Re: LAB as working/interchange space
Re: LAB as working/interchange space
- Subject: Re: LAB as working/interchange space
- From: Henrik Holmegaard <email@hidden>
- Date: Fri, 25 May 2001 05:33:14 +0200
The total LAB space is huge -- and sRGB takes up only a small
percentage of the LAB space. So converting sRGB to LAB means you're
filling up only that percentage of the LAB space covered by sRGB. So
16.7 million distinct sRGB colors get converted to 2.47 million
distinct LAB colors.
At the ECI session at FOGRA where the user feedback scheme first came
up, I blurted out something about folks sitting down and figuring out
a remake of ICC Lab that had a reasonable size, given the actual
output gamuts we have. ICC printer profiles are saddled with an
unreasonable job at the moment, and as David pointed out some time
ago, Adobe RGB has a corner point somewhere that goes outside ICC
Lab, so I wonder how the printer profile is supposed to handle that?
IMO it's better to balance the size of ICC Lab for actual gamuts
(defining a 'little Lab' or 'internal Lab'), and the size of RGB
working spaces for the size of an ICC 'little' Lab (whether anybody
uses it as working space or not I suppose is another matter). There
are arguments for RGB working spaces and there are arguments for Lab
as working space. I can follow both chains of argument, but surely
the relative sizing of the three dimensional working spaces to each
other, and the three dimensional working spaces to the four
dimensional output spaces is worth a thought.
I'm looking at this in terms of my favourite angle, proof reading.
There is an inconsistency which is conceptual, so get folks to adress
the conceptual inconsistency, and things become more consistent, not
just for the technical developers pushing whatever envelopes they
want to push, but at the end of the day also for all those out in the
tranches who are trying to figure out how to manage color -:).