X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/e8482f24cfc00645607f8526fde38d39e0dcaa63..4430ec36d1707fa18b28590bceb5d5d7cacbc337:/contrib/src/mmedia/g723_24.cpp diff --git a/contrib/src/mmedia/g723_24.cpp b/contrib/src/mmedia/g723_24.cpp index fe581c3de8..6e31578df1 100644 --- a/contrib/src/mmedia/g723_24.cpp +++ b/contrib/src/mmedia/g723_24.cpp @@ -37,24 +37,24 @@ * of workstation attributes, such as hardware 2's complement arithmetic. * */ -#include +#include "wx/wxprec.h" #include "wx/mmedia/internal/g72x.h" /* * Maps G.723_24 code word to reconstructed scale factor normalized log * magnitude values. */ -static short _dqlntab[8] = {-2048, 135, 273, 373, 373, 273, 135, -2048}; +static short _dqlntab[8] = {-2048, 135, 273, 373, 373, 273, 135, -2048}; /* Maps G.723_24 code word to log of scale factor multiplier. */ -static short _witab[8] = {-128, 960, 4384, 18624, 18624, 4384, 960, -128}; +static short _witab[8] = {-128, 960, 4384, 18624, 18624, 4384, 960, -128}; /* * Maps G.723_24 code words to a set of values whose long and short * term averages are computed and then compared to give an indication * how stationary (steady state) the signal is. */ -static short _fitab[8] = {0, 0x200, 0x400, 0xE00, 0xE00, 0x400, 0x200, 0}; +static short _fitab[8] = {0, 0x200, 0x400, 0xE00, 0xE00, 0x400, 0x200, 0}; static short qtab_723_24[3] = {8, 218, 331}; @@ -66,50 +66,50 @@ static short qtab_723_24[3] = {8, 218, 331}; */ int g723_24_encoder( - int sl, - int in_coding, - struct g72x_state *state_ptr) + int sl, + int in_coding, + struct g72x_state *state_ptr) { - short sei, sezi, se, sez; /* ACCUM */ - short d; /* SUBTA */ - short y; /* MIX */ - short sr; /* ADDB */ - short dqsez; /* ADDC */ - short dq, i; - - switch (in_coding) { /* linearize input sample to 14-bit PCM */ - case AUDIO_ENCODING_ALAW: - sl = alaw2linear(sl) >> 2; - break; - case AUDIO_ENCODING_ULAW: - sl = ulaw2linear(sl) >> 2; - break; - case AUDIO_ENCODING_LINEAR: - sl = ((short)sl) >> 2; /* sl of 14-bit dynamic range */ - break; - default: - return (-1); - } - - sezi = predictor_zero(state_ptr); - sez = sezi >> 1; - sei = sezi + predictor_pole(state_ptr); - se = sei >> 1; /* se = estimated signal */ - - d = sl - se; /* d = estimation diff. */ - - /* quantize prediction difference d */ - y = step_size(state_ptr); /* quantizer step size */ - i = quantize(d, y, qtab_723_24, 3); /* i = ADPCM code */ - dq = reconstruct(i & 4, _dqlntab[i], y); /* quantized diff. */ - - sr = (dq < 0) ? se - (dq & 0x3FFF) : se + dq; /* reconstructed signal */ - - dqsez = sr + sez - se; /* pole prediction diff. */ - - update(3, y, _witab[i], _fitab[i], dq, sr, dqsez, state_ptr); - - return (i); + short sei, sezi, se, sez; /* ACCUM */ + short d; /* SUBTA */ + short y; /* MIX */ + short sr; /* ADDB */ + short dqsez; /* ADDC */ + short dq, i; + + switch (in_coding) { /* linearize input sample to 14-bit PCM */ + case AUDIO_ENCODING_ALAW: + sl = alaw2linear(sl) >> 2; + break; + case AUDIO_ENCODING_ULAW: + sl = ulaw2linear(sl) >> 2; + break; + case AUDIO_ENCODING_LINEAR: + sl = ((short)sl) >> 2; /* sl of 14-bit dynamic range */ + break; + default: + return (-1); + } + + sezi = predictor_zero(state_ptr); + sez = sezi >> 1; + sei = sezi + predictor_pole(state_ptr); + se = sei >> 1; /* se = estimated signal */ + + d = sl - se; /* d = estimation diff. */ + + /* quantize prediction difference d */ + y = step_size(state_ptr); /* quantizer step size */ + i = quantize(d, y, qtab_723_24, 3); /* i = ADPCM code */ + dq = reconstruct(i & 4, _dqlntab[i], y); /* quantized diff. */ + + sr = (dq < 0) ? se - (dq & 0x3FFF) : se + dq; /* reconstructed signal */ + + dqsez = sr + sez - se; /* pole prediction diff. */ + + update(3, y, _witab[i], _fitab[i], dq, sr, dqsez, state_ptr); + + return (i); } /* @@ -121,39 +121,39 @@ g723_24_encoder( */ int g723_24_decoder( - int i, - int out_coding, - struct g72x_state *state_ptr) + int i, + int out_coding, + struct g72x_state *state_ptr) { - short sezi, sei, sez, se; /* ACCUM */ - short y; /* MIX */ - short sr; /* ADDB */ - short dq; - short dqsez; - - i &= 0x07; /* mask to get proper bits */ - sezi = predictor_zero(state_ptr); - sez = sezi >> 1; - sei = sezi + predictor_pole(state_ptr); - se = sei >> 1; /* se = estimated signal */ - - y = step_size(state_ptr); /* adaptive quantizer step size */ - dq = reconstruct(i & 0x04, _dqlntab[i], y); /* unquantize pred diff */ - - sr = (dq < 0) ? (se - (dq & 0x3FFF)) : (se + dq); /* reconst. signal */ - - dqsez = sr - se + sez; /* pole prediction diff. */ - - update(3, y, _witab[i], _fitab[i], dq, sr, dqsez, state_ptr); - - switch (out_coding) { - case AUDIO_ENCODING_ALAW: - return (tandem_adjust_alaw(sr, se, y, i, 4, qtab_723_24)); - case AUDIO_ENCODING_ULAW: - return (tandem_adjust_ulaw(sr, se, y, i, 4, qtab_723_24)); - case AUDIO_ENCODING_LINEAR: - return (sr << 2); /* sr was of 14-bit dynamic range */ - default: - return (-1); - } + short sezi, sei, sez, se; /* ACCUM */ + short y; /* MIX */ + short sr; /* ADDB */ + short dq; + short dqsez; + + i &= 0x07; /* mask to get proper bits */ + sezi = predictor_zero(state_ptr); + sez = sezi >> 1; + sei = sezi + predictor_pole(state_ptr); + se = sei >> 1; /* se = estimated signal */ + + y = step_size(state_ptr); /* adaptive quantizer step size */ + dq = reconstruct(i & 0x04, _dqlntab[i], y); /* unquantize pred diff */ + + sr = (dq < 0) ? (se - (dq & 0x3FFF)) : (se + dq); /* reconst. signal */ + + dqsez = sr - se + sez; /* pole prediction diff. */ + + update(3, y, _witab[i], _fitab[i], dq, sr, dqsez, state_ptr); + + switch (out_coding) { + case AUDIO_ENCODING_ALAW: + return (tandem_adjust_alaw(sr, se, y, i, 4, qtab_723_24)); + case AUDIO_ENCODING_ULAW: + return (tandem_adjust_ulaw(sr, se, y, i, 4, qtab_723_24)); + case AUDIO_ENCODING_LINEAR: + return (sr << 2); /* sr was of 14-bit dynamic range */ + default: + return (-1); + } }