[FFmpeg-soc] [soc]: r5821 - libavfilter/vf_overlay.c

[Stefano Sabatini]

On date Tuesday 2010-06-08 15:21:07 -0700, Baptiste Coudurier encoded:
> On 6/5/10 4:19 PM, Stefano Sabatini wrote:
> >On date Saturday 2010-06-05 15:33:21 -0700, Baptiste Coudurier encoded:
> >>On 6/5/10 2:33 PM, Stefano Sabatini wrote:
> >>>On date Saturday 2010-06-05 14:02:32 -0700, Baptiste Coudurier encoded:
> >>>>Hi Stefano,
> >>>>
> >>>>On 6/5/10 6:38 AM, Stefano Sabatini wrote:
> >>>>>On date Friday 2010-06-04 22:15:35 +0200, bcoudurier encoded:
> >>>>>>Author: bcoudurier
> >>>>>>Date: Fri Jun  4 22:15:35 2010
> >>>>>>New Revision: 5821
> >>>>>>
> >>>>>>Log:
> >>>>>>Direct rendering in overlay filter.
> >>>>>>RGB24 support.
> >>>>>>Doesn't work with movie in movie yet, needs loop input feature for logos
> >>>>>>either in movie filter or here.
> >>>>>
> >>>>>Overlay filter is now badly broken in a weird funny way.
> >>>>>
> >>>>>May I ask you to avoid to commit features which cause regressions?
> >>>>
> >>>>What's broken exactly ?
> >>>
> >>>ffplay -vf "movie=0:png:logo.png, scale=100:-1 [logo]; [in][logo] overlay=10:main_h-overlay_h-10:1 [out]" slow.flv
> >>>
> >>>and sorry for the rude reply, now I see that you test only with ffmpeg
> >>>and not with ffplay (also the fade filter has serious problems when
> >>>used with ffplay).
> >>
> >>That's ok, it seems you improperly translated to the new API.
> >
> >100l to me.
> >
> >>It works fine here now.
> >
> >Anyway the problem is another one, try with a MPEG video based codec
> >and you should see. What I'm observing is that the overlay is applied
> >to the source filter *before* the motion compensation, the
> >(weird|funny) result is that the logo tends to stain the image
> >sort-alike ink as the video scrolls.
> 
> Stefano, do you still have the issue ?
> I believe the overlay filter should be commited to svn.
> Michael, could you please review it ? Or do you want me to send it
> to ffmpeg-devel ?

Yes, the problem depends on the ffplay implementation (which doesn't
rely on av_vsrc_buffer), I still didn't looked at it, that for sure
should be addressed before to commit it to SVN.

Also please post the patch anyway, there are some points I want to
address in the review (mainly trivial stuff).

Regards.