[FFmpeg-devel] [PATCH v1] avformat/mux: Set AV_PKT_FLAG_KEY for is_intra_only packet

[Limin Wang]

On Fri, Apr 03, 2020 at 05:19:08PM +0200, Andreas Rheinhardt wrote:
> lance.lmwang at gmail.com:
> > From: Limin Wang <lance.lmwang at gmail.com>
> > 
> > Signed-off-by: Limin Wang <lance.lmwang at gmail.com>
> > ---
> >  libavformat/mux.c                | 16 ++++++++
> >  tests/ref/fate/binsub-movtextenc |  2 +-
> >  tests/ref/fate/movenc            | 50 +++++++++++------------
> >  tests/ref/fate/sub2video         | 86 ++++++++++++++++++++--------------------
> >  4 files changed, 85 insertions(+), 69 deletions(-)
> > 
> > diff --git a/libavformat/mux.c b/libavformat/mux.c
> > index cc2d1e2..dc897c2 100644
> > --- a/libavformat/mux.c
> > +++ b/libavformat/mux.c
> > @@ -548,6 +548,17 @@ fail:
> >      return ret;
> >  }
> >  
> > +static int is_intra_only(enum AVCodecID id)
> > +{
> > +    const AVCodecDescriptor *d = avcodec_descriptor_get(id);
> > +    if (!d)
> > +        return 0;
> > +    if ((d->type == AVMEDIA_TYPE_VIDEO || d->type == AVMEDIA_TYPE_AUDIO) &&
> > +        !(d->props & AV_CODEC_PROP_INTRA_ONLY))
> > +        return 0;
> > +    return 1;
> > +}
> > +
> >  #define AV_PKT_FLAG_UNCODED_FRAME 0x2000
> >  
> >  /* Note: using sizeof(AVFrame) from outside lavu is unsafe in general, but
> > @@ -664,6 +675,11 @@ static int compute_muxer_pkt_fields(AVFormatContext *s, AVStream *st, AVPacket *
> >          frac_add(st->internal->priv_pts, (int64_t)st->time_base.den * st->time_base.num);
> >          break;
> >      }
> > +
> > +    /* update flags */
> > +    if (is_intra_only(st->codecpar->codec_id))
> > +        pkt->flags |= AV_PKT_FLAG_KEY;
> > +
> >      return 0;
> >  }
> >  FF_ENABLE_DEPRECATION_WARNINGS
> 
> You have added the check to code that is deprecated and slated to be
> removed. Maybe it would be better to add it to prepare_input_packet()
> instead?

OK, I'll move the check to prepare_input_packet.

> 
> > diff --git a/tests/ref/fate/binsub-movtextenc b/tests/ref/fate/binsub-movtextenc
> > index 78c05f4..a8f94b7 100644
> > --- a/tests/ref/fate/binsub-movtextenc
> > +++ b/tests/ref/fate/binsub-movtextenc
> > @@ -1 +1 @@
> > -35adf776cd73e808186ae7124445f4b8
> > +fc6d07679ac1f718aa50de687924cd97
> > diff --git a/tests/ref/fate/movenc b/tests/ref/fate/movenc
> > index 637a347..fb39b98 100644
> > --- a/tests/ref/fate/movenc
> > +++ b/tests/ref/fate/movenc
> > @@ -2,17 +2,17 @@ write_data len 36, time nopts, type header atom ftyp
> >  write_data len 2389, time nopts, type header atom -
> >  write_data len 788, time 1000000, type sync atom moof
> >  write_data len 110, time nopts, type trailer atom -
> > -66cf48604f039aa9a51711786f5c8778 3323 non-empty-moov
> > +5f401347fc3c771b819e2449d69d4861 3323 non-empty-moov
> >  write_data len 36, time nopts, type header atom ftyp
> >  write_data len 2721, time nopts, type header atom -
> >  write_data len 908, time 966667, type sync atom moof
> >  write_data len 110, time nopts, type trailer atom -
> > -04b2e86f455af94f9258b8d66dbf71f5 3775 non-empty-moov-elst
> > +4267feee527adf8cd4f7b36ac0fc0872 3775 non-empty-moov-elst
> >  write_data len 36, time nopts, type header atom ftyp
> >  write_data len 2629, time nopts, type header atom -
> >  write_data len 908, time 1000000, type sync atom moof
> >  write_data len 110, time nopts, type trailer atom -
> > -e9f6fa032d6d8265d67aef5de81a48bf 3683 non-empty-moov-no-elst
> > +44077b9ad45f3e16fafe4e5ada54e9b0 3683 non-empty-moov-no-elst
> >  write_data len 24, time nopts, type header atom ftyp
> [...]
> > -1,  225640000,  225640000,  2127000,     2133, 0x670c11a5, F=0x0
> > +1,  225640000,  225640000,  2127000,     2133, 0x670c11a5
> >  0,       1139,       1139,        1,   518400, 0xbab197ea
> > -1,  227834000,  227834000,  1262000,     1264, 0xc1d9fc57, F=0x0
> > +1,  227834000,  227834000,  1262000,     1264, 0xc1d9fc57
> >  0,       1140,       1140,        1,   518400, 0xb046dd30
> >  0,       1145,       1145,        1,   518400, 0xbab197ea
> > 
> Have you checked whether all these changes are beneficial?

I think it's caused by the keyframe flags, so have no clue how to check
the beneficial for these fate testing.


> 
> - 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".

-- 
Thanks,
Limin Wang