Browse Source

dca: decode directly to the user-provided AVFrame

tags/n1.2
Justin Ruggles 12 years ago
parent
commit
182821cff4
1 changed files with 7 additions and 11 deletions
  1. +7
    -11
      libavcodec/dcadec.c

+ 7
- 11
libavcodec/dcadec.c View File

@@ -285,7 +285,6 @@ static av_always_inline int get_bitalloc(GetBitContext *gb, BitAlloc *ba,


typedef struct { typedef struct {
AVCodecContext *avctx; AVCodecContext *avctx;
AVFrame frame;
/* Frame header */ /* Frame header */
int frame_type; ///< type of the current frame int frame_type; ///< type of the current frame
int samples_deficit; ///< deficit sample count int samples_deficit; ///< deficit sample count
@@ -1653,6 +1652,7 @@ static void dca_exss_parse_header(DCAContext *s)
static int dca_decode_frame(AVCodecContext *avctx, void *data, static int dca_decode_frame(AVCodecContext *avctx, void *data,
int *got_frame_ptr, AVPacket *avpkt) int *got_frame_ptr, AVPacket *avpkt)
{ {
AVFrame *frame = data;
const uint8_t *buf = avpkt->data; const uint8_t *buf = avpkt->data;
int buf_size = avpkt->size; int buf_size = avpkt->size;


@@ -1835,17 +1835,17 @@ static int dca_decode_frame(AVCodecContext *avctx, void *data,
avctx->channels = channels; avctx->channels = channels;


/* get output buffer */ /* get output buffer */
s->frame.nb_samples = 256 * (s->sample_blocks / 8);
if ((ret = ff_get_buffer(avctx, &s->frame)) < 0) {
frame->nb_samples = 256 * (s->sample_blocks / 8);
if ((ret = ff_get_buffer(avctx, frame)) < 0) {
av_log(avctx, AV_LOG_ERROR, "get_buffer() failed\n"); av_log(avctx, AV_LOG_ERROR, "get_buffer() failed\n");
return ret; return ret;
} }
samples_flt = (float **) s->frame.extended_data;
samples_flt = (float **)frame->extended_data;


/* allocate buffer for extra channels if downmixing */ /* allocate buffer for extra channels if downmixing */
if (avctx->channels < full_channels) { if (avctx->channels < full_channels) {
ret = av_samples_get_buffer_size(NULL, full_channels - channels, ret = av_samples_get_buffer_size(NULL, full_channels - channels,
s->frame.nb_samples,
frame->nb_samples,
avctx->sample_fmt, 0); avctx->sample_fmt, 0);
if (ret < 0) if (ret < 0)
return ret; return ret;
@@ -1858,7 +1858,7 @@ static int dca_decode_frame(AVCodecContext *avctx, void *data,
ret = av_samples_fill_arrays((uint8_t **)s->extra_channels, NULL, ret = av_samples_fill_arrays((uint8_t **)s->extra_channels, NULL,
s->extra_channels_buffer, s->extra_channels_buffer,
full_channels - channels, full_channels - channels,
s->frame.nb_samples, avctx->sample_fmt, 0);
frame->nb_samples, avctx->sample_fmt, 0);
if (ret < 0) if (ret < 0)
return ret; return ret;
} }
@@ -1890,8 +1890,7 @@ static int dca_decode_frame(AVCodecContext *avctx, void *data,
for (i = 0; i < 2 * s->lfe * 4; i++) for (i = 0; i < 2 * s->lfe * 4; i++)
s->lfe_data[i] = s->lfe_data[i + lfe_samples]; s->lfe_data[i] = s->lfe_data[i + lfe_samples];


*got_frame_ptr = 1;
*(AVFrame *) data = s->frame;
*got_frame_ptr = 1;


return buf_size; return buf_size;
} }
@@ -1925,9 +1924,6 @@ static av_cold int dca_decode_init(AVCodecContext *avctx)
avctx->channels = avctx->request_channels; avctx->channels = avctx->request_channels;
} }


avcodec_get_frame_defaults(&s->frame);
avctx->coded_frame = &s->frame;

return 0; return 0;
} }




Loading…
Cancel
Save