diff options
Diffstat (limited to 'plugins/MacVST/TapeDelay/source/TapeDelayProc.cpp')
-rwxr-xr-x | plugins/MacVST/TapeDelay/source/TapeDelayProc.cpp | 363 |
1 files changed, 363 insertions, 0 deletions
diff --git a/plugins/MacVST/TapeDelay/source/TapeDelayProc.cpp b/plugins/MacVST/TapeDelay/source/TapeDelayProc.cpp new file mode 100755 index 0000000..43fa743 --- /dev/null +++ b/plugins/MacVST/TapeDelay/source/TapeDelayProc.cpp @@ -0,0 +1,363 @@ +/* ======================================== + * TapeDelay - TapeDelay.h + * Copyright (c) 2016 airwindows, All rights reserved + * ======================================== */ + +#ifndef __TapeDelay_H +#include "TapeDelay.h" +#endif + +void TapeDelay::processReplacing(float **inputs, float **outputs, VstInt32 sampleFrames) +{ + float* in1 = inputs[0]; + float* in2 = inputs[1]; + float* out1 = outputs[0]; + float* out2 = outputs[1]; + + double dry = pow(A,2); + double wet = pow(B,2); + int targetdelay = (int)(44000*C); + double feedback = (D*1.3); + double leanfat = ((E*2.0)-1.0); + double fatwet = fabs(leanfat); + int fatness = (int)floor((F*29.0)+3.0); + int count; + + double storedelayL; + double storedelayR; + double sumL = 0.0; + double sumR = 0.0; + double floattotalL = 0.0; + double floattotalR = 0.0; + int sumtotalL = 0; + int sumtotalR = 0; + + + while (--sampleFrames >= 0) + { + long double inputSampleL = *in1; + long double inputSampleR = *in2; + + static int noisesourceL = 0; + static int noisesourceR = 850010; + int residue; + double applyresidue; + + noisesourceL = noisesourceL % 1700021; noisesourceL++; + residue = noisesourceL * noisesourceL; + residue = residue % 170003; residue *= residue; + residue = residue % 17011; residue *= residue; + residue = residue % 1709; residue *= residue; + residue = residue % 173; residue *= residue; + residue = residue % 17; + applyresidue = residue; + applyresidue *= 0.00000001; + applyresidue *= 0.00000001; + inputSampleL += applyresidue; + if (inputSampleL<1.2e-38 && -inputSampleL<1.2e-38) { + inputSampleL -= applyresidue; + } + + noisesourceR = noisesourceR % 1700021; noisesourceR++; + residue = noisesourceR * noisesourceR; + residue = residue % 170003; residue *= residue; + residue = residue % 17011; residue *= residue; + residue = residue % 1709; residue *= residue; + residue = residue % 173; residue *= residue; + residue = residue % 17; + applyresidue = residue; + applyresidue *= 0.00000001; + applyresidue *= 0.00000001; + inputSampleR += applyresidue; + if (inputSampleR<1.2e-38 && -inputSampleR<1.2e-38) { + inputSampleR -= applyresidue; + } + //for live air, we always apply the dither noise. Then, if our result is + //effectively digital black, we'll subtract it again. We want a 'air' hiss + + if (gcount < 0 || gcount > 128) {gcount = 128;} + count = gcount; + if (delay < 0 || delay > maxdelay) {delay = maxdelay;} + + sumL = inputSampleL + (dL[delay]*feedback); + sumR = inputSampleR + (dR[delay]*feedback); + pL[count+128] = pL[count] = sumtotalL = (int)(sumL*8388608.0); + pR[count+128] = pR[count] = sumtotalR = (int)(sumR*8388608.0); + + switch (fatness) + { + case 32: sumtotalL += pL[count+127]; sumtotalR += pR[count+127]; //note NO break statement. + case 31: sumtotalL += pL[count+113]; sumtotalR += pR[count+113]; //This jumps to the relevant tap + case 30: sumtotalL += pL[count+109]; sumtotalR += pR[count+109]; //and then includes all smaller taps. + case 29: sumtotalL += pL[count+107]; sumtotalR += pR[count+107]; + case 28: sumtotalL += pL[count+103]; sumtotalR += pR[count+103]; + case 27: sumtotalL += pL[count+101]; sumtotalR += pR[count+101]; + case 26: sumtotalL += pL[count+97]; sumtotalR += pR[count+97]; + case 25: sumtotalL += pL[count+89]; sumtotalR += pR[count+89]; + case 24: sumtotalL += pL[count+83]; sumtotalR += pR[count+83]; + case 23: sumtotalL += pL[count+79]; sumtotalR += pR[count+79]; + case 22: sumtotalL += pL[count+73]; sumtotalR += pR[count+73]; + case 21: sumtotalL += pL[count+71]; sumtotalR += pR[count+71]; + case 20: sumtotalL += pL[count+67]; sumtotalR += pR[count+67]; + case 19: sumtotalL += pL[count+61]; sumtotalR += pR[count+61]; + case 18: sumtotalL += pL[count+59]; sumtotalR += pR[count+59]; + case 17: sumtotalL += pL[count+53]; sumtotalR += pR[count+53]; + case 16: sumtotalL += pL[count+47]; sumtotalR += pR[count+47]; + case 15: sumtotalL += pL[count+43]; sumtotalR += pR[count+43]; + case 14: sumtotalL += pL[count+41]; sumtotalR += pR[count+41]; + case 13: sumtotalL += pL[count+37]; sumtotalR += pR[count+37]; + case 12: sumtotalL += pL[count+31]; sumtotalR += pR[count+31]; + case 11: sumtotalL += pL[count+29]; sumtotalR += pR[count+29]; + case 10: sumtotalL += pL[count+23]; sumtotalR += pR[count+23]; + case 9: sumtotalL += pL[count+19]; sumtotalR += pR[count+19]; + case 8: sumtotalL += pL[count+17]; sumtotalR += pR[count+17]; + case 7: sumtotalL += pL[count+13]; sumtotalR += pR[count+13]; + case 6: sumtotalL += pL[count+11]; sumtotalR += pR[count+11]; + case 5: sumtotalL += pL[count+7]; sumtotalR += pR[count+7]; + case 4: sumtotalL += pL[count+5]; sumtotalR += pR[count+5]; + case 3: sumtotalL += pL[count+3]; sumtotalR += pR[count+3]; + case 2: sumtotalL += pL[count+2]; sumtotalR += pR[count+2]; + case 1: sumtotalL += pL[count+1]; sumtotalR += pR[count+1]; + } + + floattotalL = (double)(sumtotalL/fatness+1); + floattotalR = (double)(sumtotalR/fatness+1); + floattotalL /= 8388608.0; + floattotalR /= 8388608.0; + floattotalL *= fatwet; + floattotalR *= fatwet; + if (leanfat < 0) {storedelayL = sumL-floattotalL; storedelayR = sumR-floattotalR;} + else {storedelayL = (sumL * (1-fatwet))+floattotalL; storedelayR = (sumR * (1-fatwet))+floattotalR;} + + chase += abs(maxdelay - targetdelay); + if (chase > 9000) + { + if (maxdelay > targetdelay) { + dL[delay] = storedelayL; dR[delay] = storedelayR; + maxdelay -= 1; delay -= 1; if (delay < 0) {delay = maxdelay;} + dL[delay] = storedelayL; dR[delay] = storedelayR; + } + if (maxdelay < targetdelay) { + maxdelay += 1; delay += 1; if (delay > maxdelay) {delay = 0;} + dL[delay] = storedelayL; dR[delay] = storedelayR; + } + chase = 0; + } + else + { + dL[delay] = storedelayL; dR[delay] = storedelayR; + } + + gcount--; + delay--; + if (delay < 0 || delay > maxdelay) {delay = maxdelay;} + //yes this is a second bounds check. it's cheap, check EVERY time + + inputSampleL = (inputSampleL * dry) + (dL[delay] * wet); + inputSampleR = (inputSampleR * dry) + (dR[delay] * wet); + + //noise shaping to 32-bit floating point + float fpTemp = inputSampleL; + fpNShapeL += (inputSampleL-fpTemp); + inputSampleL += fpNShapeL; + //if this confuses you look at the wordlength for fpTemp :) + fpTemp = inputSampleR; + fpNShapeR += (inputSampleR-fpTemp); + inputSampleR += fpNShapeR; + //for deeper space and warmth, we try a non-oscillating noise shaping + //that is kind of ruthless: it will forever retain the rounding errors + //except we'll dial it back a hair at the end of every buffer processed + //end noise shaping on 32 bit output + + *out1 = inputSampleL; + *out2 = inputSampleR; + + *in1++; + *in2++; + *out1++; + *out2++; + } + fpNShapeL *= 0.999999; + fpNShapeR *= 0.999999; + //we will just delicately dial back the FP noise shaping, not even every sample + //this is a good place to put subtle 'no runaway' calculations, though bear in mind + //that it will be called more often when you use shorter sample buffers in the DAW. + //So, very low latency operation will call these calculations more often. +} + +void TapeDelay::processDoubleReplacing(double **inputs, double **outputs, VstInt32 sampleFrames) +{ + double* in1 = inputs[0]; + double* in2 = inputs[1]; + double* out1 = outputs[0]; + double* out2 = outputs[1]; + + double dry = pow(A,2); + double wet = pow(B,2); + int targetdelay = (int)(44000*C); + double feedback = (D*1.3); + double leanfat = ((E*2.0)-1.0); + double fatwet = fabs(leanfat); + int fatness = (int)floor((F*29.0)+3.0); + int count; + + double storedelayL; + double storedelayR; + double sumL = 0.0; + double sumR = 0.0; + double floattotalL = 0.0; + double floattotalR = 0.0; + int sumtotalL = 0; + int sumtotalR = 0; + + while (--sampleFrames >= 0) + { + long double inputSampleL = *in1; + long double inputSampleR = *in2; + + static int noisesourceL = 0; + static int noisesourceR = 850010; + int residue; + double applyresidue; + + noisesourceL = noisesourceL % 1700021; noisesourceL++; + residue = noisesourceL * noisesourceL; + residue = residue % 170003; residue *= residue; + residue = residue % 17011; residue *= residue; + residue = residue % 1709; residue *= residue; + residue = residue % 173; residue *= residue; + residue = residue % 17; + applyresidue = residue; + applyresidue *= 0.00000001; + applyresidue *= 0.00000001; + inputSampleL += applyresidue; + if (inputSampleL<1.2e-38 && -inputSampleL<1.2e-38) { + inputSampleL -= applyresidue; + } + + noisesourceR = noisesourceR % 1700021; noisesourceR++; + residue = noisesourceR * noisesourceR; + residue = residue % 170003; residue *= residue; + residue = residue % 17011; residue *= residue; + residue = residue % 1709; residue *= residue; + residue = residue % 173; residue *= residue; + residue = residue % 17; + applyresidue = residue; + applyresidue *= 0.00000001; + applyresidue *= 0.00000001; + inputSampleR += applyresidue; + if (inputSampleR<1.2e-38 && -inputSampleR<1.2e-38) { + inputSampleR -= applyresidue; + } + //for live air, we always apply the dither noise. Then, if our result is + //effectively digital black, we'll subtract it again. We want a 'air' hiss + + if (gcount < 0 || gcount > 128) {gcount = 128;} + count = gcount; + if (delay < 0 || delay > maxdelay) {delay = maxdelay;} + + sumL = inputSampleL + (dL[delay]*feedback); + sumR = inputSampleR + (dR[delay]*feedback); + pL[count+128] = pL[count] = sumtotalL = (int)(sumL*8388608.0); + pR[count+128] = pR[count] = sumtotalR = (int)(sumR*8388608.0); + + switch (fatness) + { + case 32: sumtotalL += pL[count+127]; sumtotalR += pR[count+127]; //note NO break statement. + case 31: sumtotalL += pL[count+113]; sumtotalR += pR[count+113]; //This jumps to the relevant tap + case 30: sumtotalL += pL[count+109]; sumtotalR += pR[count+109]; //and then includes all smaller taps. + case 29: sumtotalL += pL[count+107]; sumtotalR += pR[count+107]; + case 28: sumtotalL += pL[count+103]; sumtotalR += pR[count+103]; + case 27: sumtotalL += pL[count+101]; sumtotalR += pR[count+101]; + case 26: sumtotalL += pL[count+97]; sumtotalR += pR[count+97]; + case 25: sumtotalL += pL[count+89]; sumtotalR += pR[count+89]; + case 24: sumtotalL += pL[count+83]; sumtotalR += pR[count+83]; + case 23: sumtotalL += pL[count+79]; sumtotalR += pR[count+79]; + case 22: sumtotalL += pL[count+73]; sumtotalR += pR[count+73]; + case 21: sumtotalL += pL[count+71]; sumtotalR += pR[count+71]; + case 20: sumtotalL += pL[count+67]; sumtotalR += pR[count+67]; + case 19: sumtotalL += pL[count+61]; sumtotalR += pR[count+61]; + case 18: sumtotalL += pL[count+59]; sumtotalR += pR[count+59]; + case 17: sumtotalL += pL[count+53]; sumtotalR += pR[count+53]; + case 16: sumtotalL += pL[count+47]; sumtotalR += pR[count+47]; + case 15: sumtotalL += pL[count+43]; sumtotalR += pR[count+43]; + case 14: sumtotalL += pL[count+41]; sumtotalR += pR[count+41]; + case 13: sumtotalL += pL[count+37]; sumtotalR += pR[count+37]; + case 12: sumtotalL += pL[count+31]; sumtotalR += pR[count+31]; + case 11: sumtotalL += pL[count+29]; sumtotalR += pR[count+29]; + case 10: sumtotalL += pL[count+23]; sumtotalR += pR[count+23]; + case 9: sumtotalL += pL[count+19]; sumtotalR += pR[count+19]; + case 8: sumtotalL += pL[count+17]; sumtotalR += pR[count+17]; + case 7: sumtotalL += pL[count+13]; sumtotalR += pR[count+13]; + case 6: sumtotalL += pL[count+11]; sumtotalR += pR[count+11]; + case 5: sumtotalL += pL[count+7]; sumtotalR += pR[count+7]; + case 4: sumtotalL += pL[count+5]; sumtotalR += pR[count+5]; + case 3: sumtotalL += pL[count+3]; sumtotalR += pR[count+3]; + case 2: sumtotalL += pL[count+2]; sumtotalR += pR[count+2]; + case 1: sumtotalL += pL[count+1]; sumtotalR += pR[count+1]; + } + + floattotalL = (double)(sumtotalL/fatness+1); + floattotalR = (double)(sumtotalR/fatness+1); + floattotalL /= 8388608.0; + floattotalR /= 8388608.0; + floattotalL *= fatwet; + floattotalR *= fatwet; + if (leanfat < 0) {storedelayL = sumL-floattotalL; storedelayR = sumR-floattotalR;} + else {storedelayL = (sumL * (1-fatwet))+floattotalL; storedelayR = (sumR * (1-fatwet))+floattotalR;} + + chase += abs(maxdelay - targetdelay); + if (chase > 9000) + { + if (maxdelay > targetdelay) { + dL[delay] = storedelayL; dR[delay] = storedelayR; + maxdelay -= 1; delay -= 1; if (delay < 0) {delay = maxdelay;} + dL[delay] = storedelayL; dR[delay] = storedelayR; + } + if (maxdelay < targetdelay) { + maxdelay += 1; delay += 1; if (delay > maxdelay) {delay = 0;} + dL[delay] = storedelayL; dR[delay] = storedelayR; + } + chase = 0; + } + else + { + dL[delay] = storedelayL; dR[delay] = storedelayR; + } + + gcount--; + delay--; + if (delay < 0 || delay > maxdelay) {delay = maxdelay;} + //yes this is a second bounds check. it's cheap, check EVERY time + + inputSampleL = (inputSampleL * dry) + (dL[delay] * wet); + inputSampleR = (inputSampleR * dry) + (dR[delay] * wet); + + //noise shaping to 64-bit floating point + double fpTemp = inputSampleL; + fpNShapeL += (inputSampleL-fpTemp); + inputSampleL += fpNShapeL; + //if this confuses you look at the wordlength for fpTemp :) + fpTemp = inputSampleR; + fpNShapeR += (inputSampleR-fpTemp); + inputSampleR += fpNShapeR; + //for deeper space and warmth, we try a non-oscillating noise shaping + //that is kind of ruthless: it will forever retain the rounding errors + //except we'll dial it back a hair at the end of every buffer processed + //end noise shaping on 64 bit output + + *out1 = inputSampleL; + *out2 = inputSampleR; + + *in1++; + *in2++; + *out1++; + *out2++; + } + fpNShapeL *= 0.999999; + fpNShapeR *= 0.999999; + //we will just delicately dial back the FP noise shaping, not even every sample + //this is a good place to put subtle 'no runaway' calculations, though bear in mind + //that it will be called more often when you use shorter sample buffers in the DAW. + //So, very low latency operation will call these calculations more often. +} |