[FFmpeg-devel] Corrupted blocks and seeking issues in H264 disc sources

Guillaume Poirier gpoirier
Sat Jul 7 15:40:42 CEST 2007


Hi,

Le 7 juil. 07 ? 01:23, infernix a ?crit :

> Hi,
>
> Thanks for the speedy color issue bugfix! Current SVN looks perfect  
> here.
>
> Unfortunately I've ran into a more severe case of the blocking I
> mentioned before, with current SVN for mplayer and ffmpeg. I have
> uploaded two samples from a blu-ray source to ffh264_blocking and they
> are also available at http://dx.infernix.net/ffh264/
>
> ffh264_issue5.264 contains blocks in frames 44, 48, 64, 68, 92, 124,
> 126, and 166 (using mplayer -vo png to generate the pngs). I've made a
> screenshot here: http://dx.infernix.net/ffh264/ffh264_issue5.png (the
> blue indicators point to the blocks in the other frames). Although  
> they
> seem minor, when playing back they are quite noticeable.
>
> ffh264_issue4.264 contains just one frame with a corrupted block,  
> shown
> here: http://dx.infernix.net/ffh264/ffh264_issue4.png
>
> I am still hunting this source for more blocks as I recall there's one
> occasion where it's even more severe, but I haven't been able to trace
> it yet.
>
> Aside from the blocking issues, neither of the samples are seekable,
> neither as ES nor in the m2ts container.
>
> Andreas, perhaps this new sample can give you more insight as to what
> the problem is? If I find more blocks, I'll upload them and post a  
> followup.


Could it be that your sample is corrupted so that it's normal that  
decoding shows artifacts? Do you have a way to check with other  
decoders if the same sample decodes OK there?

Guillaume



More information about the ffmpeg-devel mailing list