[FFmpeg-devel] lavfi state of affairs
Thu Jan 22 21:55:37 CET 2009
On date Wednesday 2009-01-21 20:50:48 +0100, Stefano Sabatini encoded:
> On date Tuesday 2009-01-20 09:37:21 +0100, V?ctor Paesa encoded:
> > Hi,
> > Stefano Sabatini said:
> > > On date Monday 2009-01-19 21:04:13 -0800, Baptiste Coudurier encoded:
> > >> How is the situation, concerning full integration in svn ?
> > >> How useable the filters are ? I mean the situation seems to work for
> > >> you, and in many way it seems to be at least better than the current
> > >> vhook implementation. What's left for the merge ?
> > >
> > > The main issues is to fix the regression tests for ffmpeg, YUV tests
> > > seem to break regression, I still didn't investigate on it, mainly
> > > because I'm lazy and write filters is so much fun ;-).
> Do you have any hint regarding what could be the cause of it?
With the help of Vitor I discovered that the scale filter was using
different flags for the scaler, the attached patch, which is not meant
to be reviewed, fixes regression.
> > > For the rest I think lavfi is infinitely better than VHOOK, there are
> > > still some problems to be addressed (e.g. variable-size support,
> > > request-frame slicification) but nothing which can't be postponed to the
> > > FFmpeg SVN integration.
> > There is also a problem for filters that remove frames, as ffmpeg
> > uses/expects the fps of the input video (instead of the filtered one) and
> > generates duplicated frames to reach that.
> > For example, if input.avi has 25fps, then output.avi will also have
> > 25fps, despite the fps filter:
> > ffmpeg -i input.avi -vfilters 'fps=1' output.avi
> Another problem to signal is that slicification seems to work only
> with yuv420p (-vfilters "slicify, format=FMT").
FFmpeg = Fiendish Faboulous Monstrous Plastic Esoteric Guru
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 4171 bytes
Desc: not available
More information about the ffmpeg-devel