[Rawstudio-dev] Any efforts to control moiré?

Klaus Post klauspost at gmail.com
Sun Jul 18 12:46:41 CEST 2010


Hi Martin

> Well, there's a clear difference in files processed with Rawstudio and
> Raw Therapee respectively. Damn, by comparison RS is slow and, at least
> the 3.0 alpha, unstable. But then it also seems that RT applies many
> more different filters.

Yes - that seems quite nasty. In general the demosaic does a pretty
good job, but it does seem to break down on high-detail/aliased areas,
when the camera itself does little to prevent aliasing.

I have filed a bug a while ago for a better high-quality demosaic
routine, which you can follow:

http://bugzilla.rawstudio.org/show_bug.cgi?id=290

We cannot copy the Rawtherapy demosaic routine due to license issues
(GPLv2 vs GPLv3) - but I'd like to re-implement it anyway to see if we
can squeeze more speed out of it.


> I've uploaded a raw (moire-test.nef) and RT's output
> (moire-test-rawtherapee.jpg). I chose this particular picture because
> RT's output shows very little moiré while Rawstudio's output has quite a
> lot.

Thanks - it is always good to have good test images! No test chart can
beat real-world examples :)


Regards, Klaus Post

http://www.klauspost.com



More information about the Rawstudio-dev mailing list