[FFmpeg-devel] [PATCH] avformat/mov: compute dts_shift with trun cts

[Clément Bœsch]

On Fri, May 31, 2013 at 05:19:16PM +0200, Alexandre Sicard wrote:
> Some movies have negative composition time offsets in their trun, causing pts <
> dts errors. This patch makes use of dts_shift to handle them.
> 
> Signed-off-by: Alexandre Sicard <alexandre.sicard at smartjog.com>
> ---
>  libavformat/mov.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/libavformat/mov.c b/libavformat/mov.c
> index 1c03b15..c43e6e3 100644
> --- a/libavformat/mov.c
> +++ b/libavformat/mov.c
> @@ -2562,6 +2562,9 @@ static int mov_read_trun(MOVContext *c, AVIOContext *pb, MOVAtom atom)
>          sc->ctts_data[sc->ctts_count].count = 1;
>          sc->ctts_data[sc->ctts_count].duration = (flags & MOV_TRUN_SAMPLE_CTS) ?
>                                                    avio_rb32(pb) : 0;
> +        if (sc->ctts_data[sc->ctts_count].duration < 0) {
> +            sc->dts_shift = FFMAX(sc->dts_shift, -sc->ctts_data[sc->ctts_count].duration);
> +        }
>          sc->ctts_count++;
>          if (st->codec->codec_type == AVMEDIA_TYPE_AUDIO)
>              keyframe = 1;

Do you have a sample for this?

This reminds me various issues we had with random samples. Check for
instance:
  e74527219668b720aefb7ecb79a44a8137aafa60
  be401448e51f209a13e95d30f17d85d0168e1b40
  4093220029a4d77f272c491e9299680480a08c00

Such patch needs to be tested with the files mentioned in those
commits/associated tickets.

-- 
Clément B.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 490 bytes
Desc: not available
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20130531/938db85d/attachment.asc>