[argyllcms] Re: relativ volume of gamut -- vrml comparison?

  • From: "Preben Soeberg" <prsodk@xxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Sun, 6 Jan 2008 15:59:10 +0700

Thanks for all the answers.

There is one thing about the "camera" profile, I should have been more
clear about. The discussions above has made me aware about the
confusion, it has given rise to.

The actual profile, though tagged as an input profile, it is actually
not so. It is an abstract profile, used to convert the colors from the
demosaiced camera data, let's call these data's colorspace "Neutral".
So, in Canons DPP  (Digital Photo Professional) program, you can
choose between different abstract profiles (Normal, Portrait,
Landscape, Faithful, Monochrome) or "Neutral". You must also choose a
work color space (sRGB, AdobeRGB, WideGamutRGB, AppleRGB or

Without any knowledge of the camera profile giving the "Neutral" color
space, I assume the gamut of camera/neutral to be the the full RGB
cube and work on from here. The gamut (of the output we have access
to) can not be bigger than that.

Part of the confusion comes from the gamut volume output from iccgamut:
iccgamut -v 6071_F.icc
Total volume of gamut is 1065704.933829 cubic colorspace units
I took this number for granted, but now, I guess, this gamut comes
from converting the PCS in some wide gamut workspace and seeing if it
is within gamut here. (I have not tried to find it in the source code,
but it seems a plausible explanation for how to find "the" gamut for a
input device).

A comparison between AdobeRGB and ProPhotoRGB is shown here:
The colored area on the top is where AdobeRGB's gamut exceeds ProPhotoRGB's.
So my explanation why the Faithful\AdobeRGB gamut intersection exceeds
the Faithful\ProPhotoRGB gamut intersection is that Canon simply has
targeted AdobeRGB, when they made the profile for "Faithful".


Other related posts: