2007.10.23 21:38 "RE: [Tiff] Floating-point 32bit sample data", by Chris Cox
HDR and CIE L*a*b* really don't mix.
If you have HDR data, stick to grayscale, RGB or XYZ.
HDR data _really_ needs to be gamma 1.0, and floating point image data is best represented as gamma 1.0.
Chris
-----Original Message-----
From: tiff-bounces@lists.maptools.org on behalf of Kai-Uwe Behrmann
Sent: Tue 10/23/2007 2:02 PM
To: Joris
Cc: libtiff - Liste
Subject: Re: [Tiff] Floating-point 32bit sample data
Am 23.10.07, 19:21 +0200 schrieb Joris:
Are floating point RGB data stored gamma corrected or linearized?
The data is best stored in a linear matrix space, much like camera RAW. OpenEXR even suggests now to use XYZ instead of RGB.
Only, there isn't a photometric value for it.
RGB, without embedded ICC profile, is mostly agreed upon to mean sRGB. Another and probably more unambigious way to store color data, even if I feel RGB is regarded sRGB by just about anyone, is CIE L*a*b*, and there's a proper photometric for that one and a totally unambigious scale and meaning for floating point CIE L*a*b* data.
How would you handle the different gammas in CIE*Lab for typical floating point HDR content?
Of course you are right to point out, that a ICC profile is needed to give a meaning to the image colours.
kind regards