[MPlayer-users] Why is it "Skipping frame!"s?

[D Richard Felker III]

On Sat, Feb 07, 2004 at 05:54:45AM +0100, Amadeus wrote:
> > > for i in 1 2; do
> > > for j in /dev/null filename.avi; do
> > >          nice -19 mencoder -aid 128 -oac pcm -ovc lavc \
> > >      -lavcopts vcodec=mpeg4:mbd=2:trell:vbitrate=509:vpass=$i \
> > >      -sws 2 -vf crop=672:316:24:130,scale=544:240 \
> > >      -dvd-device some_image.bin dvd://1
> > >      -o $j
> > >  done
> > > done
> >
> > ROTFL, this will do both passes twice! :) Not quite what you had in
> > mind....hehe.
> 
> Argh. You are right. A little nasty hack takes care of that=)
> 
> i=1
> for j in /dev/null filename.avi; do
>          nice -19 mencoder -aid 128 -oac pcm -ovc lavc \
>      -lavcopts vcodec=mpeg4:mbd=2:trell:vbitrate=509:vpass=$i \
>      -sws 2 -vf crop=672:316:24:130,scale=544:240 \
>      -dvd-device some_image.bin dvd://1
>       -o $j
> i=2
>   done
> 
> 
> > Anyway the hd access time won't really have much impact unless you're
> > encoding _very_ low resolution and low quality at high bitrates. I
> > usually like to be able to preview the first pass to make sure a/v
> > sync is ok and stuff before spending time on the second. If you really
> > want your method to work, though, you could change it to -o $i.avi and
> > make 1.avi a symlink to /dev/null... :) Or make a fancier script.
> 
> What if the first pass was made with -nosound. Would that effect the
> divx2pass.log file?

This will make the 2pass process basically worthless. In theory, you
_should_ be able to use -oac pcm during the first pass, but due to
horrible bugs in mencoder's a/v sync code, this will result in
slightly different frame skips/duplicates than with mp3lame, so it's
still bad. Basically, mencoder totally sucks on the inside; it's a
miracle that it roughly works most of the time. Wait for
mencoder-g2... :))))

Rich