[MPlayer-DOCS] CVS: main/DOCS/man/en mplayer.1,1.515,1.516

[Michael Niedermayer]

Hi

On Tuesday 06 January 2004 06:46, D Richard Felker III wrote:
> On Tue, Jan 06, 2004 at 12:44:53AM +0100, Michael Niedermayer CVS wrote:
> > Update of /cvsroot/mplayer/main/DOCS/man/en
> > In directory mail:/tmp/cvs-serv1019/DOCS/man/en
> >
> > Modified Files:
> > 	mplayer.1
> > Log Message:
> > ildctcmp, VSAD, VSSE
> >
> >
> > Index: mplayer.1
> > ===================================================================
> > RCS file: /cvsroot/mplayer/main/DOCS/man/en/mplayer.1,v
> > retrieving revision 1.515
> > retrieving revision 1.516
> > diff -u -r1.515 -r1.516
> > --- mplayer.1	5 Jan 2004 21:03:09 -0000	1.515
> > +++ mplayer.1	5 Jan 2004 23:44:51 -0000	1.516
> > @@ -4608,6 +4608,9 @@
> >  use quarter pel motion compensation
> >  Tip: This seems only useful for high bitrate encodings.
> >  .TP
> > +.B ildctcmp=<0\-2000>
> > +comparison function for interlaced dct decision
> > +.TP
> >  .B precmp=<0\-2000>
> >  comparison function for motion estimation pre pass
> >  .TP
> > @@ -4637,6 +4640,10 @@
> >  rate distortion optimal, slow
> >  .IPs "7 (ZERO)"
> >  0
> > +.IPs "8 (VSAD)"
> > +sum of absolute vertical differences
> > +.IPs "9 (VSSE)"
> > +sum of squared vertical differences
>
> Could you explain these? I'm just interested because it seems like
> deciding whether a MB is interlaced or not for doing ildct is similar
> to making the decision for deinterlacing/inverse telecine purposes.
VSAD:
 sum x=0..15,y=0..6 (abs(pixel(x,y) - pixel(x,y+1)))
VSSE:
 sum x=0..15,y=0..6 ((pixel(x,y) - pixel(x,y+1))^2)

the decision is than based upon the comparission of the scores of the 
progressive vs interlaced blocks
compare(0,0,stride) + compare(0,8,stride) < compare(0,0,2*stride) + 
compare(0,1,2*stride) + threshold

and for inter MBs pixel is the difference which will be encoded not the actual 
pixel value

btw, maybe using abs(pixel(x,y) - 2pixel(x,y+1) + pixel(x,y+2)) would be 
interresting, it wasnt any better for the ildct decision though

[...]
-- 
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