Re: ICC user feedback request
Re: ICC user feedback request
- Subject: Re: ICC user feedback request
- From: Rudy Vonk <email@hidden>
- Date: Tue, 17 Apr 2001 18:03:56 +0200
I am not in the Henrik Holmegaard league of technical expertise, but
following his post forwarding Prof. Brues' invitation, which explicitly included:
>
Please don't hesitate to provide even 'low-level' kind of input
>
whenever you feel this is of importance for your individual color
>
management solution,
I would modestly float the following, fully expecting to be politely and
scientifically shot down by our more knowledgeable friends :-)
It seems to me mathematically feasible to take an image file's color
space (input profile) and contents (actually present RGB, CMYK or LAB
values), and calculate a new ad hoc "profile" that shrink-wraps around
the actual colors in the file. This could then be used if (and only if)
perceptual rendering intent is asked for, to reduce the compression that
would result from using the original input profile. This would ensure
that the full gamut of a smaller output color space would always be
available, whereas currently, colors that are outside the output gamut
are likely to fall well inside the output gamut's borders.
As an extreme example, let's assume we have a scan of an oil painting
containing exclusively tones of blueish grey and greyish blue. Suppose
it were in Ektaspace, where the blue corner is in a different universe.
Applying current perceptual rendering to an offset space would compress
everything into a "blue-gray hole". My suggested approach would
effectively turn perceptual into colorimetric and preserve the original colors.
My current workaround, on a file-by-file basis, is to check the gamut
warning in Photoshop with colorimetric, and base a decision on that.
This, however, is totally unworkable in a production environment.
--
__________________________________
Rudy Vonk
Oviedo, Spain
<email@hidden>
+34 607 354100
You can't always want what you get.
__________________________________