On 11 March 2015 at 15:10, Clément Bœsch <u...@pkh.me> wrote: > On Wed, Mar 11, 2015 at 07:12:16AM +0000, Carl Eugen Hoyos wrote: > > Timothy Gu <timothygu99 <at> gmail.com> writes: > > > > > > This is an exact inverse of the telecine filter > > > > unlike previously existing pullup and fieldmatch > > > > ones. > > > > > > What's the difference between the three seemingly > > > similar filters? > > > > The other two are similar (speed and license are > > different): They try to detect the telecined > > frames and reconstruct the original progressive > > stream even if cuts were made. They sometimes miss > > telecined frames. > > This filter does not detect anything, it applies a > > fixed pattern to the input stream: It fails very > > badly if the input was cut (and the pattern > > changed), > > > it works perfectly for streams that > > contain a constant pattern. > > Did this happen even once with real material? > > We aim to have it work perfectly for such streams. It has not been tested on a proper telecine stream yet. As the description mentions, this work is still under progress. I am collecting suitable test samples and would like to discuss the algorithm for any suggestions. Should I add the algorithm in the description to gain feedback on it?
> > > > Carl Eugen > > > > -- > Clément B. > > _______________________________________________ > ffmpeg-devel mailing list > ffmpeg-devel@ffmpeg.org > http://ffmpeg.org/mailman/listinfo/ffmpeg-devel > > _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel