[FFmpeg-devel] Remaining problems in H.264 handling
Fri Mar 27 19:56:59 CET 2009
I changed the topic of the thread, since it doesn't have to do anything
with swscale anymore...
Carl Eugen Hoyos wrote:
> Ivan Schreter <schreter <at> gmx.net> writes:
>> I provided a patch for this, which wasn't accepted by Michael.
> Could you point me to the patch?
> You had sent quite a lot...
Well, the original patches didn't work anymore, so I attach a new one,
just for testing (doesn't increase minor of lavf/lavc).
>> BTW, can you give me a link to a sample which doesn't play properly, so
>> I can look at it (if/when I get some time)?
> Note that I do not know of any PAFF sample that does not work fine with ffmpeg
> and ffplay. (I did not test AVCHD samples you mentioned, only the sony 9sec one.)
> For MPlayer:
> But I do not know of any PAFF sample that plays with currect MPlayer svn.
Looking at the samples, yes, this is the case of MPEG-TS file with field
pictures (at least the second one), which declares double frame rate. My
patch should fix it, since frame rate is detected correctly then (though
I don't have time to test with mplayer).
The first sample seems badly broken at the beginning, so our framerate
detection routine is confused. Possibly, there are also some other issues.
IMHO, frame rate for H.264 (and probably also MPEG video) should be
always set to 1/2 tbc (and set to reliable, if timing info is provided),
since this maps the best to all possible combinations of picture
structures. This would also most probably make the first sample work
without problem. But I didn't do enough convinction work to convince
Michael yet ;-).
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 3984 bytes
Desc: not available
More information about the ffmpeg-devel