[Ffmpeg-cvslog] r6213 - in trunk: Changelog MAINTAINERS doc/ffmpeg-doc.texi libavcodec/Makefile libavcodec/allcodecs.c libavcodec/avcodec.h libavcodec/vp5.c libavcodec/vp56.c libavcodec/vp56.h libavcodec/vp56data.c libavcodec/vp56data.h libavcodec/vp5data.h libavcodec/vp6.c libavcodec/vp6data.h libavformat/flvdec.c libavformat/nsvdec.c libavformat/riff.c libavformat/swf.c

Aurelien Jacobs aurel
Fri Sep 15 01:24:02 CEST 2006


On Thu, 14 Sep 2006 00:28:10 +0100
M?ns Rullg?rd <mru at inprovide.com> wrote:

> Baptiste Coudurier <baptiste.coudurier at smartjog.com> writes:
> 
> > M?ns Rullg?rd wrote:
> >>>>>>>> [...]
> >>>>>>>> Is defining a new codec id the only way to identify VP6 in flash ? It
> >>>>>>>> seems it uses the same function to decode. Why not using CODEC_ID_VP6 ?
> >>>>>>>>
> >>>>>>>>  
> >>>>>>>>
> >>>>>>> The there is a horizontally flipped relation between them. And there
> >>>>>>> isn't a clean way of joining them to one codec id so the simplest
> >>>>>>> solution is to use 2 id's.
> >>>>>>>
> >>>>>>> MvH
> >>>>>>> Benjamin Larsson
> >>>>>>>
> >>>>>> Humm, it seems this issue has been discussed already. What was decided ?
> >>>>>> I can't get if it was decided to use a new codec id or not.
> >>>>>>
> >>>>>> Personnaly I don't like that VP6F codec id. I would be more in favor of
> >>>>>> setting a field or sub id or a flag in avctx, or maybe put those 2 bytes
> >>>>>> from the flv in extradata or something to identify it.
> >>>>>>
> >>>>>> What do you guys think ?
> >>>>>>
> >>>>> Nobody cares ? Oh well...
> >>>> I do.  If it is the same codec, it should use the same ID.  What was
> >>>> the difference again?
> >>> One is vertically flipped, the other is not.
> >>> And there is nothing in the codec bitstream to detect if it is flipped.
> >> 
> >> Time for a new flag, perhaps.
> >
> > I would be more in favor, also maybe using codec_tag from avi to handle
> > that special case (or is it flv) ?
> 
> A CODEC_FLAG_FLIPPED or similar would be much more generic.

Do you think about something like the attached patch ?
I'm still not sure how this could be handled in mplayer (and all other
player using their own demuxer)...
Also I wonder how we should handle the removal of CODEC_ID_VP6F ? To
be absolutely secure, a dummy entry should be added in the CodecID enum
to ensure binary compatibility.

Aurel
-------------- next part --------------
A non-text attachment was scrubbed...
Name: vp_flip.diff
Type: text/x-diff
Size: 7566 bytes
Desc: not available
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-cvslog/attachments/20060915/78c4dfc2/attachment.diff>



More information about the ffmpeg-cvslog mailing list