mirror of https://git.ffmpeg.org/ffmpeg.git
vmdaudio: fix decoding of 16-bit audio format.
The initial sample of each block is raw 16-bit PCM, not DPCM. Fixes decoding of all samples in: http://streams.videolan.org/samples/game-formats/sierra-vmd/Lighthouse/
This commit is contained in:
parent
bb416bd68c
commit
4568c2bf97
|
@ -417,9 +417,8 @@ static av_cold int vmdvideo_decode_end(AVCodecContext *avctx)
|
||||||
#define BLOCK_TYPE_SILENCE 3
|
#define BLOCK_TYPE_SILENCE 3
|
||||||
|
|
||||||
typedef struct VmdAudioContext {
|
typedef struct VmdAudioContext {
|
||||||
AVCodecContext *avctx;
|
|
||||||
int out_bps;
|
int out_bps;
|
||||||
int predictors[2];
|
int chunk_size;
|
||||||
} VmdAudioContext;
|
} VmdAudioContext;
|
||||||
|
|
||||||
static const uint16_t vmdaudio_table[128] = {
|
static const uint16_t vmdaudio_table[128] = {
|
||||||
|
@ -442,13 +441,23 @@ static av_cold int vmdaudio_decode_init(AVCodecContext *avctx)
|
||||||
{
|
{
|
||||||
VmdAudioContext *s = avctx->priv_data;
|
VmdAudioContext *s = avctx->priv_data;
|
||||||
|
|
||||||
s->avctx = avctx;
|
if (avctx->channels < 1 || avctx->channels > 2) {
|
||||||
|
av_log(avctx, AV_LOG_ERROR, "invalid number of channels\n");
|
||||||
|
return AVERROR(EINVAL);
|
||||||
|
}
|
||||||
|
if (avctx->block_align < 1) {
|
||||||
|
av_log(avctx, AV_LOG_ERROR, "invalid block align\n");
|
||||||
|
return AVERROR(EINVAL);
|
||||||
|
}
|
||||||
|
|
||||||
if (avctx->bits_per_coded_sample == 16)
|
if (avctx->bits_per_coded_sample == 16)
|
||||||
avctx->sample_fmt = AV_SAMPLE_FMT_S16;
|
avctx->sample_fmt = AV_SAMPLE_FMT_S16;
|
||||||
else
|
else
|
||||||
avctx->sample_fmt = AV_SAMPLE_FMT_U8;
|
avctx->sample_fmt = AV_SAMPLE_FMT_U8;
|
||||||
s->out_bps = av_get_bytes_per_sample(avctx->sample_fmt);
|
s->out_bps = av_get_bytes_per_sample(avctx->sample_fmt);
|
||||||
|
|
||||||
|
s->chunk_size = avctx->block_align + avctx->channels * (s->out_bps == 2);
|
||||||
|
|
||||||
av_log(avctx, AV_LOG_DEBUG, "%d channels, %d bits/sample, "
|
av_log(avctx, AV_LOG_DEBUG, "%d channels, %d bits/sample, "
|
||||||
"block align = %d, sample rate = %d\n",
|
"block align = %d, sample rate = %d\n",
|
||||||
avctx->channels, avctx->bits_per_coded_sample, avctx->block_align,
|
avctx->channels, avctx->bits_per_coded_sample, avctx->block_align,
|
||||||
|
@ -457,52 +466,47 @@ static av_cold int vmdaudio_decode_init(AVCodecContext *avctx)
|
||||||
return 0;
|
return 0;
|
||||||
}
|
}
|
||||||
|
|
||||||
static void vmdaudio_decode_audio(VmdAudioContext *s, unsigned char *data,
|
static void decode_audio_s16(int16_t *out, const uint8_t *buf, int buf_size,
|
||||||
const uint8_t *buf, int buf_size, int stereo)
|
int channels)
|
||||||
{
|
{
|
||||||
int i;
|
int ch;
|
||||||
int chan = 0;
|
const uint8_t *buf_end = buf + buf_size;
|
||||||
int16_t *out = (int16_t*)data;
|
int predictor[2];
|
||||||
|
int st = channels - 1;
|
||||||
|
|
||||||
for(i = 0; i < buf_size; i++) {
|
/* decode initial raw sample */
|
||||||
if(buf[i] & 0x80)
|
for (ch = 0; ch < channels; ch++) {
|
||||||
s->predictors[chan] -= vmdaudio_table[buf[i] & 0x7F];
|
predictor[ch] = (int16_t)AV_RL16(buf);
|
||||||
|
buf += 2;
|
||||||
|
*out++ = predictor[ch];
|
||||||
|
}
|
||||||
|
|
||||||
|
/* decode DPCM samples */
|
||||||
|
ch = 0;
|
||||||
|
while (buf < buf_end) {
|
||||||
|
uint8_t b = *buf++;
|
||||||
|
if (b & 0x80)
|
||||||
|
predictor[ch] -= vmdaudio_table[b & 0x7F];
|
||||||
else
|
else
|
||||||
s->predictors[chan] += vmdaudio_table[buf[i]];
|
predictor[ch] += vmdaudio_table[b];
|
||||||
s->predictors[chan] = av_clip_int16(s->predictors[chan]);
|
predictor[ch] = av_clip_int16(predictor[ch]);
|
||||||
out[i] = s->predictors[chan];
|
*out++ = predictor[ch];
|
||||||
chan ^= stereo;
|
ch ^= st;
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
static int vmdaudio_loadsound(VmdAudioContext *s, unsigned char *data,
|
|
||||||
const uint8_t *buf, int silent_chunks, int data_size)
|
|
||||||
{
|
|
||||||
int silent_size = s->avctx->block_align * silent_chunks * s->out_bps;
|
|
||||||
|
|
||||||
if (silent_chunks) {
|
|
||||||
memset(data, s->out_bps == 2 ? 0x00 : 0x80, silent_size);
|
|
||||||
data += silent_size;
|
|
||||||
}
|
|
||||||
if (s->avctx->bits_per_coded_sample == 16)
|
|
||||||
vmdaudio_decode_audio(s, data, buf, data_size, s->avctx->channels == 2);
|
|
||||||
else {
|
|
||||||
/* just copy the data */
|
|
||||||
memcpy(data, buf, data_size);
|
|
||||||
}
|
|
||||||
|
|
||||||
return silent_size + data_size * s->out_bps;
|
|
||||||
}
|
|
||||||
|
|
||||||
static int vmdaudio_decode_frame(AVCodecContext *avctx,
|
static int vmdaudio_decode_frame(AVCodecContext *avctx,
|
||||||
void *data, int *data_size,
|
void *data, int *data_size,
|
||||||
AVPacket *avpkt)
|
AVPacket *avpkt)
|
||||||
{
|
{
|
||||||
const uint8_t *buf = avpkt->data;
|
const uint8_t *buf = avpkt->data;
|
||||||
|
const uint8_t *buf_end;
|
||||||
int buf_size = avpkt->size;
|
int buf_size = avpkt->size;
|
||||||
VmdAudioContext *s = avctx->priv_data;
|
VmdAudioContext *s = avctx->priv_data;
|
||||||
int block_type, silent_chunks;
|
int block_type, silent_chunks, audio_chunks;
|
||||||
unsigned char *output_samples = (unsigned char *)data;
|
int nb_samples, out_size;
|
||||||
|
uint8_t *output_samples_u8 = data;
|
||||||
|
int16_t *output_samples_s16 = data;
|
||||||
|
|
||||||
if (buf_size < 16) {
|
if (buf_size < 16) {
|
||||||
av_log(avctx, AV_LOG_WARNING, "skipping small junk packet\n");
|
av_log(avctx, AV_LOG_WARNING, "skipping small junk packet\n");
|
||||||
|
@ -518,10 +522,16 @@ static int vmdaudio_decode_frame(AVCodecContext *avctx,
|
||||||
buf += 16;
|
buf += 16;
|
||||||
buf_size -= 16;
|
buf_size -= 16;
|
||||||
|
|
||||||
|
/* get number of silent chunks */
|
||||||
silent_chunks = 0;
|
silent_chunks = 0;
|
||||||
if (block_type == BLOCK_TYPE_INITIAL) {
|
if (block_type == BLOCK_TYPE_INITIAL) {
|
||||||
uint32_t flags = AV_RB32(buf);
|
uint32_t flags;
|
||||||
silent_chunks = av_popcount(flags);
|
if (buf_size < 4) {
|
||||||
|
av_log(avctx, AV_LOG_ERROR, "packet is too small\n");
|
||||||
|
return AVERROR(EINVAL);
|
||||||
|
}
|
||||||
|
flags = AV_RB32(buf);
|
||||||
|
silent_chunks = av_popcount(flags);
|
||||||
buf += 4;
|
buf += 4;
|
||||||
buf_size -= 4;
|
buf_size -= 4;
|
||||||
} else if (block_type == BLOCK_TYPE_SILENCE) {
|
} else if (block_type == BLOCK_TYPE_SILENCE) {
|
||||||
|
@ -530,11 +540,41 @@ static int vmdaudio_decode_frame(AVCodecContext *avctx,
|
||||||
}
|
}
|
||||||
|
|
||||||
/* ensure output buffer is large enough */
|
/* ensure output buffer is large enough */
|
||||||
if (*data_size < (avctx->block_align*silent_chunks + buf_size) * s->out_bps)
|
audio_chunks = buf_size / s->chunk_size;
|
||||||
|
nb_samples = ((silent_chunks + audio_chunks) * avctx->block_align) / avctx->channels;
|
||||||
|
out_size = nb_samples * avctx->channels * s->out_bps;
|
||||||
|
if (*data_size < out_size)
|
||||||
return -1;
|
return -1;
|
||||||
|
|
||||||
*data_size = vmdaudio_loadsound(s, output_samples, buf, silent_chunks, buf_size);
|
/* decode silent chunks */
|
||||||
|
if (silent_chunks > 0) {
|
||||||
|
int silent_size = avctx->block_align * silent_chunks;
|
||||||
|
if (s->out_bps == 2) {
|
||||||
|
memset(output_samples_s16, 0x00, silent_size * 2);
|
||||||
|
output_samples_s16 += silent_size;
|
||||||
|
} else {
|
||||||
|
memset(output_samples_u8, 0x80, silent_size);
|
||||||
|
output_samples_u8 += silent_size;
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
|
/* decode audio chunks */
|
||||||
|
if (audio_chunks > 0) {
|
||||||
|
buf_end = buf + buf_size;
|
||||||
|
while (buf < buf_end) {
|
||||||
|
if (s->out_bps == 2) {
|
||||||
|
decode_audio_s16(output_samples_s16, buf, s->chunk_size,
|
||||||
|
avctx->channels);
|
||||||
|
output_samples_s16 += avctx->block_align;
|
||||||
|
} else {
|
||||||
|
memcpy(output_samples_u8, buf, s->chunk_size);
|
||||||
|
output_samples_u8 += avctx->block_align;
|
||||||
|
}
|
||||||
|
buf += s->chunk_size;
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
|
*data_size = out_size;
|
||||||
return avpkt->size;
|
return avpkt->size;
|
||||||
}
|
}
|
||||||
|
|
||||||
|
|
Loading…
Reference in New Issue