Reimar Döffinger <[email protected]> added the comment:
On Tue, Aug 25, 2009 at 01:53:00PM -0700, Baptiste Coudurier wrote:
> On 08/25/2009 02:12 PM, Reimar Döffinger wrote:
> > Reimar Döffinger<[email protected]> added the comment:
> > On Tue, Aug 25, 2009 at 08:34:56PM +0000, Ronald S. Bultje wrote:
> >> Ronald S. Bultje<[email protected]> added the comment:
> >>
> >> Passing stsd isn't going to make it remuxable to other codecs...
> >
> > I would, though it might be a non-standard format (if there exists
> > a standard format for other containers, in that case IMO
> > it would be the mov demuxer's (or whichever's format is in the
> > minority) responsibility to do the conversion.
> > Apart from that, providing the whole atom would "fix" it in so
> > far as at least the information is available.
> > Detecting which format is actually used and converting between
> > them would still have to be implemented in MPlayer.
> > How hard this is depends on whether the mov demuxer guarantees
> > e.g. that the atoms in the extradata are in a certain order, if
> > a specific one will always be the first one etc.
>
> I was thinking that a "container_data" field in AVCodecContext or a
> "codec_data" field in AVStream could be introduced, used to store the
> full stsd atom, or the full bmpheader or waveformat header from avi/wav,
> or any structure used by the container to describe the codec.
Sounds to me like extradata just with the "advantage" of making remuxing
impossible. IOW I'd consider it the opposite of what FFmpeg supposed to be
there for ("normalizing" container data onto a common format).
_____________________________________________________
FFmpeg issue tracker <[email protected]>
<https://roundup.ffmpeg.org/roundup/ffmpeg/issue1322>
_____________________________________________________
