[MPlayer-cvslog] CVS: main/libmpcodecs vf_pp7.c, NONE, 1.1 vf.c, 1.115, 1.116 Makefile, 1.144, 1.145

Diego Biurrun diego at biurrun.de
Tue Jul 26 15:33:23 CEST 2005


On Tue, Jul 26, 2005 at 10:40:28AM +0200, Michael Niedermayer wrote:
> 
> On Monday 25 July 2005 20:49, Diego Biurrun wrote:
> > On Sat, Jul 09, 2005 at 09:49:53AM +0200, Michael Niedermayer CVS wrote:
> > > Added Files:
> > > 	vf_pp7.c
> > > Log Message:
> > > pp7 filter (spp=6 filter with 7 point dct where only the center sample is
> > > used after idct) these differences from spp lead to a few nice symmetries
> > > which significantly reduce the computational cost
> >
> > Michael, is this a correct description of your filter?
> 
> probably

OK, committed.

> >   pp7[=qp[:mode]]
> >     Variant of the spp filter, similar to spp=6 with 7 point DCT where
> >     only the center sample is used after IDCT.
> >
> >     <qp>
> >       Force quantization parameter (default: 0, use QP from video).
> >
> >     <mode>
> >       0: hard thresholding
> >       1: soft thresholding (better deringing, but blurrier)
> >       2: medium thresholding (default)
> >
> > Maybe you can fill in a comment about medium thresholding..
> 
> hmm 2. looked better then the others in my very limited tests :)
> 
> heres some technical description of what all 3 do with the dct coefficients
> 
> no     thresholding: 9 8 7 6 5 4 3 2 1 0
> hard   thresholding: 9 8 7 6 5 4 0 0 0 0 
> soft   thresholding: 6 5 4 3 2 1 0 0 0 0
> medium thresholding: 9 8 7 6 4 2 0 0 0 0

I've shortened this to 

2: medium thresholding (default, good results)

:)

Diego




More information about the MPlayer-cvslog mailing list