[MPlayer-users] -vop cropdetect weird values

[Bohdan Horst]

On Wed, May 15, 2002 at 10:47:15PM +0200, Arpi wrote:
> [Automatic answer: RTFM (read DOCS, FAQ), also read DOCS/bugreports.html]
> Hi,
> 
> > I've tried the -vop cropdetect and get some strange values :
> 
> what is strange with these? it looks ok to me
> 
> > crop area: X: 0..479  Y: 92..482  (-vop crop=478:390:0:92)0.048 [710:64]
> > crop area: X: 0..479  Y: 92..482  (-vop crop=478:390:0:92)0.048 [718:65]
> > crop area: X: 0..479  Y: 92..482  (-vop crop=478:390:0:92)0.048 [722:66]
> > crop area: X: 0..479  Y: 92..482  (-vop crop=478:390:0:92)0.052 [728:67]
> > 
> > if I take 0 instead of 478 then its very nice , I think you should now this 
> 
> why?
> what is the original dimensions of the movie, and what are the good cropping
> values ?
> 

in my tests:
*** [cropdetect] Exporting mp_image_t, 768x576x12bpp YUV planar, 663552 bytes                       

results:
crop area: X: 0..767  Y: 81..493  (-vop crop=766:410:0:82)

and imho should be -vop crop=768:412     (x+2,y+2)


anyway when X size ist'y  8 multiply  Xv  output is crap ;) (-vo x11
works fine)

(same effect with -vop scale - maybe this is my grahics card/driver
"feature" ..)

best regards


-- 
 / irl:Bohdan 'Nexus' Horst | mailto:nexus at irc.pl | irc:Nexus \
{---------------------------^----------v----------^------------}
 \ http://www.physd.amu.edu.pl/~nexus/ | Instytut Fizyki  UAM /