You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 
 
 

1109 lines
40 KiB

/*
* Copyright (c) 2001-2003 The ffmpeg Project
*
* This file is part of Libav.
*
* Libav is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* Libav is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with Libav; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "avcodec.h"
#include "get_bits.h"
#include "put_bits.h"
#include "bytestream.h"
#include "adpcm.h"
#include "adpcm_data.h"
/**
* @file
* ADPCM decoders
* First version by Francois Revol (revol@free.fr)
* Fringe ADPCM codecs (e.g., DK3, DK4, Westwood)
* by Mike Melanson (melanson@pcisys.net)
* CD-ROM XA ADPCM codec by BERO
* EA ADPCM decoder by Robin Kay (komadori@myrealbox.com)
* EA ADPCM R1/R2/R3 decoder by Peter Ross (pross@xvid.org)
* EA IMA EACS decoder by Peter Ross (pross@xvid.org)
* EA IMA SEAD decoder by Peter Ross (pross@xvid.org)
* EA ADPCM XAS decoder by Peter Ross (pross@xvid.org)
* MAXIS EA ADPCM decoder by Robert Marston (rmarston@gmail.com)
* THP ADPCM decoder by Marco Gerards (mgerards@xs4all.nl)
*
* Features and limitations:
*
* Reference documents:
* http://wiki.multimedia.cx/index.php?title=Category:ADPCM_Audio_Codecs
* http://www.pcisys.net/~melanson/codecs/simpleaudio.html [dead]
* http://www.geocities.com/SiliconValley/8682/aud3.txt [dead]
* http://openquicktime.sourceforge.net/
* XAnim sources (xa_codec.c) http://xanim.polter.net/
* http://www.cs.ucla.edu/~leec/mediabench/applications.html [dead]
* SoX source code http://sox.sourceforge.net/
*
* CD-ROM XA:
* http://ku-www.ss.titech.ac.jp/~yatsushi/xaadpcm.html [dead]
* vagpack & depack http://homepages.compuserve.de/bITmASTER32/psx-index.html [dead]
* readstr http://www.geocities.co.jp/Playtown/2004/
*/
/* These are for CD-ROM XA ADPCM */
static const int xa_adpcm_table[5][2] = {
{ 0, 0 },
{ 60, 0 },
{ 115, -52 },
{ 98, -55 },
{ 122, -60 }
};
static const int ea_adpcm_table[] = {
0, 240, 460, 392,
0, 0, -208, -220,
0, 1, 3, 4,
7, 8, 10, 11,
0, -1, -3, -4
};
// padded to zero where table size is less then 16
static const int swf_index_tables[4][16] = {
/*2*/ { -1, 2 },
/*3*/ { -1, -1, 2, 4 },
/*4*/ { -1, -1, -1, -1, 2, 4, 6, 8 },
/*5*/ { -1, -1, -1, -1, -1, -1, -1, -1, 1, 2, 4, 6, 8, 10, 13, 16 }
};
/* end of tables */
typedef struct ADPCMDecodeContext {
ADPCMChannelStatus status[6];
} ADPCMDecodeContext;
static av_cold int adpcm_decode_init(AVCodecContext * avctx)
{
ADPCMDecodeContext *c = avctx->priv_data;
unsigned int max_channels = 2;
switch(avctx->codec->id) {
case CODEC_ID_ADPCM_EA_R1:
case CODEC_ID_ADPCM_EA_R2:
case CODEC_ID_ADPCM_EA_R3:
case CODEC_ID_ADPCM_EA_XAS:
max_channels = 6;
break;
}
if(avctx->channels > max_channels){
return -1;
}
switch(avctx->codec->id) {
case CODEC_ID_ADPCM_CT:
c->status[0].step = c->status[1].step = 511;
break;
case CODEC_ID_ADPCM_IMA_WAV:
if (avctx->bits_per_coded_sample != 4) {
av_log(avctx, AV_LOG_ERROR, "Only 4-bit ADPCM IMA WAV files are supported\n");
return -1;
}
break;
case CODEC_ID_ADPCM_IMA_WS:
if (avctx->extradata && avctx->extradata_size == 2 * 4) {
c->status[0].predictor = AV_RL32(avctx->extradata);
c->status[1].predictor = AV_RL32(avctx->extradata + 4);
}
break;
default:
break;
}
avctx->sample_fmt = AV_SAMPLE_FMT_S16;
return 0;
}
static inline short adpcm_ima_expand_nibble(ADPCMChannelStatus *c, char nibble, int shift)
{
int step_index;
int predictor;
int sign, delta, diff, step;
step = ff_adpcm_step_table[c->step_index];
step_index = c->step_index + ff_adpcm_index_table[(unsigned)nibble];
if (step_index < 0) step_index = 0;
else if (step_index > 88) step_index = 88;
sign = nibble & 8;
delta = nibble & 7;
/* perform direct multiplication instead of series of jumps proposed by
* the reference ADPCM implementation since modern CPUs can do the mults
* quickly enough */
diff = ((2 * delta + 1) * step) >> shift;
predictor = c->predictor;
if (sign) predictor -= diff;
else predictor += diff;
c->predictor = av_clip_int16(predictor);
c->step_index = step_index;
return (short)c->predictor;
}
static inline int adpcm_ima_qt_expand_nibble(ADPCMChannelStatus *c, int nibble, int shift)
{
int step_index;
int predictor;
int diff, step;
step = ff_adpcm_step_table[c->step_index];
step_index = c->step_index + ff_adpcm_index_table[nibble];
step_index = av_clip(step_index, 0, 88);
diff = step >> 3;
if (nibble & 4) diff += step;
if (nibble & 2) diff += step >> 1;
if (nibble & 1) diff += step >> 2;
if (nibble & 8)
predictor = c->predictor - diff;
else
predictor = c->predictor + diff;
c->predictor = av_clip_int16(predictor);
c->step_index = step_index;
return c->predictor;
}
static inline short adpcm_ms_expand_nibble(ADPCMChannelStatus *c, char nibble)
{
int predictor;
predictor = (((c->sample1) * (c->coeff1)) + ((c->sample2) * (c->coeff2))) / 64;
predictor += (signed)((nibble & 0x08)?(nibble - 0x10):(nibble)) * c->idelta;
c->sample2 = c->sample1;
c->sample1 = av_clip_int16(predictor);
c->idelta = (ff_adpcm_AdaptationTable[(int)nibble] * c->idelta) >> 8;
if (c->idelta < 16) c->idelta = 16;
return c->sample1;
}
static inline short adpcm_ct_expand_nibble(ADPCMChannelStatus *c, char nibble)
{
int sign, delta, diff;
int new_step;
sign = nibble & 8;
delta = nibble & 7;
/* perform direct multiplication instead of series of jumps proposed by
* the reference ADPCM implementation since modern CPUs can do the mults
* quickly enough */
diff = ((2 * delta + 1) * c->step) >> 3;
/* predictor update is not so trivial: predictor is multiplied on 254/256 before updating */
c->predictor = ((c->predictor * 254) >> 8) + (sign ? -diff : diff);
c->predictor = av_clip_int16(c->predictor);
/* calculate new step and clamp it to range 511..32767 */
new_step = (ff_adpcm_AdaptationTable[nibble & 7] * c->step) >> 8;
c->step = av_clip(new_step, 511, 32767);
return (short)c->predictor;
}
static inline short adpcm_sbpro_expand_nibble(ADPCMChannelStatus *c, char nibble, int size, int shift)
{
int sign, delta, diff;
sign = nibble & (1<<(size-1));
delta = nibble & ((1<<(size-1))-1);
diff = delta << (7 + c->step + shift);
/* clamp result */
c->predictor = av_clip(c->predictor + (sign ? -diff : diff), -16384,16256);
/* calculate new step */
if (delta >= (2*size - 3) && c->step < 3)
c->step++;
else if (delta == 0 && c->step > 0)
c->step--;
return (short) c->predictor;
}
static inline short adpcm_yamaha_expand_nibble(ADPCMChannelStatus *c, unsigned char nibble)
{
if(!c->step) {
c->predictor = 0;
c->step = 127;
}
c->predictor += (c->step * ff_adpcm_yamaha_difflookup[nibble]) / 8;
c->predictor = av_clip_int16(c->predictor);
c->step = (c->step * ff_adpcm_yamaha_indexscale[nibble]) >> 8;
c->step = av_clip(c->step, 127, 24567);
return c->predictor;
}
static void xa_decode(short *out, const unsigned char *in,
ADPCMChannelStatus *left, ADPCMChannelStatus *right, int inc)
{
int i, j;
int shift,filter,f0,f1;
int s_1,s_2;
int d,s,t;
for(i=0;i<4;i++) {
shift = 12 - (in[4+i*2] & 15);
filter = in[4+i*2] >> 4;
f0 = xa_adpcm_table[filter][0];
f1 = xa_adpcm_table[filter][1];
s_1 = left->sample1;
s_2 = left->sample2;
for(j=0;j<28;j++) {
d = in[16+i+j*4];
t = (signed char)(d<<4)>>4;
s = ( t<<shift ) + ((s_1*f0 + s_2*f1+32)>>6);
s_2 = s_1;
s_1 = av_clip_int16(s);
*out = s_1;
out += inc;
}
if (inc==2) { /* stereo */
left->sample1 = s_1;
left->sample2 = s_2;
s_1 = right->sample1;
s_2 = right->sample2;
out = out + 1 - 28*2;
}
shift = 12 - (in[5+i*2] & 15);
filter = in[5+i*2] >> 4;
f0 = xa_adpcm_table[filter][0];
f1 = xa_adpcm_table[filter][1];
for(j=0;j<28;j++) {
d = in[16+i+j*4];
t = (signed char)d >> 4;
s = ( t<<shift ) + ((s_1*f0 + s_2*f1+32)>>6);
s_2 = s_1;
s_1 = av_clip_int16(s);
*out = s_1;
out += inc;
}
if (inc==2) { /* stereo */
right->sample1 = s_1;
right->sample2 = s_2;
out -= 1;
} else {
left->sample1 = s_1;
left->sample2 = s_2;
}
}
}
/* DK3 ADPCM support macro */
#define DK3_GET_NEXT_NIBBLE() \
if (decode_top_nibble_next) \
{ \
nibble = last_byte >> 4; \
decode_top_nibble_next = 0; \
} \
else \
{ \
last_byte = *src++; \
if (src >= buf + buf_size) break; \
nibble = last_byte & 0x0F; \
decode_top_nibble_next = 1; \
}
static int adpcm_decode_frame(AVCodecContext *avctx,
void *data, int *data_size,
AVPacket *avpkt)
{
const uint8_t *buf = avpkt->data;
int buf_size = avpkt->size;
ADPCMDecodeContext *c = avctx->priv_data;
ADPCMChannelStatus *cs;
int n, m, channel, i;
short *samples;
short *samples_end;
const uint8_t *src;
int st; /* stereo */
uint32_t samples_in_chunk;
int count1, count2;
if (!buf_size)
return 0;
//should protect all 4bit ADPCM variants
//8 is needed for CODEC_ID_ADPCM_IMA_WAV with 2 channels
//
if(*data_size/4 < buf_size + 8)
return -1;
samples = data;
samples_end= samples + *data_size/2;
*data_size= 0;
src = buf;
st = avctx->channels == 2 ? 1 : 0;
switch(avctx->codec->id) {
case CODEC_ID_ADPCM_IMA_QT:
/* In QuickTime, IMA is encoded by chunks of 34 bytes (=64 samples).
Channel data is interleaved per-chunk. */
if (buf_size / 34 < avctx->channels) {
av_log(avctx, AV_LOG_ERROR, "packet is too small\n");
return AVERROR(EINVAL);
}
for (channel = 0; channel < avctx->channels; channel++) {
int16_t predictor;
int step_index;
cs = &(c->status[channel]);
/* (pppppp) (piiiiiii) */
/* Bits 15-7 are the _top_ 9 bits of the 16-bit initial predictor value */
predictor = AV_RB16(src);
step_index = predictor & 0x7F;
predictor &= 0xFF80;
src += 2;
if (cs->step_index == step_index) {
int diff = (int)predictor - cs->predictor;
if (diff < 0)
diff = - diff;
if (diff > 0x7f)
goto update;
} else {
update:
cs->step_index = step_index;
cs->predictor = predictor;
}
if (cs->step_index > 88){
av_log(avctx, AV_LOG_ERROR, "ERROR: step_index = %i\n", cs->step_index);
cs->step_index = 88;
}
samples = (short*)data + channel;
for (m = 0; m < 32; m++) {
*samples = adpcm_ima_qt_expand_nibble(cs, src[0] & 0x0F, 3);
samples += avctx->channels;
*samples = adpcm_ima_qt_expand_nibble(cs, src[0] >> 4 , 3);
samples += avctx->channels;
src ++;
}
}
if (st)
samples--;
break;
case CODEC_ID_ADPCM_IMA_WAV:
if (avctx->block_align != 0 && buf_size > avctx->block_align)
buf_size = avctx->block_align;
// samples_per_block= (block_align-4*chanels)*8 / (bits_per_sample * chanels) + 1;
for(i=0; i<avctx->channels; i++){
cs = &(c->status[i]);
cs->predictor = *samples++ = (int16_t)bytestream_get_le16(&src);
cs->step_index = *src++;
if (cs->step_index > 88){
av_log(avctx, AV_LOG_ERROR, "ERROR: step_index = %i\n", cs->step_index);
cs->step_index = 88;
}
if (*src++) av_log(avctx, AV_LOG_ERROR, "unused byte should be null but is %d!!\n", src[-1]); /* unused */
}
while(src < buf + buf_size){
for (i = 0; i < avctx->channels; i++) {
cs = &c->status[i];
for (m = 0; m < 4; m++) {
uint8_t v = *src++;
*samples = adpcm_ima_expand_nibble(cs, v & 0x0F, 3);
samples += avctx->channels;
*samples = adpcm_ima_expand_nibble(cs, v >> 4 , 3);
samples += avctx->channels;
}
samples -= 8 * avctx->channels - 1;
}
samples += 7 * avctx->channels;
}
break;
case CODEC_ID_ADPCM_4XM:
for (i = 0; i < avctx->channels; i++)
c->status[i].predictor= (int16_t)bytestream_get_le16(&src);
for (i = 0; i < avctx->channels; i++) {
c->status[i].step_index= (int16_t)bytestream_get_le16(&src);
c->status[i].step_index = av_clip(c->status[i].step_index, 0, 88);
}
m= (buf_size - (src - buf))>>st;
for (i = 0; i < avctx->channels; i++) {
samples = (short*)data + i;
cs = &c->status[i];
for (n = 0; n < m; n++) {
uint8_t v = *src++;
*samples = adpcm_ima_expand_nibble(cs, v & 0x0F, 4);
samples += avctx->channels;
*samples = adpcm_ima_expand_nibble(cs, v >> 4 , 4);
samples += avctx->channels;
}
}
samples -= (avctx->channels - 1);
break;
case CODEC_ID_ADPCM_MS:
{
int block_predictor;
if (avctx->block_align != 0 && buf_size > avctx->block_align)
buf_size = avctx->block_align;
n = buf_size - 7 * avctx->channels;
if (n < 0)
return -1;
block_predictor = av_clip(*src++, 0, 6);
c->status[0].coeff1 = ff_adpcm_AdaptCoeff1[block_predictor];
c->status[0].coeff2 = ff_adpcm_AdaptCoeff2[block_predictor];
if (st) {
block_predictor = av_clip(*src++, 0, 6);
c->status[1].coeff1 = ff_adpcm_AdaptCoeff1[block_predictor];
c->status[1].coeff2 = ff_adpcm_AdaptCoeff2[block_predictor];
}
c->status[0].idelta = (int16_t)bytestream_get_le16(&src);
if (st){
c->status[1].idelta = (int16_t)bytestream_get_le16(&src);
}
c->status[0].sample1 = bytestream_get_le16(&src);
if (st) c->status[1].sample1 = bytestream_get_le16(&src);
c->status[0].sample2 = bytestream_get_le16(&src);
if (st) c->status[1].sample2 = bytestream_get_le16(&src);
*samples++ = c->status[0].sample2;
if (st) *samples++ = c->status[1].sample2;
*samples++ = c->status[0].sample1;
if (st) *samples++ = c->status[1].sample1;
for(;n>0;n--) {
*samples++ = adpcm_ms_expand_nibble(&c->status[0 ], src[0] >> 4 );
*samples++ = adpcm_ms_expand_nibble(&c->status[st], src[0] & 0x0F);
src ++;
}
break;
}
case CODEC_ID_ADPCM_IMA_DK4:
if (avctx->block_align != 0 && buf_size > avctx->block_align)
buf_size = avctx->block_align;
n = buf_size - 4 * avctx->channels;
if (n < 0) {
av_log(avctx, AV_LOG_ERROR, "packet is too small\n");
return AVERROR(EINVAL);
}
for (channel = 0; channel < avctx->channels; channel++) {
cs = &c->status[channel];
cs->predictor = (int16_t)bytestream_get_le16(&src);
cs->step_index = *src++;
src++;
*samples++ = cs->predictor;
}
while (n-- > 0) {
uint8_t v = *src++;
*samples++ = adpcm_ima_expand_nibble(&c->status[0 ], v >> 4 , 3);
*samples++ = adpcm_ima_expand_nibble(&c->status[st], v & 0x0F, 3);
}
break;
case CODEC_ID_ADPCM_IMA_DK3:
{
unsigned char last_byte = 0;
unsigned char nibble;
int decode_top_nibble_next = 0;
int diff_channel;
if (avctx->block_align != 0 && buf_size > avctx->block_align)
buf_size = avctx->block_align;
if(buf_size + 16 > (samples_end - samples)*3/8)
return -1;
c->status[0].predictor = (int16_t)AV_RL16(src + 10);
c->status[1].predictor = (int16_t)AV_RL16(src + 12);
c->status[0].step_index = src[14];
c->status[1].step_index = src[15];
/* sign extend the predictors */
src += 16;
diff_channel = c->status[1].predictor;
/* the DK3_GET_NEXT_NIBBLE macro issues the break statement when
* the buffer is consumed */
while (1) {
/* for this algorithm, c->status[0] is the sum channel and
* c->status[1] is the diff channel */
/* process the first predictor of the sum channel */
DK3_GET_NEXT_NIBBLE();
adpcm_ima_expand_nibble(&c->status[0], nibble, 3);
/* process the diff channel predictor */
DK3_GET_NEXT_NIBBLE();
adpcm_ima_expand_nibble(&c->status[1], nibble, 3);
/* process the first pair of stereo PCM samples */
diff_channel = (diff_channel + c->status[1].predictor) / 2;
*samples++ = c->status[0].predictor + c->status[1].predictor;
*samples++ = c->status[0].predictor - c->status[1].predictor;
/* process the second predictor of the sum channel */
DK3_GET_NEXT_NIBBLE();
adpcm_ima_expand_nibble(&c->status[0], nibble, 3);
/* process the second pair of stereo PCM samples */
diff_channel = (diff_channel + c->status[1].predictor) / 2;
*samples++ = c->status[0].predictor + c->status[1].predictor;
*samples++ = c->status[0].predictor - c->status[1].predictor;
}
break;
}
case CODEC_ID_ADPCM_IMA_ISS:
n = buf_size - 4 * avctx->channels;
if (n < 0) {
av_log(avctx, AV_LOG_ERROR, "packet is too small\n");
return AVERROR(EINVAL);
}
for (channel = 0; channel < avctx->channels; channel++) {
cs = &c->status[channel];
cs->predictor = (int16_t)bytestream_get_le16(&src);
cs->step_index = *src++;
src++;
}
while (n-- > 0) {
uint8_t v1, v2;
uint8_t v = *src++;
/* nibbles are swapped for mono */
if (st) {
v1 = v >> 4;
v2 = v & 0x0F;
} else {
v2 = v >> 4;
v1 = v & 0x0F;
}
*samples++ = adpcm_ima_expand_nibble(&c->status[0 ], v1, 3);
*samples++ = adpcm_ima_expand_nibble(&c->status[st], v2, 3);
}
break;
case CODEC_ID_ADPCM_IMA_WS:
while (src < buf + buf_size) {
uint8_t v = *src++;
*samples++ = adpcm_ima_expand_nibble(&c->status[0], v >> 4 , 3);
*samples++ = adpcm_ima_expand_nibble(&c->status[st], v & 0x0F, 3);
}
break;
case CODEC_ID_ADPCM_XA:
while (buf_size >= 128) {
xa_decode(samples, src, &c->status[0], &c->status[1],
avctx->channels);
src += 128;
samples += 28 * 8;
buf_size -= 128;
}
break;
case CODEC_ID_ADPCM_IMA_EA_EACS:
samples_in_chunk = bytestream_get_le32(&src) >> (1-st);
if (samples_in_chunk > buf_size-4-(8<<st)) {
src += buf_size - 4;
break;
}
for (i=0; i<=st; i++)
c->status[i].step_index = bytestream_get_le32(&src);
for (i=0; i<=st; i++)
c->status[i].predictor = bytestream_get_le32(&src);
for (; samples_in_chunk; samples_in_chunk--, src++) {
*samples++ = adpcm_ima_expand_nibble(&c->status[0], *src>>4, 3);
*samples++ = adpcm_ima_expand_nibble(&c->status[st], *src&0x0F, 3);
}
break;
case CODEC_ID_ADPCM_IMA_EA_SEAD:
for (; src < buf+buf_size; src++) {
*samples++ = adpcm_ima_expand_nibble(&c->status[0], src[0] >> 4, 6);
*samples++ = adpcm_ima_expand_nibble(&c->status[st],src[0]&0x0F, 6);
}
break;
case CODEC_ID_ADPCM_EA:
{
int32_t previous_left_sample, previous_right_sample;
int32_t current_left_sample, current_right_sample;
int32_t next_left_sample, next_right_sample;
int32_t coeff1l, coeff2l, coeff1r, coeff2r;
uint8_t shift_left, shift_right;
/* Each EA ADPCM frame has a 12-byte header followed by 30-byte pieces,
each coding 28 stereo samples. */
if (buf_size < 12) {
av_log(avctx, AV_LOG_ERROR, "frame too small\n");
return AVERROR(EINVAL);
}
samples_in_chunk = AV_RL32(src);
if (samples_in_chunk / 28 > (buf_size - 12) / 30) {
av_log(avctx, AV_LOG_ERROR, "invalid frame\n");
return AVERROR(EINVAL);
}
src += 4;
current_left_sample = (int16_t)bytestream_get_le16(&src);
previous_left_sample = (int16_t)bytestream_get_le16(&src);
current_right_sample = (int16_t)bytestream_get_le16(&src);
previous_right_sample = (int16_t)bytestream_get_le16(&src);
for (count1 = 0; count1 < samples_in_chunk/28;count1++) {
coeff1l = ea_adpcm_table[ *src >> 4 ];
coeff2l = ea_adpcm_table[(*src >> 4 ) + 4];
coeff1r = ea_adpcm_table[*src & 0x0F];
coeff2r = ea_adpcm_table[(*src & 0x0F) + 4];
src++;
shift_left = (*src >> 4 ) + 8;
shift_right = (*src & 0x0F) + 8;
src++;
for (count2 = 0; count2 < 28; count2++) {
next_left_sample = (int32_t)((*src & 0xF0) << 24) >> shift_left;
next_right_sample = (int32_t)((*src & 0x0F) << 28) >> shift_right;
src++;
next_left_sample = (next_left_sample +
(current_left_sample * coeff1l) +
(previous_left_sample * coeff2l) + 0x80) >> 8;
next_right_sample = (next_right_sample +
(current_right_sample * coeff1r) +
(previous_right_sample * coeff2r) + 0x80) >> 8;
previous_left_sample = current_left_sample;
current_left_sample = av_clip_int16(next_left_sample);
previous_right_sample = current_right_sample;
current_right_sample = av_clip_int16(next_right_sample);
*samples++ = (unsigned short)current_left_sample;
*samples++ = (unsigned short)current_right_sample;
}
}
if (src - buf == buf_size - 2)
src += 2; // Skip terminating 0x0000
break;
}
case CODEC_ID_ADPCM_EA_MAXIS_XA:
{
int coeff[2][2], shift[2];
for(channel = 0; channel < avctx->channels; channel++) {
for (i=0; i<2; i++)
coeff[channel][i] = ea_adpcm_table[(*src >> 4) + 4*i];
shift[channel] = (*src & 0x0F) + 8;
src++;
}
for (count1 = 0; count1 < (buf_size - avctx->channels) / avctx->channels; count1++) {
for(i = 4; i >= 0; i-=4) { /* Pairwise samples LL RR (st) or LL LL (mono) */
for(channel = 0; channel < avctx->channels; channel++) {
int32_t sample = (int32_t)(((*(src+channel) >> i) & 0x0F) << 0x1C) >> shift[channel];
sample = (sample +
c->status[channel].sample1 * coeff[channel][0] +
c->status[channel].sample2 * coeff[channel][1] + 0x80) >> 8;
c->status[channel].sample2 = c->status[channel].sample1;
c->status[channel].sample1 = av_clip_int16(sample);
*samples++ = c->status[channel].sample1;
}
}
src+=avctx->channels;
}
break;
}
case CODEC_ID_ADPCM_EA_R1:
case CODEC_ID_ADPCM_EA_R2:
case CODEC_ID_ADPCM_EA_R3: {
/* channel numbering
2chan: 0=fl, 1=fr
4chan: 0=fl, 1=rl, 2=fr, 3=rr
6chan: 0=fl, 1=c, 2=fr, 3=rl, 4=rr, 5=sub */
const int big_endian = avctx->codec->id == CODEC_ID_ADPCM_EA_R3;
int32_t previous_sample, current_sample, next_sample;
int32_t coeff1, coeff2;
uint8_t shift;
unsigned int channel;
uint16_t *samplesC;
const uint8_t *srcC;
const uint8_t *src_end = buf + buf_size;
samples_in_chunk = (big_endian ? bytestream_get_be32(&src)
: bytestream_get_le32(&src)) / 28;
if (samples_in_chunk > UINT32_MAX/(28*avctx->channels) ||
28*samples_in_chunk*avctx->channels > samples_end-samples) {
src += buf_size - 4;
break;
}
for (channel=0; channel<avctx->channels; channel++) {
int32_t offset = (big_endian ? bytestream_get_be32(&src)
: bytestream_get_le32(&src))
+ (avctx->channels-channel-1) * 4;
if ((offset < 0) || (offset >= src_end - src - 4)) break;
srcC = src + offset;
samplesC = samples + channel;
if (avctx->codec->id == CODEC_ID_ADPCM_EA_R1) {
current_sample = (int16_t)bytestream_get_le16(&srcC);
previous_sample = (int16_t)bytestream_get_le16(&srcC);
} else {
current_sample = c->status[channel].predictor;
previous_sample = c->status[channel].prev_sample;
}
for (count1=0; count1<samples_in_chunk; count1++) {
if (*srcC == 0xEE) { /* only seen in R2 and R3 */
srcC++;
if (srcC > src_end - 30*2) break;
current_sample = (int16_t)bytestream_get_be16(&srcC);
previous_sample = (int16_t)bytestream_get_be16(&srcC);
for (count2=0; count2<28; count2++) {
*samplesC = (int16_t)bytestream_get_be16(&srcC);
samplesC += avctx->channels;
}
} else {
coeff1 = ea_adpcm_table[ *srcC>>4 ];
coeff2 = ea_adpcm_table[(*srcC>>4) + 4];
shift = (*srcC++ & 0x0F) + 8;
if (srcC > src_end - 14) break;
for (count2=0; count2<28; count2++) {
if (count2 & 1)
next_sample = (int32_t)((*srcC++ & 0x0F) << 28) >> shift;
else
next_sample = (int32_t)((*srcC & 0xF0) << 24) >> shift;
next_sample += (current_sample * coeff1) +
(previous_sample * coeff2);
next_sample = av_clip_int16(next_sample >> 8);
previous_sample = current_sample;
current_sample = next_sample;
*samplesC = current_sample;
samplesC += avctx->channels;
}
}
}
if (avctx->codec->id != CODEC_ID_ADPCM_EA_R1) {
c->status[channel].predictor = current_sample;
c->status[channel].prev_sample = previous_sample;
}
}
src = src + buf_size - (4 + 4*avctx->channels);
samples += 28 * samples_in_chunk * avctx->channels;
break;
}
case CODEC_ID_ADPCM_EA_XAS:
if (samples_end-samples < 32*4*avctx->channels
|| buf_size < (4+15)*4*avctx->channels) {
src += buf_size;
break;
}
for (channel=0; channel<avctx->channels; channel++) {
int coeff[2][4], shift[4];
short *s2, *s = &samples[channel];
for (n=0; n<4; n++, s+=32*avctx->channels) {
for (i=0; i<2; i++)
coeff[i][n] = ea_adpcm_table[(src[0]&0x0F)+4*i];
shift[n] = (src[2]&0x0F) + 8;
for (s2=s, i=0; i<2; i++, src+=2, s2+=avctx->channels)
s2[0] = (src[0]&0xF0) + (src[1]<<8);
}
for (m=2; m<32; m+=2) {
s = &samples[m*avctx->channels + channel];
for (n=0; n<4; n++, src++, s+=32*avctx->channels) {
for (s2=s, i=0; i<8; i+=4, s2+=avctx->channels) {
int level = (int32_t)((*src & (0xF0>>i)) << (24+i)) >> shift[n];
int pred = s2[-1*avctx->channels] * coeff[0][n]
+ s2[-2*avctx->channels] * coeff[1][n];
s2[0] = av_clip_int16((level + pred + 0x80) >> 8);
}
}
}
}
samples += 32*4*avctx->channels;
break;
case CODEC_ID_ADPCM_IMA_AMV:
case CODEC_ID_ADPCM_IMA_SMJPEG:
c->status[0].predictor = (int16_t)bytestream_get_le16(&src);
c->status[0].step_index = bytestream_get_le16(&src);
if (avctx->codec->id == CODEC_ID_ADPCM_IMA_AMV)
src+=4;
while (src < buf + buf_size) {
char hi, lo;
lo = *src & 0x0F;
hi = *src >> 4;
if (avctx->codec->id == CODEC_ID_ADPCM_IMA_AMV)
FFSWAP(char, hi, lo);
*samples++ = adpcm_ima_expand_nibble(&c->status[0],
lo, 3);
*samples++ = adpcm_ima_expand_nibble(&c->status[0],
hi, 3);
src++;
}
break;
case CODEC_ID_ADPCM_CT:
while (src < buf + buf_size) {
uint8_t v = *src++;
*samples++ = adpcm_ct_expand_nibble(&c->status[0 ], v >> 4 );
*samples++ = adpcm_ct_expand_nibble(&c->status[st], v & 0x0F);
}
break;
case CODEC_ID_ADPCM_SBPRO_4:
case CODEC_ID_ADPCM_SBPRO_3:
case CODEC_ID_ADPCM_SBPRO_2:
if (!c->status[0].step_index) {
/* the first byte is a raw sample */
*samples++ = 128 * (*src++ - 0x80);
if (st)
*samples++ = 128 * (*src++ - 0x80);
c->status[0].step_index = 1;
}
if (avctx->codec->id == CODEC_ID_ADPCM_SBPRO_4) {
while (src < buf + buf_size) {
*samples++ = adpcm_sbpro_expand_nibble(&c->status[0],
src[0] >> 4, 4, 0);
*samples++ = adpcm_sbpro_expand_nibble(&c->status[st],
src[0] & 0x0F, 4, 0);
src++;
}
} else if (avctx->codec->id == CODEC_ID_ADPCM_SBPRO_3) {
while (src < buf + buf_size && samples + 2 < samples_end) {
*samples++ = adpcm_sbpro_expand_nibble(&c->status[0],
src[0] >> 5 , 3, 0);
*samples++ = adpcm_sbpro_expand_nibble(&c->status[0],
(src[0] >> 2) & 0x07, 3, 0);
*samples++ = adpcm_sbpro_expand_nibble(&c->status[0],
src[0] & 0x03, 2, 0);
src++;
}
} else {
while (src < buf + buf_size && samples + 3 < samples_end) {
*samples++ = adpcm_sbpro_expand_nibble(&c->status[0],
src[0] >> 6 , 2, 2);
*samples++ = adpcm_sbpro_expand_nibble(&c->status[st],
(src[0] >> 4) & 0x03, 2, 2);
*samples++ = adpcm_sbpro_expand_nibble(&c->status[0],
(src[0] >> 2) & 0x03, 2, 2);
*samples++ = adpcm_sbpro_expand_nibble(&c->status[st],
src[0] & 0x03, 2, 2);
src++;
}
}
break;
case CODEC_ID_ADPCM_SWF:
{
GetBitContext gb;
const int *table;
int k0, signmask, nb_bits, count;
int size = buf_size*8;
init_get_bits(&gb, buf, size);
//read bits & initial values
nb_bits = get_bits(&gb, 2)+2;
//av_log(NULL,AV_LOG_INFO,"nb_bits: %d\n", nb_bits);
table = swf_index_tables[nb_bits-2];
k0 = 1 << (nb_bits-2);
signmask = 1 << (nb_bits-1);
while (get_bits_count(&gb) <= size - 22*avctx->channels) {
for (i = 0; i < avctx->channels; i++) {
*samples++ = c->status[i].predictor = get_sbits(&gb, 16);
c->status[i].step_index = get_bits(&gb, 6);
}
for (count = 0; get_bits_count(&gb) <= size - nb_bits*avctx->channels && count < 4095; count++) {
int i;
for (i = 0; i < avctx->channels; i++) {
// similar to IMA adpcm
int delta = get_bits(&gb, nb_bits);
int step = ff_adpcm_step_table[c->status[i].step_index];
long vpdiff = 0; // vpdiff = (delta+0.5)*step/4
int k = k0;
do {
if (delta & k)
vpdiff += step;
step >>= 1;
k >>= 1;
} while(k);
vpdiff += step;
if (delta & signmask)
c->status[i].predictor -= vpdiff;
else
c->status[i].predictor += vpdiff;
c->status[i].step_index += table[delta & (~signmask)];
c->status[i].step_index = av_clip(c->status[i].step_index, 0, 88);
c->status[i].predictor = av_clip_int16(c->status[i].predictor);
*samples++ = c->status[i].predictor;
if (samples >= samples_end) {
av_log(avctx, AV_LOG_ERROR, "allocated output buffer is too small\n");
return -1;
}
}
}
}
src += buf_size;
break;
}
case CODEC_ID_ADPCM_YAMAHA:
while (src < buf + buf_size) {
uint8_t v = *src++;
*samples++ = adpcm_yamaha_expand_nibble(&c->status[0 ], v & 0x0F);
*samples++ = adpcm_yamaha_expand_nibble(&c->status[st], v >> 4 );
}
break;
case CODEC_ID_ADPCM_THP:
{
int table[2][16];
unsigned int samplecnt;
int prev[2][2];
int ch;
if (buf_size < 80) {
av_log(avctx, AV_LOG_ERROR, "frame too small\n");
return -1;
}
src+=4;
samplecnt = bytestream_get_be32(&src);
for (i = 0; i < 32; i++)
table[0][i] = (int16_t)bytestream_get_be16(&src);
/* Initialize the previous sample. */
for (i = 0; i < 4; i++)
prev[0][i] = (int16_t)bytestream_get_be16(&src);
if (samplecnt >= (samples_end - samples) / (st + 1)) {
av_log(avctx, AV_LOG_ERROR, "allocated output buffer is too small\n");
return -1;
}
for (ch = 0; ch <= st; ch++) {
samples = (unsigned short *) data + ch;
/* Read in every sample for this channel. */
for (i = 0; i < samplecnt / 14; i++) {
int index = (*src >> 4) & 7;
unsigned int exp = 28 - (*src++ & 15);
int factor1 = table[ch][index * 2];
int factor2 = table[ch][index * 2 + 1];
/* Decode 14 samples. */
for (n = 0; n < 14; n++) {
int32_t sampledat;
if(n&1) sampledat= *src++ <<28;
else sampledat= (*src&0xF0)<<24;
sampledat = ((prev[ch][0]*factor1
+ prev[ch][1]*factor2) >> 11) + (sampledat>>exp);
*samples = av_clip_int16(sampledat);
prev[ch][1] = prev[ch][0];
prev[ch][0] = *samples++;
/* In case of stereo, skip one sample, this sample
is for the other channel. */
samples += st;
}
}
}
/* In the previous loop, in case stereo is used, samples is
increased exactly one time too often. */
samples -= st;
break;
}
default:
return -1;
}
*data_size = (uint8_t *)samples - (uint8_t *)data;
return src - buf;
}
#define ADPCM_DECODER(id_, name_, long_name_) \
AVCodec ff_ ## name_ ## _decoder = { \
.name = #name_, \
.type = AVMEDIA_TYPE_AUDIO, \
.id = id_, \
.priv_data_size = sizeof(ADPCMDecodeContext), \
.init = adpcm_decode_init, \
.decode = adpcm_decode_frame, \
.long_name = NULL_IF_CONFIG_SMALL(long_name_), \
}
/* Note: Do not forget to add new entries to the Makefile as well. */
ADPCM_DECODER(CODEC_ID_ADPCM_4XM, adpcm_4xm, "ADPCM 4X Movie");
ADPCM_DECODER(CODEC_ID_ADPCM_CT, adpcm_ct, "ADPCM Creative Technology");
ADPCM_DECODER(CODEC_ID_ADPCM_EA, adpcm_ea, "ADPCM Electronic Arts");
ADPCM_DECODER(CODEC_ID_ADPCM_EA_MAXIS_XA, adpcm_ea_maxis_xa, "ADPCM Electronic Arts Maxis CDROM XA");
ADPCM_DECODER(CODEC_ID_ADPCM_EA_R1, adpcm_ea_r1, "ADPCM Electronic Arts R1");
ADPCM_DECODER(CODEC_ID_ADPCM_EA_R2, adpcm_ea_r2, "ADPCM Electronic Arts R2");
ADPCM_DECODER(CODEC_ID_ADPCM_EA_R3, adpcm_ea_r3, "ADPCM Electronic Arts R3");
ADPCM_DECODER(CODEC_ID_ADPCM_EA_XAS, adpcm_ea_xas, "ADPCM Electronic Arts XAS");
ADPCM_DECODER(CODEC_ID_ADPCM_IMA_AMV, adpcm_ima_amv, "ADPCM IMA AMV");
ADPCM_DECODER(CODEC_ID_ADPCM_IMA_DK3, adpcm_ima_dk3, "ADPCM IMA Duck DK3");
ADPCM_DECODER(CODEC_ID_ADPCM_IMA_DK4, adpcm_ima_dk4, "ADPCM IMA Duck DK4");
ADPCM_DECODER(CODEC_ID_ADPCM_IMA_EA_EACS, adpcm_ima_ea_eacs, "ADPCM IMA Electronic Arts EACS");
ADPCM_DECODER(CODEC_ID_ADPCM_IMA_EA_SEAD, adpcm_ima_ea_sead, "ADPCM IMA Electronic Arts SEAD");
ADPCM_DECODER(CODEC_ID_ADPCM_IMA_ISS, adpcm_ima_iss, "ADPCM IMA Funcom ISS");
ADPCM_DECODER(CODEC_ID_ADPCM_IMA_QT, adpcm_ima_qt, "ADPCM IMA QuickTime");
ADPCM_DECODER(CODEC_ID_ADPCM_IMA_SMJPEG, adpcm_ima_smjpeg, "ADPCM IMA Loki SDL MJPEG");
ADPCM_DECODER(CODEC_ID_ADPCM_IMA_WAV, adpcm_ima_wav, "ADPCM IMA WAV");
ADPCM_DECODER(CODEC_ID_ADPCM_IMA_WS, adpcm_ima_ws, "ADPCM IMA Westwood");
ADPCM_DECODER(CODEC_ID_ADPCM_MS, adpcm_ms, "ADPCM Microsoft");
ADPCM_DECODER(CODEC_ID_ADPCM_SBPRO_2, adpcm_sbpro_2, "ADPCM Sound Blaster Pro 2-bit");
ADPCM_DECODER(CODEC_ID_ADPCM_SBPRO_3, adpcm_sbpro_3, "ADPCM Sound Blaster Pro 2.6-bit");
ADPCM_DECODER(CODEC_ID_ADPCM_SBPRO_4, adpcm_sbpro_4, "ADPCM Sound Blaster Pro 4-bit");
ADPCM_DECODER(CODEC_ID_ADPCM_SWF, adpcm_swf, "ADPCM Shockwave Flash");
ADPCM_DECODER(CODEC_ID_ADPCM_THP, adpcm_thp, "ADPCM Nintendo Gamecube THP");
ADPCM_DECODER(CODEC_ID_ADPCM_XA, adpcm_xa, "ADPCM CDROM XA");
ADPCM_DECODER(CODEC_ID_ADPCM_YAMAHA, adpcm_yamaha, "ADPCM Yamaha");