mirror of
https://github.com/FFmpeg/FFmpeg.git
synced 2025-01-24 13:56:33 +02:00
roqaudioenc: use AVCodecContext.frame_size correctly.
It is not allowed to change mid-stream like it does currently. Instead we need to buffer the first 8 frames before returning them as a single packet, then only return single frame packets after that.
This commit is contained in:
parent
56279f1d61
commit
cf57c78b77
@ -25,9 +25,8 @@
|
|||||||
#include "avcodec.h"
|
#include "avcodec.h"
|
||||||
#include "bytestream.h"
|
#include "bytestream.h"
|
||||||
|
|
||||||
#define ROQ_FIRST_FRAME_SIZE (735*8)
|
|
||||||
#define ROQ_FRAME_SIZE 735
|
#define ROQ_FRAME_SIZE 735
|
||||||
|
#define ROQ_HEADER_SIZE 8
|
||||||
|
|
||||||
#define MAX_DPCM (127*127)
|
#define MAX_DPCM (127*127)
|
||||||
|
|
||||||
@ -35,11 +34,26 @@
|
|||||||
typedef struct
|
typedef struct
|
||||||
{
|
{
|
||||||
short lastSample[2];
|
short lastSample[2];
|
||||||
|
int input_frames;
|
||||||
|
int buffered_samples;
|
||||||
|
int16_t *frame_buffer;
|
||||||
} ROQDPCMContext;
|
} ROQDPCMContext;
|
||||||
|
|
||||||
|
|
||||||
|
static av_cold int roq_dpcm_encode_close(AVCodecContext *avctx)
|
||||||
|
{
|
||||||
|
ROQDPCMContext *context = avctx->priv_data;
|
||||||
|
|
||||||
|
av_freep(&avctx->coded_frame);
|
||||||
|
av_freep(&context->frame_buffer);
|
||||||
|
|
||||||
|
return 0;
|
||||||
|
}
|
||||||
|
|
||||||
static av_cold int roq_dpcm_encode_init(AVCodecContext *avctx)
|
static av_cold int roq_dpcm_encode_init(AVCodecContext *avctx)
|
||||||
{
|
{
|
||||||
ROQDPCMContext *context = avctx->priv_data;
|
ROQDPCMContext *context = avctx->priv_data;
|
||||||
|
int ret;
|
||||||
|
|
||||||
if (avctx->channels > 2) {
|
if (avctx->channels > 2) {
|
||||||
av_log(avctx, AV_LOG_ERROR, "Audio must be mono or stereo\n");
|
av_log(avctx, AV_LOG_ERROR, "Audio must be mono or stereo\n");
|
||||||
@ -50,15 +64,27 @@ static av_cold int roq_dpcm_encode_init(AVCodecContext *avctx)
|
|||||||
return -1;
|
return -1;
|
||||||
}
|
}
|
||||||
|
|
||||||
avctx->frame_size = ROQ_FIRST_FRAME_SIZE;
|
avctx->frame_size = ROQ_FRAME_SIZE;
|
||||||
|
|
||||||
|
context->frame_buffer = av_malloc(8 * ROQ_FRAME_SIZE * avctx->channels *
|
||||||
|
sizeof(*context->frame_buffer));
|
||||||
|
if (!context->frame_buffer) {
|
||||||
|
ret = AVERROR(ENOMEM);
|
||||||
|
goto error;
|
||||||
|
}
|
||||||
|
|
||||||
context->lastSample[0] = context->lastSample[1] = 0;
|
context->lastSample[0] = context->lastSample[1] = 0;
|
||||||
|
|
||||||
avctx->coded_frame= avcodec_alloc_frame();
|
avctx->coded_frame= avcodec_alloc_frame();
|
||||||
if (!avctx->coded_frame)
|
if (!avctx->coded_frame) {
|
||||||
return AVERROR(ENOMEM);
|
ret = AVERROR(ENOMEM);
|
||||||
|
goto error;
|
||||||
|
}
|
||||||
|
|
||||||
return 0;
|
return 0;
|
||||||
|
error:
|
||||||
|
roq_dpcm_encode_close(avctx);
|
||||||
|
return ret;
|
||||||
}
|
}
|
||||||
|
|
||||||
static unsigned char dpcm_predict(short *previous, short current)
|
static unsigned char dpcm_predict(short *previous, short current)
|
||||||
@ -104,25 +130,45 @@ static unsigned char dpcm_predict(short *previous, short current)
|
|||||||
static int roq_dpcm_encode_frame(AVCodecContext *avctx,
|
static int roq_dpcm_encode_frame(AVCodecContext *avctx,
|
||||||
unsigned char *frame, int buf_size, void *data)
|
unsigned char *frame, int buf_size, void *data)
|
||||||
{
|
{
|
||||||
int i, samples, stereo, ch;
|
int i, stereo, data_size;
|
||||||
const short *in;
|
const int16_t *in = data;
|
||||||
unsigned char *out;
|
uint8_t *out = frame;
|
||||||
|
|
||||||
ROQDPCMContext *context = avctx->priv_data;
|
ROQDPCMContext *context = avctx->priv_data;
|
||||||
|
|
||||||
stereo = (avctx->channels == 2);
|
stereo = (avctx->channels == 2);
|
||||||
|
|
||||||
|
if (!data && context->input_frames >= 8)
|
||||||
|
return 0;
|
||||||
|
|
||||||
|
if (data && context->input_frames < 8) {
|
||||||
|
memcpy(&context->frame_buffer[context->buffered_samples * avctx->channels],
|
||||||
|
in, avctx->frame_size * avctx->channels * sizeof(*in));
|
||||||
|
context->buffered_samples += avctx->frame_size;
|
||||||
|
if (context->input_frames < 7) {
|
||||||
|
context->input_frames++;
|
||||||
|
return 0;
|
||||||
|
}
|
||||||
|
in = context->frame_buffer;
|
||||||
|
}
|
||||||
|
|
||||||
if (stereo) {
|
if (stereo) {
|
||||||
context->lastSample[0] &= 0xFF00;
|
context->lastSample[0] &= 0xFF00;
|
||||||
context->lastSample[1] &= 0xFF00;
|
context->lastSample[1] &= 0xFF00;
|
||||||
}
|
}
|
||||||
|
|
||||||
out = frame;
|
if (context->input_frames == 7 || !data)
|
||||||
in = data;
|
data_size = avctx->channels * context->buffered_samples;
|
||||||
|
else
|
||||||
|
data_size = avctx->channels * avctx->frame_size;
|
||||||
|
|
||||||
|
if (buf_size < ROQ_HEADER_SIZE + data_size) {
|
||||||
|
av_log(avctx, AV_LOG_ERROR, "output buffer is too small\n");
|
||||||
|
return AVERROR(EINVAL);
|
||||||
|
}
|
||||||
|
|
||||||
bytestream_put_byte(&out, stereo ? 0x21 : 0x20);
|
bytestream_put_byte(&out, stereo ? 0x21 : 0x20);
|
||||||
bytestream_put_byte(&out, 0x10);
|
bytestream_put_byte(&out, 0x10);
|
||||||
bytestream_put_le32(&out, avctx->frame_size*avctx->channels);
|
bytestream_put_le32(&out, data_size);
|
||||||
|
|
||||||
if (stereo) {
|
if (stereo) {
|
||||||
bytestream_put_byte(&out, (context->lastSample[1])>>8);
|
bytestream_put_byte(&out, (context->lastSample[1])>>8);
|
||||||
@ -131,23 +177,15 @@ static int roq_dpcm_encode_frame(AVCodecContext *avctx,
|
|||||||
bytestream_put_le16(&out, context->lastSample[0]);
|
bytestream_put_le16(&out, context->lastSample[0]);
|
||||||
|
|
||||||
/* Write the actual samples */
|
/* Write the actual samples */
|
||||||
samples = avctx->frame_size;
|
for (i = 0; i < data_size; i++)
|
||||||
for (i=0; i<samples; i++)
|
*out++ = dpcm_predict(&context->lastSample[i & 1], *in++);
|
||||||
for (ch=0; ch<avctx->channels; ch++)
|
|
||||||
*out++ = dpcm_predict(&context->lastSample[ch], *in++);
|
|
||||||
|
|
||||||
/* Use smaller frames from now on */
|
context->input_frames++;
|
||||||
avctx->frame_size = ROQ_FRAME_SIZE;
|
if (!data)
|
||||||
|
context->input_frames = FFMAX(context->input_frames, 8);
|
||||||
|
|
||||||
/* Return the result size */
|
/* Return the result size */
|
||||||
return out - frame;
|
return ROQ_HEADER_SIZE + data_size;
|
||||||
}
|
|
||||||
|
|
||||||
static av_cold int roq_dpcm_encode_close(AVCodecContext *avctx)
|
|
||||||
{
|
|
||||||
av_freep(&avctx->coded_frame);
|
|
||||||
|
|
||||||
return 0;
|
|
||||||
}
|
}
|
||||||
|
|
||||||
AVCodec ff_roq_dpcm_encoder = {
|
AVCodec ff_roq_dpcm_encoder = {
|
||||||
@ -158,6 +196,7 @@ AVCodec ff_roq_dpcm_encoder = {
|
|||||||
.init = roq_dpcm_encode_init,
|
.init = roq_dpcm_encode_init,
|
||||||
.encode = roq_dpcm_encode_frame,
|
.encode = roq_dpcm_encode_frame,
|
||||||
.close = roq_dpcm_encode_close,
|
.close = roq_dpcm_encode_close,
|
||||||
|
.capabilities = CODEC_CAP_DELAY,
|
||||||
.sample_fmts = (const enum AVSampleFormat[]){AV_SAMPLE_FMT_S16,AV_SAMPLE_FMT_NONE},
|
.sample_fmts = (const enum AVSampleFormat[]){AV_SAMPLE_FMT_S16,AV_SAMPLE_FMT_NONE},
|
||||||
.long_name = NULL_IF_CONFIG_SMALL("id RoQ DPCM"),
|
.long_name = NULL_IF_CONFIG_SMALL("id RoQ DPCM"),
|
||||||
};
|
};
|
||||||
|
Loading…
x
Reference in New Issue
Block a user