[FFmpeg-devel] [FFmpeg-devel-irc] IRC log for 2010-03-28

[Carl Eugen Hoyos]

Michael Niedermayer <michaelni <at> gmx.at> writes:

> the problem is matroska + h264 in general, its not needed to be PAFF
> though that surely makes it worse

Am I understanding you correctly that the H264 timestamp issue we know of does
also exist for non-PAFF samples? Do you know the specific reference stream you
are referring to from out of your head?

> > (So the question is: Which "reports" do you mean?)
> 
> there was someone who submited a patch to do some very broken timestamp
> guessing in matroska*.c, this looked alot like an attempt to implement the
> timestamp calculation we are missing just completly wrongly at the complete
> wrong place.

Again:
This is issue 807, I am convinced it has nothing to do with H264 timestamps
(which I agree have bugs, but they are not related to the issue or the patch you
think of).

Carl Eugen