[FFmpeg-devel] [PATCH] avformat/dv: always set audio packet duration

[Paul B Mahol]

On Mon, Sep 6, 2021 at 8:29 PM Andreas Rheinhardt <
andreas.rheinhardt at outlook.com> wrote:

> Paul B Mahol:
> > If audio packet is present in DV stream it have duration of 1 in DV
> timebase units.
> >
> > Signed-off-by: Paul B Mahol <onemda at gmail.com>
> > ---
> >  libavformat/dv.c | 4 ++++
> >  1 file changed, 4 insertions(+)
> >
> > diff --git a/libavformat/dv.c b/libavformat/dv.c
> > index d7909683c3..b2b74162df 100644
> > --- a/libavformat/dv.c
> > +++ b/libavformat/dv.c
> > @@ -48,6 +48,7 @@ struct DVPacket {
> >      int      stream_index;
> >      int      flags;
> >      int64_t  pos;
> > +    int64_t  duration;
> >  };
> >
> >  struct DVDemuxContext {
> > @@ -276,6 +277,7 @@ static int dv_extract_audio_info(DVDemuxContext *c,
> const uint8_t *frame)
> >              c->audio_pkt[i].stream_index = c->ast[i]->index;
> >              c->audio_pkt[i].flags       |= AV_PKT_FLAG_KEY;
> >              c->audio_pkt[i].pts          = AV_NOPTS_VALUE;
> > +            c->audio_pkt[i].duration     = 0;
> >              c->audio_pkt[i].pos          = -1;
> >          }
> >          c->ast[i]->codecpar->sample_rate    = dv_audio_frequency[freq];
> > @@ -374,6 +376,7 @@ int avpriv_dv_get_packet(DVDemuxContext *c, AVPacket
> *pkt)
> >              pkt->stream_index = c->audio_pkt[i].stream_index;
> >              pkt->flags        = c->audio_pkt[i].flags;
> >              pkt->pts          = c->audio_pkt[i].pts;
> > +            pkt->duration     = c->audio_pkt[i].duration;
> >              pkt->pos          = c->audio_pkt[i].pos;
> >
> >              c->audio_pkt[i].size = 0;
> > @@ -404,6 +407,7 @@ int avpriv_dv_produce_packet(DVDemuxContext *c,
> AVPacket *pkt,
> >          c->audio_pkt[i].pos  = pos;
> >          c->audio_pkt[i].size = size;
> >          c->audio_pkt[i].pts  = (c->sys->height == 720) ? (c->frames &
> ~1) : c->frames;
> > +        c->audio_pkt[i].duration = 1;
> >          ppcm[i] = c->audio_buf[i];
> >      }
> >      if (c->ach)
> >
> Sure about that? According to the code, the packets contain slightly
> different amounts of samples; see dv_extract_audio_info(), in particular
> lines 242 and 289.
> (It seems to me that one audio packet is supposed to contain the audio
> for one video frame, yet the standard sample rates and standard video
> framerates (in particular NTSC) are incompatible, so a slight variation
> in the number of samples per frame is used.)
> IMO one should set the audio timebase to be the inverse of the
> samplerate and set the duration based upon the number of samples. But I
> have zero experience with dv.
>

I had exact same thoughts, but DV have timestamps actually, and timestamps
are in different units than sample rate.
And audio timestamps are linked to video timestamps.


>
> - 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".
>