[FFmpeg-devel] pixel format convertion yuv422p->yuv420p->yuv411p->yuv422p degrade picture

Michael Niedermayer michaelni
Tue Jul 15 19:26:59 CEST 2008

On Tue, Jul 15, 2008 at 05:44:10PM +0100, M?ns Rullg?rd wrote:
> Maksym Veremeyenko wrote:
> > Michael Niedermayer ???????(??):
> >> On Tue, Jul 15, 2008 at 04:22:33PM +0100, M?ns Rullg?rd wrote:
> > 50 iterations was only for proving that convertions degrade picture...
> >
> > So, as conclusion, that a normal?
> The conclusion is that it's probably not worthwhile doing anything about
> it.  It would make the code slower with no real benefit in practical use
> cases.

It should be possible to add some ordered dither in some of the SIMD
converters with no speed effect. Our RGB15/16 output already applies dither
as it does make a vissible difference.
In C it could be done by unrolling the code, but yes i agree it has no
practical benefit for 24bit->24bit. Then again if someone wants it and
theres no speed loss and its cleanly implementd ...
iam not opposed to reviewing and applying a patch to swscale.

That btw leads to the question if the greeniness is from swscale or the
old converter? And if SWS_ACCURATE_RND was used?


Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

I know you won't believe me, but the highest form of Human Excellence is
to question oneself and others. -- Socrates
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20080715/d5b7851d/attachment.pgp>

More information about the ffmpeg-devel mailing list