Re: [dvd::rip] Results of Linux vs. Windows benchmarking

On Thu, Jan 22, 2004 at 09:38:30PM -0800, it was written by Florin Andrei:

> > a) transcode's routines for resizing and deinterlacing eat up a lot more CPU
> >    than AviSynth's.  Especially at Pass 1 this is apparent -- I'd like to
> >    know why (wrong setup?)
> 
> Not necessarily related to resizing/deinterlacing, but try and tweak the
> -u parameter. In some cases, bumping up (at least) the first numeric
> argument increased the processing speed quite a bit for me. Don't
> remember details, though.

I'll give it a try, thanks.

> > d) Probably it's worthwhile to port AviSynth's resizing/deinterlacing over
> >    into transcode.  The AviSynth website states that VirtualDub's routines 
> > are
> >    even better optimized (they explicitely talk about resizing there even). 
> >    Unfortunately I'm not able to do that I'm afraid...
> 
> Make sure you're comparing similar algorithms. A faster resizer usually
> has poorer quality, or the other way round.
> E.g. see y4mscaler which performs very high quality rescaling, but it's
> not quite so fast.

For deinterlacing I don't have sufficient information about what exactly
AviSynth is performing there (I looked into the docs yesterday, maybe not
close enough :).  As for resizing, both implementations are using "Lanzcos
resizing", now I don't know how freely this algorithm is defined but just
assumed they probably both yield similar quality.  One would have to compare
them separately, with deinterlacing turned off.

--Frank

____________________________________________________
http://lists.exit1.org/mailman/listinfo/dvdrip-users
How To Ask Questions The Smart Way:
http://catb.org/~esr/faqs/smart-questions.html