ref: 4beea686fd27e6c6ba19ec52ea8d80c2a7cc56eb
dir: /src/pt2_blep.c/
// These BLEP routines were coded by aciddose/adejr #include <stdint.h> #include "pt2_blep.h" #include "pt2_helpers.h" /* Why this table is not represented as readable floating-point numbers: ** Accurate float (double) representation in string format requires at least 14 digits and normalized ** (scientific) notation, notwithstanding compiler issues with precision or rounding error. ** Also, don't touch this table ever, just keep it exactly identical! */ // TODO: get a proper double-precision table. This one is converted from float. static const uint64_t dBlepData[48] = { 0x3FEFFC3E20000000, 0x3FEFFAA900000000, 0x3FEFFAD460000000, 0x3FEFFA9C60000000, 0x3FEFF5B0A0000000, 0x3FEFE42A40000000, 0x3FEFB7F5C0000000, 0x3FEF599BE0000000, 0x3FEEA5E3C0000000, 0x3FED6E7080000000, 0x3FEB7F7960000000, 0x3FE8AB9E40000000, 0x3FE4DCA480000000, 0x3FE0251880000000, 0x3FD598FB80000000, 0x3FC53D0D60000000, 0x3F8383A520000000, 0xBFBC977CC0000000, 0xBFC755C080000000, 0xBFC91BDBA0000000, 0xBFC455AFC0000000, 0xBFB6461340000000, 0xBF7056C400000000, 0x3FB1028220000000, 0x3FBB5B7E60000000, 0x3FBC5903A0000000, 0x3FB55403E0000000, 0x3FA3CED340000000, 0xBF7822DAE0000000, 0xBFA2805D00000000, 0xBFA7140D20000000, 0xBFA18A7760000000, 0xBF87FF7180000000, 0x3F88CBFA40000000, 0x3F9D4AEC80000000, 0x3FA14A3AC0000000, 0x3F9D5C5AA0000000, 0x3F92558B40000000, 0x3F7C997EE0000000, 0x0000000000000000, 0x0000000000000000, 0x0000000000000000, 0x0000000000000000, 0x0000000000000000, 0x0000000000000000, 0x0000000000000000, 0x0000000000000000, 0x0000000000000000 }; void blepAdd(blep_t *b, double dOffset, double dAmplitude) { int8_t n; int32_t i; const double *dBlepSrc; double f; assert(dOffset >= 0.0 && dOffset < 1.0); f = dOffset * BLEP_SP; i = (int32_t)f; // get integer part of f dBlepSrc = (const double *)dBlepData + i + BLEP_OS; f -= i; // remove integer part from f i = b->index; n = BLEP_NS; while (n--) { b->dBuffer[i] += dAmplitude * LERP(dBlepSrc[0], dBlepSrc[1], f); i = (i + 1) & BLEP_RNS; dBlepSrc += BLEP_SP; } b->samplesLeft = BLEP_NS; } /* 8bitbubsy: simplified, faster version of blepAdd for blep'ing voice volume. ** Result is identical! (confirmed with binary comparison) */ void blepVolAdd(blep_t *b, double dAmplitude) { int8_t n; int32_t i; const double *dBlepSrc; dBlepSrc = (const double *)dBlepData + BLEP_OS; i = b->index; n = BLEP_NS; while (n--) { b->dBuffer[i] += dAmplitude * (*dBlepSrc); i = (i + 1) & BLEP_RNS; dBlepSrc += BLEP_SP; } b->samplesLeft = BLEP_NS; } double blepRun(blep_t *b) { double fBlepOutput; fBlepOutput = b->dBuffer[b->index]; b->dBuffer[b->index] = 0.0; b->index = (b->index + 1) & BLEP_RNS; b->samplesLeft--; return fBlepOutput; }