+		const float left2 = leftBuf
+			? leftBuf->values()[ f ] 
+			: m_ampControls.m_leftModel.value();
+		const float right2 = rightBuf
+			? rightBuf->values()[ f ] 
+			: m_ampControls.m_rightModel.value();
+			
+		s[0] *= left1 * left2 * 0.01;
+		s[1] *= right1 * right2 * 0.01;
 
 		buf[f][0] = d * buf[f][0] + w * s[0];
 		buf[f][1] = d * buf[f][1] + w * s[1];

plugins/Amplifier/Amplifier.h

@@ -29,7 +29,7 @@
 
 #include "Effect.h"
 #include "AmplifierControls.h"
-
+#include "ValueBuffer.h"
 
 class AmplifierEffect : public Effect
 {

plugins/peak_controller_effect/peak_controller_effect.cpp

@@ -63,6 +63,7 @@ PeakControllerEffect::PeakControllerEffect(
 	m_effectId( rand() ),
 	m_peakControls( this ),
 	m_lastSample( 0 ),
+	m_previousSample( 0 ),
 	m_lastRMS( -1 ),
 	m_lastRMSavail(false),
 	m_autoController( NULL )
@@ -89,11 +90,15 @@ namespace helpers
 {
 
 	//! returns 1.0f if val > 0.0f, -1.0 else
-	inline float sign(float val) { return -1.0f + 2.0f * (val > 0.0f); }
+	inline float sign( float val ) 
+	{ 
+		return -1.0f + 2.0f * ( val > 0.0f ); 
+	}
 
 	//! if val >= 0.0f, returns sqrtf(val), else: -sqrtf(-val)
-	inline float sqrt_neg(float val) {
-		return sqrtf(fabs(val)) * helpers::sign(val);
+	inline float sqrt_neg( float val ) 
+	{
+		return sqrtf( fabs( val ) ) * helpers::sign( val );
 	}
 
 }
@@ -113,7 +118,7 @@ bool PeakControllerEffect::processAudioBuffer( sampleFrame * _buf,
 	// RMS:
 	double sum = 0;
 
-	if(c.m_absModel.value())
+	if( c.m_absModel.value() )
 	{
 		for( int i = 0; i < _frames; ++i )
 		{
@@ -127,8 +132,8 @@ bool PeakControllerEffect::processAudioBuffer( sampleFrame * _buf,
 		{
 			// the value is absolute because of squaring,
 			// so we need to correct it
-			sum +=	_buf[i][0]*_buf[i][0]*helpers::sign(_buf[i][0])
-			 + _buf[i][1]*_buf[i][1]*helpers::sign(_buf[i][1]);
+			sum += _buf[i][0] * _buf[i][0] * helpers::sign( _buf[i][0] )
+				+ _buf[i][1] * _buf[i][1] * helpers::sign( _buf[i][1] );
 		}
 	}
 
@@ -157,6 +162,7 @@ bool PeakControllerEffect::processAudioBuffer( sampleFrame * _buf,
 	curRMS = (1-a)*curRMS + a*m_lastRMS;
 
 	const float amount = c.m_amountModel.value() * c.m_amountMultModel.value();
+	m_previousSample = m_lastSample;
 	m_lastSample = c.m_baseModel.value() + amount*curRMS;
 	m_lastRMS = curRMS;
 

plugins/peak_controller_effect/peak_controller_effect.h

@@ -23,8 +23,8 @@
  */
 
 
-#ifndef _PEAK_CONTROLLER_EFFECT_H
-#define _PEAK_CONTROLLER_EFFECT_H
+#ifndef PEAK_CONTROLLER_EFFECT_H
+#define PEAK_CONTROLLER_EFFECT_H
 
 #include "Effect.h"
 #include "peak_controller_effect_controls.h"
@@ -48,6 +48,11 @@ public:
 		return m_lastSample;
 	}
 
+	float previousSample()
+	{
+		return m_previousSample;
+	}
+
 	PeakController * controller()
 	{
 		return m_autoController;
@@ -59,6 +64,7 @@ private:
 	PeakControllerEffectControls m_peakControls;
 
 	float m_lastSample;
+	float m_previousSample;
 	float m_lastRMS;
 	bool m_lastRMSavail;