/* ========================================
 *  Infinity2 - Infinity2.h
 *  Copyright (c) 2016 airwindows, Airwindows uses the MIT license
 * ======================================== */

#ifndef __Infinity2_H
#include "Infinity2.h"
#endif

void Infinity2::processReplacing(float **inputs, float **outputs, VstInt32 sampleFrames) 
{
    float* in1  =  inputs[0];
    float* in2  =  inputs[1];
    float* out1 = outputs[0];
    float* out2 = outputs[1];
	
	biquadC[0] = biquadB[0] = biquadA[0] = ((pow(A,2)*9900.0)+100.0) / getSampleRate();
	biquadA[1] = 0.618033988749894848204586;
	biquadB[1] = (A*0.5)+0.118033988749894848204586;
    biquadC[1] = 0.5;
	
	double K = tan(M_PI * biquadA[0]); //lowpass
	double norm = 1.0 / (1.0 + K / biquadA[1] + K * K);
	biquadA[2] = K * K * norm;
	biquadA[3] = 2.0 * biquadA[2];
	biquadA[4] = biquadA[2];
	biquadA[5] = 2.0 * (K * K - 1.0) * norm;
	biquadA[6] = (1.0 - K / biquadA[1] + K * K) * norm;
	
	K = tan(M_PI * biquadA[0]);
	norm = 1.0 / (1.0 + K / biquadB[1] + K * K);
	biquadB[2] = K * K * norm;
	biquadB[3] = 2.0 * biquadB[2];
	biquadB[4] = biquadB[2];
	biquadB[5] = 2.0 * (K * K - 1.0) * norm;
	biquadB[6] = (1.0 - K / biquadB[1] + K * K) * norm;
	
	K = tan(M_PI * biquadC[0]);
	norm = 1.0 / (1.0 + K / biquadC[1] + K * K);
	biquadC[2] = K * K * norm;
	biquadC[3] = 2.0 * biquadC[2];
	biquadC[4] = biquadC[2];
	biquadC[5] = 2.0 * (K * K - 1.0) * norm;
	biquadC[6] = (1.0 - K / biquadC[1] + K * K) * norm;
	
	double size = (pow(B,2)*99.0)+1.0;
	double damping = pow(C,2)*0.5;
	double rawPass = D;
	double feedback = 1.0-(pow(1.0-E,4));
	double wet = F;
		
	delayA = 79*size;
	delayB = 73*size;
	delayC = 71*size;
	delayD = 67*size;
	delayE = 61*size;
	delayF = 59*size;
	delayG = 53*size;
	delayH = 47*size;
	
	delayI = 43*size;
	delayJ = 41*size;
	delayK = 37*size;
	delayL = 31*size;
	
    while (--sampleFrames >= 0)
    {
		double inputSampleL = *in1;
		double inputSampleR = *in2;
		if (fabs(inputSampleL)<1.18e-23) inputSampleL = fpdL * 1.18e-17;
		if (fabs(inputSampleR)<1.18e-23) inputSampleR = fpdR * 1.18e-17;
		double drySampleL = inputSampleL;
		double drySampleR = inputSampleR;
		
		double tempSampleL = (inputSampleL * biquadA[2]) + biquadA[7];
		biquadA[7] = (inputSampleL * biquadA[3]) - (tempSampleL * biquadA[5]) + biquadA[8];
		biquadA[8] = (inputSampleL * biquadA[4]) - (tempSampleL * biquadA[6]);
		inputSampleL = tempSampleL; //like mono AU, 7 and 8 store L channel
		
		double tempSampleR = (inputSampleR * biquadA[2]) + biquadA[9];
		biquadA[9] = (inputSampleR * biquadA[3]) - (tempSampleR * biquadA[5]) + biquadA[10];
		biquadA[10] = (inputSampleR * biquadA[4]) - (tempSampleR * biquadA[6]);
		inputSampleR = tempSampleR; //note: 9 and 10 store the R channel
				
		double allpassIL = inputSampleL;
		double allpassJL = inputSampleL;
		double allpassKL = inputSampleL;
		double allpassLL = inputSampleL;
		
		double allpassIR = inputSampleR;
		double allpassJR = inputSampleR;
		double allpassKR = inputSampleR;
		double allpassLR = inputSampleR;
		
		int allpasstemp = countI + 1;
		if (allpasstemp < 0 || allpasstemp > delayI) {allpasstemp = 0;}
		allpassIL -= aIL[allpasstemp]*0.5;
		aIL[countI] = allpassIL;
		allpassIL *= 0.5;
		allpassIR -= aIR[allpasstemp]*0.5;
		aIR[countI] = allpassIR;
		allpassIR *= 0.5;
		countI++; if (countI < 0 || countI > delayI) {countI = 0;}		
		allpassIL += (aIL[countI]);
		allpassIR += (aIR[countI]);
		
		allpasstemp = countJ + 1;
		if (allpasstemp < 0 || allpasstemp > delayJ) {allpasstemp = 0;}
		allpassJL -= aJL[allpasstemp]*0.5;
		aJL[countJ] = allpassJL;
		allpassJL *= 0.5;
		allpassJR -= aJR[allpasstemp]*0.5;
		aJR[countJ] = allpassJR;
		allpassJR *= 0.5;
		countJ++; if (countJ < 0 || countJ > delayJ) {countJ = 0;}		
		allpassJL += (aJL[countJ]);
		allpassJR += (aJR[countJ]);
		
		allpasstemp = countK + 1;
		if (allpasstemp < 0 || allpasstemp > delayK) {allpasstemp = 0;}
		allpassKL -= aKL[allpasstemp]*0.5;
		aKL[countK] = allpassKL;
		allpassKL *= 0.5;
		allpassKR -= aKR[allpasstemp]*0.5;
		aKR[countK] = allpassKR;
		allpassKR *= 0.5;
		countK++; if (countK < 0 || countK > delayK) {countK = 0;}		
		allpassKL += (aKL[countK]);
		allpassKR += (aKR[countK]);
		
		allpasstemp = countL + 1;
		if (allpasstemp < 0 || allpasstemp > delayL) {allpasstemp = 0;}
		allpassLL -= aLL[allpasstemp]*0.5;
		aLL[countL] = allpassLL;
		allpassLL *= 0.5;
		allpassLR -= aLR[allpasstemp]*0.5;
		aLR[countL] = allpassLR;
		allpassLR *= 0.5;
		countL++; if (countL < 0 || countL > delayL) {countL = 0;}		
		allpassLL += (aLL[countL]);		
		allpassLR += (aLR[countL]);		
		//the big allpass in front of everything
		
		if (rawPass !=1.0) {
			allpassIL = (allpassIL * rawPass) + (drySampleL * (1.0-rawPass));
			allpassJL = (allpassJL * rawPass) + (drySampleL * (1.0-rawPass));
			allpassKL = (allpassKL * rawPass) + (drySampleL * (1.0-rawPass));
			allpassLL = (allpassLL * rawPass) + (drySampleL * (1.0-rawPass));
			allpassIR = (allpassIR * rawPass) + (drySampleR * (1.0-rawPass));
			allpassJR = (allpassJR * rawPass) + (drySampleR * (1.0-rawPass));
			allpassKR = (allpassKR * rawPass) + (drySampleR * (1.0-rawPass));
			allpassLR = (allpassLR * rawPass) + (drySampleR * (1.0-rawPass));
		}		
		
		aAL[countA] = allpassIL + (feedbackAL*feedback);
		aBL[countB] = allpassJL + (feedbackBL*feedback);
		aCL[countC] = allpassKL + (feedbackCL*feedback);
		aDL[countD] = allpassLL + (feedbackDL*feedback);
		aEL[countE] = allpassIL + (feedbackEL*feedback);
		aFL[countF] = allpassJL + (feedbackFL*feedback);
		aGL[countG] = allpassKL + (feedbackGL*feedback);
		aHL[countH] = allpassLL + (feedbackHL*feedback); //L
		
		aAR[countA] = allpassIR + (feedbackAR*feedback);
		aBR[countB] = allpassJR + (feedbackBR*feedback);
		aCR[countC] = allpassKR + (feedbackCR*feedback);
		aDR[countD] = allpassLR + (feedbackDR*feedback);
		aER[countE] = allpassIR + (feedbackER*feedback);
		aFR[countF] = allpassJR + (feedbackFR*feedback);
		aGR[countG] = allpassKR + (feedbackGR*feedback);
		aHR[countH] = allpassLR + (feedbackHR*feedback); //R
		
		countA++; if (countA < 0 || countA > delayA) {countA = 0;}
		countB++; if (countB < 0 || countB > delayB) {countB = 0;}
		countC++; if (countC < 0 || countC > delayC) {countC = 0;}
		countD++; if (countD < 0 || countD > delayD) {countD = 0;}
		countE++; if (countE < 0 || countE > delayE) {countE = 0;}
		countF++; if (countF < 0 || countF > delayF) {countF = 0;}
		countG++; if (countG < 0 || countG > delayG) {countG = 0;}
		countH++; if (countH < 0 || countH > delayH) {countH = 0;}
		//the Householder matrices (shared between channels, offset is stereo)
		
		double infiniteAL = (aAL[countA-((countA > delayA)?delayA+1:0)] * (1-(damping-floor(damping))) );
		infiniteAL += (aAL[countA+1-((countA+1 > delayA)?delayA+1:0)] * ((damping-floor(damping))) );
		double infiniteBL = (aBL[countB-((countB > delayB)?delayB+1:0)] * (1-(damping-floor(damping))) );
		infiniteBL += (aBL[countB+1-((countB+1 > delayB)?delayB+1:0)] * ((damping-floor(damping))) );
		double infiniteCL = (aCL[countC-((countC > delayC)?delayC+1:0)] * (1-(damping-floor(damping))) );
		infiniteCL += (aCL[countC+1-((countC+1 > delayC)?delayC+1:0)] * ((damping-floor(damping))) );
		double infiniteDL = (aDL[countD-((countD > delayD)?delayD+1:0)] * (1-(damping-floor(damping))) );
		infiniteDL += (aDL[countD+1-((countD+1 > delayD)?delayD+1:0)] * ((damping-floor(damping))) );
		
		double infiniteAR = (aAR[countA-((countA > delayA)?delayA+1:0)] * (1-(damping-floor(damping))) );
		infiniteAR += (aAR[countA+1-((countA+1 > delayA)?delayA+1:0)] * ((damping-floor(damping))) );
		double infiniteBR = (aBR[countB-((countB > delayB)?delayB+1:0)] * (1-(damping-floor(damping))) );
		infiniteBR += (aBR[countB+1-((countB+1 > delayB)?delayB+1:0)] * ((damping-floor(damping))) );
		double infiniteCR = (aCR[countC-((countC > delayC)?delayC+1:0)] * (1-(damping-floor(damping))) );
		infiniteCR += (aCR[countC+1-((countC+1 > delayC)?delayC+1:0)] * ((damping-floor(damping))) );
		double infiniteDR = (aDR[countD-((countD > delayD)?delayD+1:0)] * (1-(damping-floor(damping))) );
		infiniteDR += (aDR[countD+1-((countD+1 > delayD)?delayD+1:0)] * ((damping-floor(damping))) );
		
		double infiniteEL = (aEL[countE-((countE > delayE)?delayE+1:0)] * (1-(damping-floor(damping))) );
		infiniteEL += (aEL[countE+1-((countE+1 > delayE)?delayE+1:0)] * ((damping-floor(damping))) );
		double infiniteFL = (aFL[countF-((countF > delayF)?delayF+1:0)] * (1-(damping-floor(damping))) );
		infiniteFL += (aFL[countF+1-((countF+1 > delayF)?delayF+1:0)] * ((damping-floor(damping))) );
		double infiniteGL = (aGL[countG-((countG > delayG)?delayG+1:0)] * (1-(damping-floor(damping))) );
		infiniteGL += (aGL[countG+1-((countG+1 > delayG)?delayG+1:0)] * ((damping-floor(damping))) );
		double infiniteHL = (aHL[countH-((countH > delayH)?delayH+1:0)] * (1-(damping-floor(damping))) );
		infiniteHL += (aHL[countH+1-((countH+1 > delayH)?delayH+1:0)] * ((damping-floor(damping))) );
		
		double infiniteER = (aER[countE-((countE > delayE)?delayE+1:0)] * (1-(damping-floor(damping))) );
		infiniteER += (aER[countE+1-((countE+1 > delayE)?delayE+1:0)] * ((damping-floor(damping))) );
		double infiniteFR = (aFR[countF-((countF > delayF)?delayF+1:0)] * (1-(damping-floor(damping))) );
		infiniteFR += (aFR[countF+1-((countF+1 > delayF)?delayF+1:0)] * ((damping-floor(damping))) );
		double infiniteGR = (aGR[countG-((countG > delayG)?delayG+1:0)] * (1-(damping-floor(damping))) );
		infiniteGR += (aGR[countG+1-((countG+1 > delayG)?delayG+1:0)] * ((damping-floor(damping))) );
		double infiniteHR = (aHR[countH-((countH > delayH)?delayH+1:0)] * (1-(damping-floor(damping))) );
		infiniteHR += (aHR[countH+1-((countH+1 > delayH)?delayH+1:0)] * ((damping-floor(damping))) );
		
		double dialBackAL = 0.5;
		double dialBackEL = 0.5;
		double dialBackDryL = 0.5;
		if (fabs(infiniteAL)>0.4) dialBackAL -= ((fabs(infiniteAL)-0.4)*0.2);
		if (fabs(infiniteEL)>0.4) dialBackEL -= ((fabs(infiniteEL)-0.4)*0.2);
		if (fabs(drySampleL)>0.4) dialBackDryL -= ((fabs(drySampleL)-0.4)*0.2);
		//we're compressing things subtly so we can feed energy in and not overload
		
		double dialBackAR = 0.5;
		double dialBackER = 0.5;
		double dialBackDryR = 0.5;
		if (fabs(infiniteAR)>0.4) dialBackAR -= ((fabs(infiniteAR)-0.4)*0.2);
		if (fabs(infiniteER)>0.4) dialBackER -= ((fabs(infiniteER)-0.4)*0.2);
		if (fabs(drySampleR)>0.4) dialBackDryR -= ((fabs(drySampleR)-0.4)*0.2);
		//we're compressing things subtly so we can feed energy in and not overload
		
		feedbackAL = (infiniteAL - (infiniteBL + infiniteCL + infiniteDL)) * dialBackAL;
		feedbackBL = (infiniteBL - (infiniteAL + infiniteCL + infiniteDL)) * dialBackDryL;
		feedbackCL = (infiniteCL - (infiniteAL + infiniteBL + infiniteDL)) * dialBackDryL;
		feedbackDL = (infiniteDL - (infiniteAL + infiniteBL + infiniteCL)) * dialBackDryL; //Householder feedback matrix, L
		
		feedbackEL = (infiniteEL - (infiniteFL + infiniteGL + infiniteHL)) * dialBackEL;
		feedbackFL = (infiniteFL - (infiniteEL + infiniteGL + infiniteHL)) * dialBackDryL;
		feedbackGL = (infiniteGL - (infiniteEL + infiniteFL + infiniteHL)) * dialBackDryL;
		feedbackHL = (infiniteHL - (infiniteEL + infiniteFL + infiniteGL)) * dialBackDryL; //Householder feedback matrix, L
		
		feedbackAR = (infiniteAR - (infiniteBR + infiniteCR + infiniteDR)) * dialBackAR;
		feedbackBR = (infiniteBR - (infiniteAR + infiniteCR + infiniteDR)) * dialBackDryR;
		feedbackCR = (infiniteCR - (infiniteAR + infiniteBR + infiniteDR)) * dialBackDryR;
		feedbackDR = (infiniteDR - (infiniteAR + infiniteBR + infiniteCR)) * dialBackDryR; //Householder feedback matrix, R
		
		feedbackER = (infiniteER - (infiniteFR + infiniteGR + infiniteHR)) * dialBackER;
		feedbackFR = (infiniteFR - (infiniteER + infiniteGR + infiniteHR)) * dialBackDryR;
		feedbackGR = (infiniteGR - (infiniteER + infiniteFR + infiniteHR)) * dialBackDryR;
		feedbackHR = (infiniteHR - (infiniteER + infiniteFR + infiniteGR)) * dialBackDryR; //Householder feedback matrix, R
		
		inputSampleL = (infiniteAL + infiniteBL + infiniteCL + infiniteDL + infiniteEL + infiniteFL + infiniteGL + infiniteHL)/8.0;
		inputSampleR = (infiniteAR + infiniteBR + infiniteCR + infiniteDR + infiniteER + infiniteFR + infiniteGR + infiniteHR)/8.0;
		
		tempSampleL = (inputSampleL * biquadB[2]) + biquadB[7];
		biquadB[7] = (inputSampleL * biquadB[3]) - (tempSampleL * biquadB[5]) + biquadB[8];
		biquadB[8] = (inputSampleL * biquadB[4]) - (tempSampleL * biquadB[6]);
		inputSampleL = tempSampleL; //like mono AU, 7 and 8 store L channel
		
		tempSampleR = (inputSampleR * biquadB[2]) + biquadB[9];
		biquadB[9] = (inputSampleR * biquadB[3]) - (tempSampleR * biquadB[5]) + biquadB[10];
		biquadB[10] = (inputSampleR * biquadB[4]) - (tempSampleR * biquadB[6]);
		inputSampleR = tempSampleR; //note: 9 and 10 store the R channel
				
		if (inputSampleL > 1.0) inputSampleL = 1.0;
		if (inputSampleL < -1.0) inputSampleL = -1.0;
		if (inputSampleR > 1.0) inputSampleR = 1.0;
		if (inputSampleR < -1.0) inputSampleR = -1.0;
		//without this, you can get a NaN condition where it spits out DC offset at full blast!
		
		inputSampleL = asin(inputSampleL);
		inputSampleR = asin(inputSampleR);		
		
		tempSampleL = (inputSampleL * biquadC[2]) + biquadC[7];
		biquadC[7] = (inputSampleL * biquadC[3]) - (tempSampleL * biquadC[5]) + biquadC[8];
		biquadC[8] = (inputSampleL * biquadC[4]) - (tempSampleL * biquadC[6]);
		inputSampleL = tempSampleL; //like mono AU, 7 and 8 store L channel
		
		tempSampleR = (inputSampleR * biquadC[2]) + biquadC[9];
		biquadC[9] = (inputSampleR * biquadC[3]) - (tempSampleR * biquadC[5]) + biquadC[10];
		biquadC[10] = (inputSampleR * biquadC[4]) - (tempSampleR * biquadC[6]);
		inputSampleR = tempSampleR; //note: 9 and 10 store the R channel
		
		if (wet !=1.0) {
			inputSampleL = (inputSampleL * wet) + (drySampleL * (1.0-wet));
			inputSampleR = (inputSampleR * wet) + (drySampleR * (1.0-wet));
		}
		
		//begin 32 bit stereo floating point dither
		int expon; frexpf((float)inputSampleL, &expon);
		fpdL ^= fpdL << 13; fpdL ^= fpdL >> 17; fpdL ^= fpdL << 5;
		inputSampleL += ((double(fpdL)-uint32_t(0x7fffffff)) * 5.5e-36l * pow(2,expon+62));
		frexpf((float)inputSampleR, &expon);
		fpdR ^= fpdR << 13; fpdR ^= fpdR >> 17; fpdR ^= fpdR << 5;
		inputSampleR += ((double(fpdR)-uint32_t(0x7fffffff)) * 5.5e-36l * pow(2,expon+62));
		//end 32 bit stereo floating point dither
		
		*out1 = inputSampleL;
		*out2 = inputSampleR;

		*in1++;
		*in2++;
		*out1++;
		*out2++;
    }
}

void Infinity2::processDoubleReplacing(double **inputs, double **outputs, VstInt32 sampleFrames) 
{
    double* in1  =  inputs[0];
    double* in2  =  inputs[1];
    double* out1 = outputs[0];
    double* out2 = outputs[1];
	
	biquadC[0] = biquadB[0] = biquadA[0] = ((pow(A,2)*9900.0)+100.0) / getSampleRate();
	biquadA[1] = 0.618033988749894848204586;
	biquadB[1] = (A*0.5)+0.118033988749894848204586;
    biquadC[1] = 0.5;
	
	double K = tan(M_PI * biquadA[0]); //lowpass
	double norm = 1.0 / (1.0 + K / biquadA[1] + K * K);
	biquadA[2] = K * K * norm;
	biquadA[3] = 2.0 * biquadA[2];
	biquadA[4] = biquadA[2];
	biquadA[5] = 2.0 * (K * K - 1.0) * norm;
	biquadA[6] = (1.0 - K / biquadA[1] + K * K) * norm;
	
	K = tan(M_PI * biquadA[0]);
	norm = 1.0 / (1.0 + K / biquadB[1] + K * K);
	biquadB[2] = K * K * norm;
	biquadB[3] = 2.0 * biquadB[2];
	biquadB[4] = biquadB[2];
	biquadB[5] = 2.0 * (K * K - 1.0) * norm;
	biquadB[6] = (1.0 - K / biquadB[1] + K * K) * norm;
	
	K = tan(M_PI * biquadC[0]);
	norm = 1.0 / (1.0 + K / biquadC[1] + K * K);
	biquadC[2] = K * K * norm;
	biquadC[3] = 2.0 * biquadC[2];
	biquadC[4] = biquadC[2];
	biquadC[5] = 2.0 * (K * K - 1.0) * norm;
	biquadC[6] = (1.0 - K / biquadC[1] + K * K) * norm;
	
	double size = (pow(B,2)*99.0)+1.0;
	double damping = pow(C,2)*0.5;
	double rawPass = D;
	double feedback = 1.0-(pow(1.0-E,4));
	double wet = F;
	
	delayA = 79*size;
	delayB = 73*size;
	delayC = 71*size;
	delayD = 67*size;
	delayE = 61*size;
	delayF = 59*size;
	delayG = 53*size;
	delayH = 47*size;
	
	delayI = 43*size;
	delayJ = 41*size;
	delayK = 37*size;
	delayL = 31*size;
	
    while (--sampleFrames >= 0)
    {
		double inputSampleL = *in1;
		double inputSampleR = *in2;
		if (fabs(inputSampleL)<1.18e-23) inputSampleL = fpdL * 1.18e-17;
		if (fabs(inputSampleR)<1.18e-23) inputSampleR = fpdR * 1.18e-17;
		double drySampleL = inputSampleL;
		double drySampleR = inputSampleR;
		
		double tempSampleL = (inputSampleL * biquadA[2]) + biquadA[7];
		biquadA[7] = (inputSampleL * biquadA[3]) - (tempSampleL * biquadA[5]) + biquadA[8];
		biquadA[8] = (inputSampleL * biquadA[4]) - (tempSampleL * biquadA[6]);
		inputSampleL = tempSampleL; //like mono AU, 7 and 8 store L channel
		
		double tempSampleR = (inputSampleR * biquadA[2]) + biquadA[9];
		biquadA[9] = (inputSampleR * biquadA[3]) - (tempSampleR * biquadA[5]) + biquadA[10];
		biquadA[10] = (inputSampleR * biquadA[4]) - (tempSampleR * biquadA[6]);
		inputSampleR = tempSampleR; //note: 9 and 10 store the R channel
				
		double allpassIL = inputSampleL;
		double allpassJL = inputSampleL;
		double allpassKL = inputSampleL;
		double allpassLL = inputSampleL;
		
		double allpassIR = inputSampleR;
		double allpassJR = inputSampleR;
		double allpassKR = inputSampleR;
		double allpassLR = inputSampleR;
		
		int allpasstemp = countI + 1;
		if (allpasstemp < 0 || allpasstemp > delayI) {allpasstemp = 0;}
		allpassIL -= aIL[allpasstemp]*0.5;
		aIL[countI] = allpassIL;
		allpassIL *= 0.5;
		allpassIR -= aIR[allpasstemp]*0.5;
		aIR[countI] = allpassIR;
		allpassIR *= 0.5;
		countI++; if (countI < 0 || countI > delayI) {countI = 0;}		
		allpassIL += (aIL[countI]);
		allpassIR += (aIR[countI]);
		
		allpasstemp = countJ + 1;
		if (allpasstemp < 0 || allpasstemp > delayJ) {allpasstemp = 0;}
		allpassJL -= aJL[allpasstemp]*0.5;
		aJL[countJ] = allpassJL;
		allpassJL *= 0.5;
		allpassJR -= aJR[allpasstemp]*0.5;
		aJR[countJ] = allpassJR;
		allpassJR *= 0.5;
		countJ++; if (countJ < 0 || countJ > delayJ) {countJ = 0;}		
		allpassJL += (aJL[countJ]);
		allpassJR += (aJR[countJ]);
		
		allpasstemp = countK + 1;
		if (allpasstemp < 0 || allpasstemp > delayK) {allpasstemp = 0;}
		allpassKL -= aKL[allpasstemp]*0.5;
		aKL[countK] = allpassKL;
		allpassKL *= 0.5;
		allpassKR -= aKR[allpasstemp]*0.5;
		aKR[countK] = allpassKR;
		allpassKR *= 0.5;
		countK++; if (countK < 0 || countK > delayK) {countK = 0;}		
		allpassKL += (aKL[countK]);
		allpassKR += (aKR[countK]);
		
		allpasstemp = countL + 1;
		if (allpasstemp < 0 || allpasstemp > delayL) {allpasstemp = 0;}
		allpassLL -= aLL[allpasstemp]*0.5;
		aLL[countL] = allpassLL;
		allpassLL *= 0.5;
		allpassLR -= aLR[allpasstemp]*0.5;
		aLR[countL] = allpassLR;
		allpassLR *= 0.5;
		countL++; if (countL < 0 || countL > delayL) {countL = 0;}		
		allpassLL += (aLL[countL]);		
		allpassLR += (aLR[countL]);		
		//the big allpass in front of everything
		
		if (rawPass !=1.0) {
			allpassIL = (allpassIL * rawPass) + (drySampleL * (1.0-rawPass));
			allpassJL = (allpassJL * rawPass) + (drySampleL * (1.0-rawPass));
			allpassKL = (allpassKL * rawPass) + (drySampleL * (1.0-rawPass));
			allpassLL = (allpassLL * rawPass) + (drySampleL * (1.0-rawPass));
			allpassIR = (allpassIR * rawPass) + (drySampleR * (1.0-rawPass));
			allpassJR = (allpassJR * rawPass) + (drySampleR * (1.0-rawPass));
			allpassKR = (allpassKR * rawPass) + (drySampleR * (1.0-rawPass));
			allpassLR = (allpassLR * rawPass) + (drySampleR * (1.0-rawPass));
		}		

		aAL[countA] = allpassIL + (feedbackAL*feedback);
		aBL[countB] = allpassJL + (feedbackBL*feedback);
		aCL[countC] = allpassKL + (feedbackCL*feedback);
		aDL[countD] = allpassLL + (feedbackDL*feedback);
		aEL[countE] = allpassIL + (feedbackEL*feedback);
		aFL[countF] = allpassJL + (feedbackFL*feedback);
		aGL[countG] = allpassKL + (feedbackGL*feedback);
		aHL[countH] = allpassLL + (feedbackHL*feedback); //L
		
		aAR[countA] = allpassIR + (feedbackAR*feedback);
		aBR[countB] = allpassJR + (feedbackBR*feedback);
		aCR[countC] = allpassKR + (feedbackCR*feedback);
		aDR[countD] = allpassLR + (feedbackDR*feedback);
		aER[countE] = allpassIR + (feedbackER*feedback);
		aFR[countF] = allpassJR + (feedbackFR*feedback);
		aGR[countG] = allpassKR + (feedbackGR*feedback);
		aHR[countH] = allpassLR + (feedbackHR*feedback); //R
		
		countA++; if (countA < 0 || countA > delayA) {countA = 0;}
		countB++; if (countB < 0 || countB > delayB) {countB = 0;}
		countC++; if (countC < 0 || countC > delayC) {countC = 0;}
		countD++; if (countD < 0 || countD > delayD) {countD = 0;}
		countE++; if (countE < 0 || countE > delayE) {countE = 0;}
		countF++; if (countF < 0 || countF > delayF) {countF = 0;}
		countG++; if (countG < 0 || countG > delayG) {countG = 0;}
		countH++; if (countH < 0 || countH > delayH) {countH = 0;}
		//the Householder matrices (shared between channels, offset is stereo)
		
		double infiniteAL = (aAL[countA-((countA > delayA)?delayA+1:0)] * (1-(damping-floor(damping))) );
		infiniteAL += (aAL[countA+1-((countA+1 > delayA)?delayA+1:0)] * ((damping-floor(damping))) );
		double infiniteBL = (aBL[countB-((countB > delayB)?delayB+1:0)] * (1-(damping-floor(damping))) );
		infiniteBL += (aBL[countB+1-((countB+1 > delayB)?delayB+1:0)] * ((damping-floor(damping))) );
		double infiniteCL = (aCL[countC-((countC > delayC)?delayC+1:0)] * (1-(damping-floor(damping))) );
		infiniteCL += (aCL[countC+1-((countC+1 > delayC)?delayC+1:0)] * ((damping-floor(damping))) );
		double infiniteDL = (aDL[countD-((countD > delayD)?delayD+1:0)] * (1-(damping-floor(damping))) );
		infiniteDL += (aDL[countD+1-((countD+1 > delayD)?delayD+1:0)] * ((damping-floor(damping))) );
		
		double infiniteAR = (aAR[countA-((countA > delayA)?delayA+1:0)] * (1-(damping-floor(damping))) );
		infiniteAR += (aAR[countA+1-((countA+1 > delayA)?delayA+1:0)] * ((damping-floor(damping))) );
		double infiniteBR = (aBR[countB-((countB > delayB)?delayB+1:0)] * (1-(damping-floor(damping))) );
		infiniteBR += (aBR[countB+1-((countB+1 > delayB)?delayB+1:0)] * ((damping-floor(damping))) );
		double infiniteCR = (aCR[countC-((countC > delayC)?delayC+1:0)] * (1-(damping-floor(damping))) );
		infiniteCR += (aCR[countC+1-((countC+1 > delayC)?delayC+1:0)] * ((damping-floor(damping))) );
		double infiniteDR = (aDR[countD-((countD > delayD)?delayD+1:0)] * (1-(damping-floor(damping))) );
		infiniteDR += (aDR[countD+1-((countD+1 > delayD)?delayD+1:0)] * ((damping-floor(damping))) );
		
		double infiniteEL = (aEL[countE-((countE > delayE)?delayE+1:0)] * (1-(damping-floor(damping))) );
		infiniteEL += (aEL[countE+1-((countE+1 > delayE)?delayE+1:0)] * ((damping-floor(damping))) );
		double infiniteFL = (aFL[countF-((countF > delayF)?delayF+1:0)] * (1-(damping-floor(damping))) );
		infiniteFL += (aFL[countF+1-((countF+1 > delayF)?delayF+1:0)] * ((damping-floor(damping))) );
		double infiniteGL = (aGL[countG-((countG > delayG)?delayG+1:0)] * (1-(damping-floor(damping))) );
		infiniteGL += (aGL[countG+1-((countG+1 > delayG)?delayG+1:0)] * ((damping-floor(damping))) );
		double infiniteHL = (aHL[countH-((countH > delayH)?delayH+1:0)] * (1-(damping-floor(damping))) );
		infiniteHL += (aHL[countH+1-((countH+1 > delayH)?delayH+1:0)] * ((damping-floor(damping))) );
		
		double infiniteER = (aER[countE-((countE > delayE)?delayE+1:0)] * (1-(damping-floor(damping))) );
		infiniteER += (aER[countE+1-((countE+1 > delayE)?delayE+1:0)] * ((damping-floor(damping))) );
		double infiniteFR = (aFR[countF-((countF > delayF)?delayF+1:0)] * (1-(damping-floor(damping))) );
		infiniteFR += (aFR[countF+1-((countF+1 > delayF)?delayF+1:0)] * ((damping-floor(damping))) );
		double infiniteGR = (aGR[countG-((countG > delayG)?delayG+1:0)] * (1-(damping-floor(damping))) );
		infiniteGR += (aGR[countG+1-((countG+1 > delayG)?delayG+1:0)] * ((damping-floor(damping))) );
		double infiniteHR = (aHR[countH-((countH > delayH)?delayH+1:0)] * (1-(damping-floor(damping))) );
		infiniteHR += (aHR[countH+1-((countH+1 > delayH)?delayH+1:0)] * ((damping-floor(damping))) );
		
		double dialBackAL = 0.5;
		double dialBackEL = 0.5;
		double dialBackDryL = 0.5;
		if (fabs(infiniteAL)>0.4) dialBackAL -= ((fabs(infiniteAL)-0.4)*0.2);
		if (fabs(infiniteEL)>0.4) dialBackEL -= ((fabs(infiniteEL)-0.4)*0.2);
		if (fabs(drySampleL)>0.4) dialBackDryL -= ((fabs(drySampleL)-0.4)*0.2);
		//we're compressing things subtly so we can feed energy in and not overload
		
		double dialBackAR = 0.5;
		double dialBackER = 0.5;
		double dialBackDryR = 0.5;
		if (fabs(infiniteAR)>0.4) dialBackAR -= ((fabs(infiniteAR)-0.4)*0.2);
		if (fabs(infiniteER)>0.4) dialBackER -= ((fabs(infiniteER)-0.4)*0.2);
		if (fabs(drySampleR)>0.4) dialBackDryR -= ((fabs(drySampleR)-0.4)*0.2);
		//we're compressing things subtly so we can feed energy in and not overload
		
		feedbackAL = (infiniteAL - (infiniteBL + infiniteCL + infiniteDL)) * dialBackAL;
		feedbackBL = (infiniteBL - (infiniteAL + infiniteCL + infiniteDL)) * dialBackDryL;
		feedbackCL = (infiniteCL - (infiniteAL + infiniteBL + infiniteDL)) * dialBackDryL;
		feedbackDL = (infiniteDL - (infiniteAL + infiniteBL + infiniteCL)) * dialBackDryL; //Householder feedback matrix, L
		
		feedbackEL = (infiniteEL - (infiniteFL + infiniteGL + infiniteHL)) * dialBackEL;
		feedbackFL = (infiniteFL - (infiniteEL + infiniteGL + infiniteHL)) * dialBackDryL;
		feedbackGL = (infiniteGL - (infiniteEL + infiniteFL + infiniteHL)) * dialBackDryL;
		feedbackHL = (infiniteHL - (infiniteEL + infiniteFL + infiniteGL)) * dialBackDryL; //Householder feedback matrix, L
		
		feedbackAR = (infiniteAR - (infiniteBR + infiniteCR + infiniteDR)) * dialBackAR;
		feedbackBR = (infiniteBR - (infiniteAR + infiniteCR + infiniteDR)) * dialBackDryR;
		feedbackCR = (infiniteCR - (infiniteAR + infiniteBR + infiniteDR)) * dialBackDryR;
		feedbackDR = (infiniteDR - (infiniteAR + infiniteBR + infiniteCR)) * dialBackDryR; //Householder feedback matrix, R
		
		feedbackER = (infiniteER - (infiniteFR + infiniteGR + infiniteHR)) * dialBackER;
		feedbackFR = (infiniteFR - (infiniteER + infiniteGR + infiniteHR)) * dialBackDryR;
		feedbackGR = (infiniteGR - (infiniteER + infiniteFR + infiniteHR)) * dialBackDryR;
		feedbackHR = (infiniteHR - (infiniteER + infiniteFR + infiniteGR)) * dialBackDryR; //Householder feedback matrix, R
		
		inputSampleL = (infiniteAL + infiniteBL + infiniteCL + infiniteDL + infiniteEL + infiniteFL + infiniteGL + infiniteHL)/8.0;
		inputSampleR = (infiniteAR + infiniteBR + infiniteCR + infiniteDR + infiniteER + infiniteFR + infiniteGR + infiniteHR)/8.0;
		
		tempSampleL = (inputSampleL * biquadB[2]) + biquadB[7];
		biquadB[7] = (inputSampleL * biquadB[3]) - (tempSampleL * biquadB[5]) + biquadB[8];
		biquadB[8] = (inputSampleL * biquadB[4]) - (tempSampleL * biquadB[6]);
		inputSampleL = tempSampleL; //like mono AU, 7 and 8 store L channel
		
		tempSampleR = (inputSampleR * biquadB[2]) + biquadB[9];
		biquadB[9] = (inputSampleR * biquadB[3]) - (tempSampleR * biquadB[5]) + biquadB[10];
		biquadB[10] = (inputSampleR * biquadB[4]) - (tempSampleR * biquadB[6]);
		inputSampleR = tempSampleR; //note: 9 and 10 store the R channel
				
		if (inputSampleL > 1.0) inputSampleL = 1.0;
		if (inputSampleL < -1.0) inputSampleL = -1.0;
		if (inputSampleR > 1.0) inputSampleR = 1.0;
		if (inputSampleR < -1.0) inputSampleR = -1.0;
		//without this, you can get a NaN condition where it spits out DC offset at full blast!
		
		inputSampleL = asin(inputSampleL);
		inputSampleR = asin(inputSampleR);		
		
		tempSampleL = (inputSampleL * biquadC[2]) + biquadC[7];
		biquadC[7] = (inputSampleL * biquadC[3]) - (tempSampleL * biquadC[5]) + biquadC[8];
		biquadC[8] = (inputSampleL * biquadC[4]) - (tempSampleL * biquadC[6]);
		inputSampleL = tempSampleL; //like mono AU, 7 and 8 store L channel
		
		tempSampleR = (inputSampleR * biquadC[2]) + biquadC[9];
		biquadC[9] = (inputSampleR * biquadC[3]) - (tempSampleR * biquadC[5]) + biquadC[10];
		biquadC[10] = (inputSampleR * biquadC[4]) - (tempSampleR * biquadC[6]);
		inputSampleR = tempSampleR; //note: 9 and 10 store the R channel
		
		if (wet !=1.0) {
			inputSampleL = (inputSampleL * wet) + (drySampleL * (1.0-wet));
			inputSampleR = (inputSampleR * wet) + (drySampleR * (1.0-wet));
		}
		
		//begin 64 bit stereo floating point dither
		//int expon; frexp((double)inputSampleL, &expon);
		fpdL ^= fpdL << 13; fpdL ^= fpdL >> 17; fpdL ^= fpdL << 5;
		//inputSampleL += ((double(fpdL)-uint32_t(0x7fffffff)) * 1.1e-44l * pow(2,expon+62));
		//frexp((double)inputSampleR, &expon);
		fpdR ^= fpdR << 13; fpdR ^= fpdR >> 17; fpdR ^= fpdR << 5;
		//inputSampleR += ((double(fpdR)-uint32_t(0x7fffffff)) * 1.1e-44l * pow(2,expon+62));
		//end 64 bit stereo floating point dither
		
		*out1 = inputSampleL;
		*out2 = inputSampleR;

		*in1++;
		*in2++;
		*out1++;
		*out2++;
    }
}