[FFmpeg-devel] Gaining access to a encoder context in avformat?

[Ronald S. Bultje]

Hi,

On Fri, Feb 20, 2015 at 6:56 AM, Kevin Wheatley <kevin.j.wheatley at gmail.com>
wrote:

> On Fri, Feb 20, 2015 at 11:44 AM, Michael Niedermayer <michaelni at gmx.at>
> wrote:
> > If you start with a existing movie and copy the packets one by one
> > there is no decoder and no encoder, so no dnxencoder structure
> >
> > if you want to put some value in the avid atom which are stored in the
> > bitstream/packets of dnxhd and not in the generic data structures
> > like width/height would be then you have to extract these from the
> > dnxhd bitstream one way or another because thats the only thing thats
> > there in the memory of your machine. I dont know what exact information
> > you want/need from the dnxhd encoder ...
>
>
> hmm, maybe I should just say that copy is not supported for DNxHD
> files then as that sounds a pain to deal with


Why wouldn't the mov demuxer parse these atoms and fill them in the same
way the dnxhd encoder does? The idea is that anything that outputs encoded
dnxhd frames should fill these data structures (where-ever they are, but
should probably be in some lavformat structure, not a lavcodec structure),
and for backwards compat you should also deal with the fact that they might
not be there.

Ronald