[XviD-devel] Scrolling motion problem..

[Michael Niedermayer]

Hi

On Friday 21 May 2004 00:33, Michael Militzer wrote:
> Quoting "Hans K. Rosbach" <hk at circlestorm.org>:
> > For some reason the initial mail i sent to the list did not arrive before
> > after
> > 2 days..  Hopefully thisone will be faster.
> >
> > > it seems like the 'jerkyness' may come from b-frames not being
> > > correctly decoded (or rather: not decoded at all). You describe that as
> > > soon as a fade occurs, the movie plays smoothly again - well, it may be
> > > very
> >
> > likely
> >
> > > that no b-frames are used during that fade and that therefore the video
> > > plays smoothly again.
> >
> > That is most probably it.
> >
> > > So now the next thing we need to figure out is: what do you use to play
> > > back this video? Are you sure that XviD is used for decoding (and not
> >
> > DivX,
> >
> > > ffdshow or something else)? What happens if you open the video in
> > > VirtualDub - still jerky? Did you use the packed bitstream option for
> > > encoding the clip or not? Depending on what you've selected so far, try
> > > en- or disabling packed bitstream.
> >
> > I use ffdshow latest alpha..  Please don't say I cannot use that.  =(
> > I simply don't know of any other codec that gives me that level of
> > control.
> >
> > At least this explains the choppyness I have been experiencing lately
> > when watching certain videoclips. I was thinking it was an HDD related
> > problem and dismissed it as such.
>
> The 'latest' ffdshow alpha is way too old. You need a newer version if you
> want to continue to use ffdshow. Try one of the newer Athos builds:
>
> http://athos.leffe.dnsalias.com/
>
> BTW: imho the decoding quality of XviD files using the native XviD decoder
> + postprocessing is better than using ffdshow (but ffdshow is faster, yes).
> But all this surely is a matter of personal taste...
could u elaborate on this? 
are u saying that the postprocessing in ffdshow is worse then the 
postprocessing used in xvid? if thats the case then the obvious fix would be 
that someone ports the xvid pp code to ffdshow
the mpeg4 decoding itself shouldnt look different as the spec requires it to 
be binary identical except the IDCT, but IIRC (i dont use windows or ffdshow) 
ffdshow supports the libmpeg2 IDCT which is AFAIK also used in xvid, so both 
xvid and ffdshow without postprocessing should have binary identical output, 
unless one is buggy of coarse ...

[...]
-- 
Michael
level[i]= get_vlc(); i+=get_vlc();		(violates patent EP0266049)
median(mv[y-1][x], mv[y][x-1], mv[y+1][x+1]);	(violates patent #5,905,535)
buf[i]= qp - buf[i-1];				(violates patent #?)
for more examples, see http://mplayerhq.hu/~michael/patent.html
stop it, see http://petition.eurolinux.org & http://petition.ffii.org/eubsa/en