siox and lab

Michael Thaler michael.thaler at physik.tu-muenchen.de
Wed Mar 29 11:04:00 CEST 2006


On Wednesday 29 March 2006 10:34, Boudewijn Rempt wrote:

> Or port the algorithm to use 16 bit integers? Or convert the lab shorts to

Probably doable, but I don't think it is a good idea. My experience with 
porting from gimp to Krita is, that it is best to change as few things as 
possible. Once you get it working, you can still optimize it.

The problem with the siox algorithm is that it is quite complex. That is 
nothing you understand in an afternoon. Probably someone who really knows the 
algorithm spent quite some time to get it working with gimp. If one changes 
the algorithm to use 16bit integers and at the end the whole thing doesn't 
work, you don't know if you did something wrong by changing the code to  use 
16bit integers or if something else is wrong.

I really want to port the algorithm from gimp and change as few things as 
possible. That will probably already be quite a lot of work. Once it works, 
we can optimize it. So maybe it is really the best idea to port the RGB to 
LAB routines from gimp. Then the tool will of course not work in anything 
else then RGB8, but making it colorspace independent can be done later once 
it works for RGB8 (and shouldn't be too hard because it uses lab internally 
anyway)

> float, but that would be costly. In any case: please first check the siox
> page itself, instead of just taking the code in the gimp.

That is probably a good idea, anyway:-)

Greetings,
Michael


More information about the kimageshop mailing list