[MPlayer-users] 2 pass divx encoding from .VOB/dvd - Targeted filesize (bitrate) ends up way way wrong

[D Richard Felker III]

On Tue, Aug 20, 2002 at 09:55:40PM +0200, Michael Niedermayer wrote:
> [Automatic answer: RTFM (read DOCS, FAQ), also read DOCS/bugreports.html]
> Hi
> 
> On Tuesday 20 August 2002 02:46, D Richard Felker III wrote:
> > [Automatic answer: RTFM (read DOCS, FAQ), also read DOCS/bugreports.html]
> >
> > On Tue, Aug 20, 2002 at 01:17:37AM +0200, Michael Niedermayer wrote:
> > > > Actually, I believe there *is* a bug in libavcodec where it won't
> > > > properly distribute bits on the second pass when the minimum qscale
> > > > (vqmin) is already reached on lots of frames, so quality still won't
> > > > be ideal, but it should be a lot better.
> > >
> > > hmm, one possible bug is with the fps, lavc ratecontrol calculates the
> > > bitrate by counting the encoded frames & using the framerate but if this
> > > isnt constant or something then the bitrate could be wrong
> > >
> > > btw, its much easyer to fix bugs which i can reproduce, sadly i cant, and
> > > noone sended me a video which triggers it :( (yes i know that most of
> > > these videos are too huge to upload)
> >
> > I thought I uploaded a relevant sample from Ghost in the Shell, but
> hmm, i cant find it on mphq, if u did upload it, then tell me the filename ...
> 
> > maybe not. Let me know if you'd like me to try to reproduce it and
> > I'll see what I can do.
> yes, a sample which triggers the bug would be nice
> 
> btw, i did mostly rewrote the rc code in my local tree, its much more flexible 
> now and allso seems to be better than divx5-window$ on a few short 
> testsamples but it isnt bugfree yet :(