[MPlayer-users] detc filter and DVDs that switch from telecine to progressive

D Richard Felker III dalias at aerifal.cx
Thu Jul 17 02:35:09 CEST 2003


On Wed, Jul 16, 2003 at 04:37:34PM -0700, Corey Hickey wrote:
> [Automatic answer: RTFM (read DOCS, FAQ), also read DOCS/bugreports.html]
> Hello,
> 
> The detc filter works wonders for me when I feed it a video that is
> "hard-telecined" for the entire length. I've never quite understood,
> however, what I should do when I have a video that, for instance, starts
> out hard-telecined, then half an hour later switches to soft-telecine,
> then back to hard, etc.
> 
> Is there a way to recover nice, clean, 23.976 progressive video from 
> such a source? Whenever I specify -vf detc=dr=1 -ofps 23.976 , mencoder
> reports lots of "1 duplicate frame(s)!" as soon as it hits the
> soft-telecined material.

What's needed is either (a) a smarter filter made to handle this
material, or (b), a way to force mplayer to apply the soft-telecine
like a hardware player would, so that the inverse telecine filter can
then reverse it. Method (b) sounds rather dumb in principle, but it's
probably the best with mplayer's current architecture. Method (a)
might work once I finish the new inverse telecine engine I'm working
on, at least with MPlayer G2...

In the mean time, if your movie doesn't switch back and forth too many
times, you might be best off just cutting it down the middle at the
point where it changes, encoding the two parts separately, then
splicing the resulting avi files together....

Another thing you could try is omitting the dr=1 parameter to detc.
BUT, this could break telecine removal, since without dr=1, mencoder
will decide which frames to drop before showing them to the filters...
:( So if you try this method, be sure to check that the output stream
is actually correct, especially around any timestamps where mencoder
says it skipped frames!

Rich



More information about the MPlayer-users mailing list