[argyllcms] Re: Digital Camera input profiles and gamma correction.
- From: Nikolay Pokhilchenko <nikolay_po@xxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Mon, 13 Apr 2009 16:33:59 +0400
I'd recommend you to try gamma=1, and "-aS -ql" or "-ag -qu" parameters for
colprof. Just try and check wither it works with other software or not. I've
got poor results with -al (LUT-type profile).
It's very important to maintain the fixed white balance coefficients, fixed
brightness and gamma as for profiling, as for later workflow.
To check the resulting profile I suggest to use xicclu utility:
xicclu -v -ir -pj -fif -g profile.icc
xicclu -v -ir -pj -fb -g profile.icc
This command produces a graph with "source values over a gray scale".
Assumption can be maid that sensors are pretty linear and their behavior are
not warped. So the curves in graph must be smooth. If there are sharp bends in
xicclu results, the profile is more likely bad.
-----Original Message-----
From: Paul Hunt <prhunt@xxxxxxxxxxxxx>
To: argyllcms@xxxxxxxxxxxxx
Date: Mon, 13 Apr 2009 15:51:27 +1000
Subject: [argyllcms] Re: Digital Camera input profiles and gamma correction.
> Gerhard,
>
> Thank you for your reply. I created profiles using the -al option. Is there
> a better option for creating camera input files?
>
> Here are sample error reports from profile building:
>
> // Target imnge file created in ufraw with gamma 0.45 and linearity 0.1
> // Wolf Faust C1 target used
>
> profile check complete, peak err = 7.398101, avg err = 1.207893
>
>
> // Target imnge file created in ufraw with gamma 1.0 and linearity 0
> // Wolf Faust C1 target used
>
> profile check complete, peak err = 9.980332, avg err = 1.570878
>
> Both created profiles work quite satisfactorily in ufraw provided that I use
> the same gamma and linearity settings used to develop the target image.
>
> (Is there any reason to choose one profile over the other?)
>
> Unfortunately, the profiles do not seem to work well in other software (eg
> digikam, rawtherapee). For example, I can not get gamma curve correct, or I
> see strange colour artefacts.
>
> Is it the case that I should process target images using the software for
> which the profile is intended? That is to say, if I want profiles for ufraw,
> I should process the target image in ufraw and profile accordingly. If I
> wish to use digikam, I should first process the target image using digikam
> and profile accodingly.
>
> Thanks for your interest and advice!
>
> Paul
>
Other related posts: