[FFmpeg-devel] [PATCH] avformat/pcmdec: add pts and dts calculation for pcmdec
Hiccup Zhu
hiccupzhu at gmail.com
Wed May 15 14:48:32 EEST 2024
Andreas:
>
> Hiccup Zhu:
> > The purpose of this patch is to calculate pts and dts when using pcmdemux.
> > Is there anything wrong with doing this, or do you have any suggestions for
> > improvement?
> >
>
> 1. Don't top-post on this list.
> 2. PTS and DTS are already produced with this demuxer. As has been said:
> If it isn't for you, open a ticket about it.
This is the case. I found that when opening a pcm file,
avformat_find_stream_info will keep reading pkt until the number >
max_ts_probe then exit.
The reason is that when demux pcm was used, valid pts and dts were not
read, and sti->first_pts was never set correctly;
This is unreasonable in some scenarios, because
avformat_find_stream_info will consume more time to read pkt, which is
especially serious in the case of network streams;
You can reproduce this problem by opening any pcm file.
Based on the above facts, I submitted this patch;
Of course, this problem can also be fixed by solving the assignment
problem of sti->first_pts, in another patch of mine:
https://patchwork.ffmpeg.org/project/ffmpeg/patch/20240515113522.1921274-1-hiccupzhu@gmail.com/
- Shiqi
> >
> >> Shiqi Zhu:
> >>> Signed-off-by: Shiqi Zhu <hiccupzhu at gmail.com>
> >>> ---
> >>> libavformat/pcmdec.c | 37 +++++++++++++++++++++++++++++++++++--
> >>> 1 file changed, 35 insertions(+), 2 deletions(-)
> >>>
> >>> diff --git a/libavformat/pcmdec.c b/libavformat/pcmdec.c
> >>> index 2f6508b75a..d879aefaad 100644
> >>> --- a/libavformat/pcmdec.c
> >>> +++ b/libavformat/pcmdec.c
> >>> @@ -36,6 +36,7 @@ typedef struct PCMAudioDemuxerContext {
> >>> AVClass *class;
> >>> int sample_rate;
> >>> AVChannelLayout ch_layout;
> >>> + int64_t nb_samples;
> >>> } PCMAudioDemuxerContext;
> >>>
> >>> static int pcm_read_header(AVFormatContext *s)
> >>> @@ -46,6 +47,7 @@ static int pcm_read_header(AVFormatContext *s)
> >>> uint8_t *mime_type = NULL;
> >>> int ret;
> >>>
> >>> + s1->nb_samples = 0;
> >>> st = avformat_new_stream(s, NULL);
> >>> if (!st)
> >>> return AVERROR(ENOMEM);
> >>> @@ -104,6 +106,37 @@ static int pcm_read_header(AVFormatContext *s)
> >>> return 0;
> >>> }
> >>>
> >>> +static int pcm_dec_read_packet(AVFormatContext *s, AVPacket *pkt)
> >>> +{
> >>> + PCMAudioDemuxerContext *s1 = s->priv_data;
> >>> + AVCodecParameters *par = s->streams[0]->codecpar;
> >>> + int ret;
> >>> +
> >>> + ret = ff_pcm_read_packet(s, pkt);
> >>> + if (ret < 0)
> >>> + return ret;
> >>> +
> >>> + pkt->time_base = s->streams[0]->time_base;
> >>> + pkt->dts = pkt->pts = s1->nb_samples;
> >>> + s1->nb_samples += pkt->size / par->block_align;
> >>> +
> >>> + return ret;
> >>> +}
> >>> +
> >>> +static int pcm_dec_read_seek(AVFormatContext *s,
> >>> + int stream_index, int64_t timestamp, int
> >> flags)
> >>> +{
> >>> + PCMAudioDemuxerContext *s1 = s->priv_data;
> >>> + int ret;
> >>> +
> >>> + ret = ff_pcm_read_seek(s, stream_index, timestamp, flags);
> >>> + if (ret < 0)
> >>> + return ret;
> >>> +
> >>> + s1->nb_samples = ffstream(s->streams[0])->cur_dts;
> >>> + return ret;
> >>> +}
> >>> +
> >>> static const AVOption pcm_options[] = {
> >>> { "sample_rate", "", offsetof(PCMAudioDemuxerContext, sample_rate),
> >> AV_OPT_TYPE_INT, {.i64 = 44100}, 0, INT_MAX, AV_OPT_FLAG_DECODING_PARAM },
> >>> { "ch_layout", "", offsetof(PCMAudioDemuxerContext, ch_layout),
> >> AV_OPT_TYPE_CHLAYOUT, {.str = "mono"}, 0, 0, AV_OPT_FLAG_DECODING_PARAM },
> >>> @@ -126,8 +159,8 @@ const FFInputFormat ff_pcm_ ## name_ ## _demuxer =
> >> { \
> >>> .p.priv_class = &pcm_demuxer_class, \
> >>> .priv_data_size = sizeof(PCMAudioDemuxerContext), \
> >>> .read_header = pcm_read_header, \
> >>> - .read_packet = ff_pcm_read_packet, \
> >>> - .read_seek = ff_pcm_read_seek, \
> >>> + .read_packet = pcm_dec_read_packet, \
> >>> + .read_seek = pcm_dec_read_seek, \
> >>> .raw_codec_id = codec, \
> >>> __VA_ARGS__ \
> >>> };
> >>
> >> A quick test shows that PTS and DTS are already set generically for pcm
> >> formats (unless the AVFMT_FLAG_NOFILLIN flag is set). If it is not in
> >> your usecase, then you should provide details about this (preferably by
> >> opening a ticket on trac).
> >>
> >> - Andreas
> >>
> >> _______________________________________________
> >> ffmpeg-devel mailing list
> >> ffmpeg-devel at ffmpeg.org
> >> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
> >>
> >> To unsubscribe, visit link above, or email
> >> ffmpeg-devel-request at ffmpeg.org with subject "unsubscribe".
> >>
> >
> >
>
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> ffmpeg-devel-request at ffmpeg.org with subject "unsubscribe".
More information about the ffmpeg-devel
mailing list