Re: Camera profiling with ICC et al
Re: Camera profiling with ICC et al
- Subject: Re: Camera profiling with ICC et al
- From: Uli Zappe <email@hidden>
- Date: Mon, 15 Sep 2008 23:18:07 +0200
Am 15.09.2008 um 22:57 schrieb email@hidden:
it can at least say that profiles that already show large color
deviations under the testing conditions are no reasonable
candidates for profiles to deliver color accuracy.
I draw nearly the opposite conclusion. In my experience, and that of
much of the technical community, camera profiles that are designed
to reproduce real scenes very well, reproduce targets less well. So
relatively large color errors on a target, does not disqualify the
profile for use with real scenes.
I don't really understand this. Let's assume a specific profile is way
off with respect to one specific color patch on the target. Now, it
might well happen (and in fact, did happen during my testing), that a
real scene contains an object of exactly this color. As a result, this
"real scene object" will have a wrong color on the image. How does
this not make the profile a "bad" profile for color accuracy?
The only possible counter-argument I can think of is to argue that the
color patches on a camera target are only a very very small and thus,
completely unrepresentative sample of possible colors, and that the
profile in my example therefore could be spectacularly exact in the
"remaining 16,699,999 colors" and as result, be an extremely good
profile overall. But how probable is this?
On the other hand, if the profile is so well optimized for targets,
then it probably wont be optimal for real scenes.
What I wrote was under the premise I articulated before (and to which
you agreed by writing "that may well be true") that for my test
images, I did find a clear positive correlation between small color
errors on the target and color accuracy for real scenes.
If a specific profile did *not* provide accurate color, in neither the
target measurements nor the real scenes of my test images, how could
it possibly become a "good" profile by another testing methodology?
Bye
Uli
________________________________________________________
Uli Zappe, Solmsstraße 5, D-65189 Wiesbaden, Germany
http://www.ritual.org
Fon: +49-700-ULIZAPPE
Fax: +49-700-ZAPPEFAX
________________________________________________________
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Colorsync-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden