On Wed, Jul 16, 2014 at 05:31:47PM +0200, Clément Bœsch wrote: > On Wed, Jul 16, 2014 at 05:06:27PM +0200, Clément Bœsch wrote: > > Also add an example exporting the MB information as a CSV stream. > > > > --- > > > > A bunch of questions stand: > > > > - Ideally the "source" for a given macroblock should be a specific > > reference to a future or paste frame (something like -4 or +2). > > Currently it's just ±1 depending on the direction. I don't see how I > > can extract that information. > > > > - how the MB "type" should be exported? Like, what "generic" type we > > need to expose for the users > > > > - Who is motivated to port -vismv & various other vis_* debug to a video > > filter? (The hard part will probably be writing its documentation...) > > > > Ah and BTW, can we make the decode step faster somehow in the example > since we're only interested in the MB and not the final picture? I was > thinking of lowres, but maybe we can even skip the whole decode step with > some of the codecs?
mc / idct could be skiped, we have a AVCodecContext.skip_idct also partitioned mpeg4 would allow decoding mb types and mvs without ac coeffs but i guess its rare enough that supporting that isnt worth it [...] -- Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB While the State exists there can be no freedom; when there is freedom there will be no State. -- Vladimir Lenin
signature.asc
Description: Digital signature
_______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
