[argyllcms] Re: Measuring a colorcheker on screen
- From: Kai-Uwe Behrmann <ku.b@xxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Thu, 24 Jan 2008 11:56:09 +0100 (CET)
Am 24.01.08, 11:07 +0100 schrieb "Gerhard Fürnkranz":
> -------- Original-Nachricht --------
> > Datum: Thu, 24 Jan 2008 09:03:01 +0100 (CET)
> > Von: Kai-Uwe Behrmann <ku.b@xxxxxx>
> > An: argyllcms@xxxxxxxxxxxxx
> > Betreff: [argyllcms] Re: Measuring a colorcheker on screen
>
> > It is not enough. The black point must be taken into accout.
> > A RGB 0,0,0 smaple on screen will have a XYZ different from 0,0,0.
> > CMM's should give XYZ 0,0,0 for ICC relative intent as it is
> > required for display profiles.
>
> Did I miss anything in the ICC spec? Could you please point me to the
> chapter/page in the spec where this requirement is stated?
ICC(v4.2) D 7.2 clears about display handling as a emmisive device:
"Thus, for practical reasons, it is important that the monitor profiles be
designed to display paper white at full-drive monitor white (R = G = B =
255 on a typical 24-bit display). Similarly, there is an expectation that
R = G = B = 0 corresponds to "black" and will be reproduced by the minimum
reflectance of the output medium. These user expectations are based on
common practice and convenience and lie outside of strict colorimetry and
colour-appearance considerations."
Or, a given PCS value of XYZ 0,0,0 shall result in a device RGB of 0,0,0
for display profiles. I would assume the same is iabove requested for the
other direction.
In the context of monitor verification, you did cut off from the privious
email, this is not sufficient as a CMM will provide not the measureable
XYZ values of the device but the 0,0,0 as required by the standard
rendering intents.
> Or do you just talk about the pseudo-intent "relative colorimetric with BPC",
> which could be computed by a CMM from the colorimetric tables in the profile
> (in order to avoid that very dark colors are clipped on the display)?
The ICC rendering intent selecton should play no role for display
profiles.
To check the monitor with the generated profile for a given RGB triple, a
given CMM must, which way ever, provide the XYZ intensities as strict
colorimetry. This is of course not handled by the assumptions in the above
citation.
I hope my words are a bit more clear about the device XYZ versus the
display PCS XYZ side ;-)
kind regards
Kai-Uwe Behrmann
--
developing for colour management
www.behrmann.name + www.oyranos.org
Other related posts:
