[FFmpeg-devel] [PATCH] Only output necessary NAL units in H.264 extradata in SDP

Baptiste Coudurier baptiste.coudurier
Sat Apr 4 06:36:57 CEST 2009


Jason Garrett-Glaser wrote:
> On Fri, Apr 3, 2009 at 9:37 AM, Michael Niedermayer <michaelni at gmx.at> wrote:
>> On Fri, Apr 03, 2009 at 07:25:06PM +0300, Martin Storsj? wrote:
>>> Hi,
>>>
>>> As noted by another user a month ago
>>> (https://lists.mplayerhq.hu/pipermail/ffmpeg-devel/2009-March/064219.html),
>>> the SDP data generated by ffmpeg for H.264 streams doesn't work with
>>> QuickTime Player. I'm not sure if this is a change in recent versions of
>>> QuickTime or a regression in ffmpeg, though.
>>>
>>> QuickTime seems to dislike that the sprop-parameter-sets contains the
>>> version and identification output from x264 (which is a quite large chunk
>>> of data compared to the SPS and PPS parts). I also tested to shrink the
>>> amount of info output by x264 into this nal unit, but even with a minimal
>>> amount of data, it wouldn't play the content. Reordering the nal units (in
>>> x264), putting this unit as the last one, didn't help either.
>>>
>>> The attached patch solves this in two possible ways, please comment on
>>> which one you find more reasonable.
>> checking nal_type IMHO
>>
>> but first question, what does the RFC say? is what x264 stores ok in SDP or
>> not?
> 
> Why not just stuff the SEI in the bitstream itself?
> 
> There is currently a rather nasty bug in the mp4 muxer where the x264
> SEI simply gets lost--which makes it almost impossible to identify
> streams for debugging.

I object, this is not a bug, it's a feature ;)
Patch welcome.

-- 
Baptiste COUDURIER                              GnuPG Key Id: 0x5C1ABAAA
Key fingerprint                 8D77134D20CC9220201FC5DB0AC9325C5C1ABAAA
checking for life_signs in -lkenny... no
FFmpeg maintainer                                  http://www.ffmpeg.org



More information about the ffmpeg-devel mailing list