RE: Finding CIELab Values from Target Chart TIF
RE: Finding CIELab Values from Target Chart TIF
- Subject: RE: Finding CIELab Values from Target Chart TIF
- From: Marco Ugolini <email@hidden>
- Date: Sat, 1 May 2010 14:32:59 -0400 (EDT)
Roger Breton wrote:
>Best is to study its repeatability over time yourself. I would take a small
>patch count chart and measure delta E's over the course of a day, week and
>month. In principle, if should gravitate around a certain set of values but
>I don't expect, as I did this to some extent for clients, that it would be
>all that stable. We did this on more expensive Xerox's, Canon's and
>Minolta's and found the profile we did fresh in the morning was no longer
>good in the afternoon, and no amount of recalibrating with the built in
>routines would ever get us back the color we'd see in the morning.
Hi, Roger and Joseph.
I wouldn't expect the linearization to be producing the same values in each and every instance. That may happen by accident, but not by design.
It seems to me that what a linearization procedure ensures is that gradations appear VISUALLY LINEAR, meaning that the tonal values for each primary, secondary, etc., be distributed on the final print in a way that appears even and smooth to the human eye.
I don't think that the linearization process must necessarily hit a given pre-decided set of values (though it possibly may, in some circumstances: the question should be posed to those who put together the linearization mechanisms in RIPs and the like).
With the ColorBurst RIP, for example, one can "re-linearize" after setting baseline values based on the initial linearization, then compare the fresh set of measurements to the initial ones. If the newer set of measurements exceeds the tolerances deemed acceptable by the RIP's internal mechanisms for accurate output, ColorBurst will alert the user. In that case, the user will have to build a wholly NEW profile from a newer set of linearization measurements.
So, my guess is that linearization results are not supposed to match any set of pre-established measurements. It's the task of the custom output ICC profile to match expected color appearances to what the printer (in its current state of linearization) is capable of producing from any given combination of color numbers in the source image file, properly color-managed.
Again, it's my sense of it that the linearization's role is solely to produces even and smooth-looking tonal distributions -- not to match any "expected" set of measurements.
Once that evenness and smoothness have been achieved, it becomes the custom output ICC profile's task to allow for accurate color output -- or at least as accurate as feasible within the existing workflow.
Marco Ugolini
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Colorsync-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden