/* ======================================== * Pear - Pear.h * Copyright (c) airwindows, Airwindows uses the MIT license * ======================================== */ #ifndef __Pear_H #include "Pear.h" #endif void Pear::processReplacing(float **inputs, float **outputs, VstInt32 sampleFrames) { float* in1 = inputs[0]; float* in2 = inputs[1]; float* out1 = outputs[0]; float* out2 = outputs[1]; double overallscale = 1.0; overallscale /= 44100.0; overallscale *= getSampleRate(); int cycleEnd = floor(overallscale); if (cycleEnd < 1) cycleEnd = 1; if (cycleEnd > 3) cycleEnd = 3; cycleEnd--; //this is going to be 0 for 44.1 or 48k, 1 for 88.2 or 96k, 2 for 176 or 192k. //DIFFERENT! Offsetting the bt shift int bitshiftFreq = (10-(A*10.0))+cycleEnd; double freq = 1.0; switch (bitshiftFreq) { case 16: freq = 0.0000152587890625; break; case 15: freq = 0.000030517578125; break; case 14: freq = 0.00006103515625; break; case 13: freq = 0.0001220703125; break; case 12: freq = 0.000244140625; break; case 11: freq = 0.00048828125; break; case 10: freq = 0.0009765625; break; case 9: freq = 0.001953125; break; case 8: freq = 0.00390625; break; case 7: freq = 0.0078125; break; case 6: freq = 0.015625; break; case 5: freq = 0.03125; break; case 4: freq = 0.0625; break; case 3: freq = 0.125; break; case 2: freq = 0.25; break; case 1: freq = 0.5; break; case 0: freq = 1.0; break; } double maxPoles = B*pear_total; double wet = (C*2.0)-1.0; //inv-dry-wet for highpass double dry = 2.0-(C*2.0); if (dry > 1.0) dry = 1.0; //full dry for use with inv, to 0.0 at full wet 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; //Pear is a variation on a Holt filter, made to act like my biquad filters: //its variables (L and R prevSample and prevSlew) contained in arrays for easier handling. //On top of that, the array is used to facilitate use of the filter in a stack which will //sharpen its very weak frequency response performance (3dB/oct per pole) and allow it to run //in a loop. Many earlier Airwindows filters did this 'unrolled', often with variables given //lettered suffixes like iirA through iirZ. The use of an array will make this a lot tidier. for (int x = 0; x < maxPoles; x += 4) { double slew = ((inputSampleL - pear[x]) + pear[x+1])*freq*0.5; pear[x] = inputSampleL = (freq * inputSampleL) + ((1.0-freq) * (pear[x] + pear[x+1])); pear[x+1] = slew; slew = ((inputSampleR - pear[x+2]) + pear[x+3])*freq*0.5; pear[x+2] = inputSampleR = (freq * inputSampleR) + ((1.0-freq) * (pear[x+2] + pear[x+3])); pear[x+3] = slew; } inputSampleL *= wet; inputSampleR *= wet; drySampleL *= dry; drySampleR *= dry; inputSampleL += drySampleL; inputSampleR += drySampleR; //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 Pear::processDoubleReplacing(double **inputs, double **outputs, VstInt32 sampleFrames) { double* in1 = inputs[0]; double* in2 = inputs[1]; double* out1 = outputs[0]; double* out2 = outputs[1]; double overallscale = 1.0; overallscale /= 44100.0; overallscale *= getSampleRate(); int cycleEnd = floor(overallscale); if (cycleEnd < 1) cycleEnd = 1; if (cycleEnd > 3) cycleEnd = 3; cycleEnd--; //this is going to be 0 for 44.1 or 48k, 1 for 88.2 or 96k, 2 for 176 or 192k. //DIFFERENT! Offsetting the bt shift int bitshiftFreq = (10-(A*10.0))+cycleEnd; double freq = 1.0; switch (bitshiftFreq) { case 16: freq = 0.0000152587890625; break; case 15: freq = 0.000030517578125; break; case 14: freq = 0.00006103515625; break; case 13: freq = 0.0001220703125; break; case 12: freq = 0.000244140625; break; case 11: freq = 0.00048828125; break; case 10: freq = 0.0009765625; break; case 9: freq = 0.001953125; break; case 8: freq = 0.00390625; break; case 7: freq = 0.0078125; break; case 6: freq = 0.015625; break; case 5: freq = 0.03125; break; case 4: freq = 0.0625; break; case 3: freq = 0.125; break; case 2: freq = 0.25; break; case 1: freq = 0.5; break; case 0: freq = 1.0; break; } double maxPoles = B*pear_total; double wet = (C*2.0)-1.0; //inv-dry-wet for highpass double dry = 2.0-(C*2.0); if (dry > 1.0) dry = 1.0; //full dry for use with inv, to 0.0 at full wet 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; //Pear is a variation on a Holt filter, made to act like my biquad filters: //its variables (L and R prevSample and prevSlew) contained in arrays for easier handling. //On top of that, the array is used to facilitate use of the filter in a stack which will //sharpen its very weak frequency response performance (3dB/oct per pole) and allow it to run //in a loop. Many earlier Airwindows filters did this 'unrolled', often with variables given //lettered suffixes like iirA through iirZ. The use of an array will make this a lot tidier. for (int x = 0; x < maxPoles; x += 4) { double slew = ((inputSampleL - pear[x]) + pear[x+1])*freq*0.5; pear[x] = inputSampleL = (freq * inputSampleL) + ((1.0-freq) * (pear[x] + pear[x+1])); pear[x+1] = slew; slew = ((inputSampleR - pear[x+2]) + pear[x+3])*freq*0.5; pear[x+2] = inputSampleR = (freq * inputSampleR) + ((1.0-freq) * (pear[x+2] + pear[x+3])); pear[x+3] = slew; } inputSampleL *= wet; inputSampleR *= wet; drySampleL *= dry; drySampleR *= dry; inputSampleL += drySampleL; inputSampleR += drySampleR; //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++; } }