/* ======================================== * Biquad - Biquad.h * Copyright (c) 2016 airwindows, All rights reserved * ======================================== */ #ifndef __Biquad_H #include "Biquad.h" #endif void Biquad::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 type = ceil((A*3.999)+0.00001); biquad[0] = ((B*B*B*0.9999)+0.0001)*0.499; if (biquad[0] < 0.0001) biquad[0] = 0.0001; biquad[1] = (C*C*C*29.99)+0.01; if (biquad[1] < 0.0001) biquad[1] = 0.0001; double wet = (D*2.0)-1.0; //biquad contains these values: //[0] is frequency: 0.000001 to 0.499999 is near-zero to near-Nyquist //[1] is resonance, 0.7071 is Butterworth. Also can't be zero //[2] is a0 but you need distinct ones for additional biquad instances so it's here //[3] is a1 but you need distinct ones for additional biquad instances so it's here //[4] is a2 but you need distinct ones for additional biquad instances so it's here //[5] is b1 but you need distinct ones for additional biquad instances so it's here //[6] is b2 but you need distinct ones for additional biquad instances so it's here //[7] is LEFT stored delayed sample (freq and res are stored so you can move them sample by sample) //[8] is LEFT stored delayed sample (you have to include the coefficient making code if you do that) //[9] is RIGHT stored delayed sample (freq and res are stored so you can move them sample by sample) //[10] is RIGHT stored delayed sample (you have to include the coefficient making code if you do that) //to build a dedicated filter, rename 'biquad' to whatever the new filter is, then //put this code either within the sample buffer (for smoothly modulating freq or res) //or in this 'read the controls' area (for letting you change freq and res with controls) //or in 'reset' if the freq and res are absolutely fixed (use GetSampleRate to define freq) if (type == 1) { //lowpass double K = tan(M_PI * biquad[0]); double norm = 1.0 / (1.0 + K / biquad[1] + K * K); biquad[2] = K * K * norm; biquad[3] = 2.0 * biquad[2]; biquad[4] = biquad[2]; biquad[5] = 2.0 * (K * K - 1.0) * norm; biquad[6] = (1.0 - K / biquad[1] + K * K) * norm; } if (type == 2) { //highpass double K = tan(M_PI * biquad[0]); double norm = 1.0 / (1.0 + K / biquad[1] + K * K); biquad[2] = norm; biquad[3] = -2.0 * biquad[2]; biquad[4] = biquad[2]; biquad[5] = 2.0 * (K * K - 1.0) * norm; biquad[6] = (1.0 - K / biquad[1] + K * K) * norm; } if (type == 3) { //bandpass double K = tan(M_PI * biquad[0]); double norm = 1.0 / (1.0 + K / biquad[1] + K * K); biquad[2] = K / biquad[1] * norm; biquad[3] = 0.0; //bandpass can simplify the biquad kernel: leave out this multiply biquad[4] = -biquad[2]; biquad[5] = 2.0 * (K * K - 1.0) * norm; biquad[6] = (1.0 - K / biquad[1] + K * K) * norm; } if (type == 4) { //notch double K = tan(M_PI * biquad[0]); double norm = 1.0 / (1.0 + K / biquad[1] + K * K); biquad[2] = (1.0 + K * K) * norm; biquad[3] = 2.0 * (K * K - 1) * norm; biquad[4] = biquad[2]; biquad[5] = biquad[3]; biquad[6] = (1.0 - K / biquad[1] + K * K) * norm; } while (--sampleFrames >= 0) { long double inputSampleL = *in1; long double inputSampleR = *in2; if (fabs(inputSampleL)<1.18e-37) inputSampleL = fpd * 1.18e-37; if (fabs(inputSampleR)<1.18e-37) inputSampleR = fpd * 1.18e-37; long double drySampleL = inputSampleL; long double drySampleR = inputSampleR; inputSampleL = sin(inputSampleL); inputSampleR = sin(inputSampleR); //encode Console5: good cleanness /* long double mid = inputSampleL + inputSampleR; long double side = inputSampleL - inputSampleR; //assign mid and side.Between these sections, you can do mid/side processing long double tempSampleM = (mid * biquad[2]) + biquad[7]; biquad[7] = (mid * biquad[3]) - (tempSampleM * biquad[5]) + biquad[8]; biquad[8] = (mid * biquad[4]) - (tempSampleM * biquad[6]); mid = tempSampleM; //like mono AU, 7 and 8 store mid channel long double tempSampleS = (side * biquad[2]) + biquad[9]; biquad[9] = (side * biquad[3]) - (tempSampleS * biquad[5]) + biquad[10]; biquad[10] = (side * biquad[4]) - (tempSampleS * biquad[6]); inputSampleR = tempSampleS; //note: 9 and 10 store the side channel inputSampleL = (mid+side)/2.0; inputSampleR = (mid-side)/2.0; //unassign mid and side */ long double tempSampleL = (inputSampleL * biquad[2]) + biquad[7]; biquad[7] = (inputSampleL * biquad[3]) - (tempSampleL * biquad[5]) + biquad[8]; biquad[8] = (inputSampleL * biquad[4]) - (tempSampleL * biquad[6]); inputSampleL = tempSampleL; //like mono AU, 7 and 8 store L channel long double tempSampleR = (inputSampleR * biquad[2]) + biquad[9]; biquad[9] = (inputSampleR * biquad[3]) - (tempSampleR * biquad[5]) + biquad[10]; biquad[10] = (inputSampleR * biquad[4]) - (tempSampleR * biquad[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); //amplitude aspect if (wet < 1.0) { inputSampleL = (inputSampleL*wet) + (drySampleL*(1.0-fabs(wet))); inputSampleR = (inputSampleR*wet) + (drySampleR*(1.0-fabs(wet))); //inv/dry/wet lets us turn LP into HP and band into notch } //begin 32 bit stereo floating point dither int expon; frexpf((float)inputSampleL, &expon); fpd ^= fpd << 13; fpd ^= fpd >> 17; fpd ^= fpd << 5; inputSampleL += ((double(fpd)-uint32_t(0x7fffffff)) * 5.5e-36l * pow(2,expon+62)); frexpf((float)inputSampleR, &expon); fpd ^= fpd << 13; fpd ^= fpd >> 17; fpd ^= fpd << 5; inputSampleR += ((double(fpd)-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 Biquad::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 type = ceil((A*3.999)+0.00001); biquad[0] = ((B*B*B*0.9999)+0.0001)*0.499; if (biquad[0] < 0.0001) biquad[0] = 0.0001; biquad[1] = (C*C*C*29.99)+0.01; if (biquad[1] < 0.0001) biquad[1] = 0.0001; double wet = (D*2.0)-1.0; //biquad contains these values: //[0] is frequency: 0.000001 to 0.499999 is near-zero to near-Nyquist //[1] is resonance, 0.7071 is Butterworth. Also can't be zero //[2] is a0 but you need distinct ones for additional biquad instances so it's here //[3] is a1 but you need distinct ones for additional biquad instances so it's here //[4] is a2 but you need distinct ones for additional biquad instances so it's here //[5] is b1 but you need distinct ones for additional biquad instances so it's here //[6] is b2 but you need distinct ones for additional biquad instances so it's here //[7] is LEFT stored delayed sample (freq and res are stored so you can move them sample by sample) //[8] is LEFT stored delayed sample (you have to include the coefficient making code if you do that) //[9] is RIGHT stored delayed sample (freq and res are stored so you can move them sample by sample) //[10] is RIGHT stored delayed sample (you have to include the coefficient making code if you do that) //to build a dedicated filter, rename 'biquad' to whatever the new filter is, then //put this code either within the sample buffer (for smoothly modulating freq or res) //or in this 'read the controls' area (for letting you change freq and res with controls) //or in 'reset' if the freq and res are absolutely fixed (use GetSampleRate to define freq) if (type == 1) { //lowpass double K = tan(M_PI * biquad[0]); double norm = 1.0 / (1.0 + K / biquad[1] + K * K); biquad[2] = K * K * norm; biquad[3] = 2.0 * biquad[2]; biquad[4] = biquad[2]; biquad[5] = 2.0 * (K * K - 1.0) * norm; biquad[6] = (1.0 - K / biquad[1] + K * K) * norm; } if (type == 2) { //highpass double K = tan(M_PI * biquad[0]); double norm = 1.0 / (1.0 + K / biquad[1] + K * K); biquad[2] = norm; biquad[3] = -2.0 * biquad[2]; biquad[4] = biquad[2]; biquad[5] = 2.0 * (K * K - 1.0) * norm; biquad[6] = (1.0 - K / biquad[1] + K * K) * norm; } if (type == 3) { //bandpass double K = tan(M_PI * biquad[0]); double norm = 1.0 / (1.0 + K / biquad[1] + K * K); biquad[2] = K / biquad[1] * norm; biquad[3] = 0.0; //bandpass can simplify the biquad kernel: leave out this multiply biquad[4] = -biquad[2]; biquad[5] = 2.0 * (K * K - 1.0) * norm; biquad[6] = (1.0 - K / biquad[1] + K * K) * norm; } if (type == 4) { //notch double K = tan(M_PI * biquad[0]); double norm = 1.0 / (1.0 + K / biquad[1] + K * K); biquad[2] = (1.0 + K * K) * norm; biquad[3] = 2.0 * (K * K - 1) * norm; biquad[4] = biquad[2]; biquad[5] = biquad[3]; biquad[6] = (1.0 - K / biquad[1] + K * K) * norm; } while (--sampleFrames >= 0) { long double inputSampleL = *in1; long double inputSampleR = *in2; if (fabs(inputSampleL)<1.18e-43) inputSampleL = fpd * 1.18e-43; if (fabs(inputSampleR)<1.18e-43) inputSampleR = fpd * 1.18e-43; long double drySampleL = inputSampleL; long double drySampleR = inputSampleR; inputSampleL = sin(inputSampleL); inputSampleR = sin(inputSampleR); //encode Console5: good cleanness /* long double mid = inputSampleL + inputSampleR; long double side = inputSampleL - inputSampleR; //assign mid and side.Between these sections, you can do mid/side processing long double tempSampleM = (mid * biquad[2]) + biquad[7]; biquad[7] = (mid * biquad[3]) - (tempSampleM * biquad[5]) + biquad[8]; biquad[8] = (mid * biquad[4]) - (tempSampleM * biquad[6]); mid = tempSampleM; //like mono AU, 7 and 8 store mid channel long double tempSampleS = (side * biquad[2]) + biquad[9]; biquad[9] = (side * biquad[3]) - (tempSampleS * biquad[5]) + biquad[10]; biquad[10] = (side * biquad[4]) - (tempSampleS * biquad[6]); inputSampleR = tempSampleS; //note: 9 and 10 store the side channel inputSampleL = (mid+side)/2.0; inputSampleR = (mid-side)/2.0; //unassign mid and side */ long double tempSampleL = (inputSampleL * biquad[2]) + biquad[7]; biquad[7] = (inputSampleL * biquad[3]) - (tempSampleL * biquad[5]) + biquad[8]; biquad[8] = (inputSampleL * biquad[4]) - (tempSampleL * biquad[6]); inputSampleL = tempSampleL; //like mono AU, 7 and 8 store L channel long double tempSampleR = (inputSampleR * biquad[2]) + biquad[9]; biquad[9] = (inputSampleR * biquad[3]) - (tempSampleR * biquad[5]) + biquad[10]; biquad[10] = (inputSampleR * biquad[4]) - (tempSampleR * biquad[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); //amplitude aspect if (wet < 1.0) { inputSampleL = (inputSampleL*wet) + (drySampleL*(1.0-fabs(wet))); inputSampleR = (inputSampleR*wet) + (drySampleR*(1.0-fabs(wet))); //inv/dry/wet lets us turn LP into HP and band into notch } //begin 64 bit stereo floating point dither int expon; frexp((double)inputSampleL, &expon); fpd ^= fpd << 13; fpd ^= fpd >> 17; fpd ^= fpd << 5; inputSampleL += ((double(fpd)-uint32_t(0x7fffffff)) * 1.1e-44l * pow(2,expon+62)); frexp((double)inputSampleR, &expon); fpd ^= fpd << 13; fpd ^= fpd >> 17; fpd ^= fpd << 5; inputSampleR += ((double(fpd)-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++; } }