[Ffmpeg-cvslog] CVS: ffmpeg/libavformat oggparseogm.c, NONE, 1.1 Makefile, 1.129, 1.130 ogg2.c, 1.13, 1.14 ogg2.h, 1.4, 1.5

Måns Rullgård mru
Fri May 12 22:08:14 CEST 2006


Rich Felker <dalias at aerifal.cx> writes:

> On Fri, May 12, 2006 at 06:48:04PM +0100, M?ns Rullg?rd wrote:
>> Diego Biurrun <diego at biurrun.de> writes:
>> 
>> > On Fri, May 12, 2006 at 01:07:25AM +0200, M?ns Rullg?rd CVS wrote:
>> >> 
>> >> Modified Files:
>> >> 	Makefile ogg2.c ogg2.h 
>> >> Added Files:
>> >> 	oggparseogm.c 
>> >> Log Message:
>> >> ogm demuxing
>> >
>> > Excellent, we're getting closer to making libavformat a replacement for
>> > other demuxer libs one step at a time.
>> >
>> > Here is a sample that still has problems:
>> >
>> > http://www.mplayerhq.hu/MPlayer/samples/ogg/basicinstinct.ogm
>> >
>> > There is no sound, only the the following message:
>> > [vorbis @ 0x102b6138]Extradata corrupt.
>> 
>> I've figured that one out.  The vorbis extradata is split across three
>> packets, and the video data starts after the first of these.  This is
>> in violation of the ogg spec, which states that all extradata of all
>> streams must precede any non-extradata packets of any stream.  I guess
>> we should still deal with it somehow.  For some reason, it works with
>> tcvp.  I'm trying to figure out why.
>
> I suppose since ogm is a broken, ill-specified format anyway we should
> deal with it, though it would be better in principle to reject

Quite.

> malformed files. :( I wonder what kind of broken crap generated this
> file..?

I wonder what kind of broken mind got the idea to use ogm in the first
place.

-- 
M?ns Rullg?rd
mru at inprovide.com





More information about the ffmpeg-cvslog mailing list