Re: Real world experience w/ GMG and Oris RIPs
Re: Real world experience w/ GMG and Oris RIPs
- Subject: Re: Real world experience w/ GMG and Oris RIPs
- From: Mike Eddington <email@hidden>
- Date: Thu, 25 Nov 2004 18:33:23 -0500 (EST)
- Organization: North American Color, Inc.
<I don't think theorizing on a product one actually does not have is bad at
all>
nor do I
<It's quite a pity that GMG does not have a demo version (time limited or
maybe limited to printing some specific files only) of its software; that
would allow Graeme, me and many others to test it and make more conclusive
comments.>
I agree
<But I believe, as Graeme pointed out, that the "secret sauce" is more in the
underlying model (which would reflect in the interpolation algorithm) than
in the iteration. And also in the fact that usage of link profiles, by
interpolating in the same space of the end result (CMYK in this case), can
increase the fitness of the interpolation algorithm compared to
interpolating in an intermediary space (Lab or XYZ) which has different
properties.>
So a comparison between a software using link profiles (GMG) and an one that uses an intermediary space (ICC based) wouldn't be a fair comparison...but it would be a "real world" comparison for someone shopping for a solution.
<As an example, there is an ICC profiling program that does (or at least did
in its previous versions) iteration also: Scitex Profile Wizard.
>From 1998 (at v1) it had the ability to select a file of target values, and
printing through the profile, read the result and iterate to minimize dE. I
remember however they recommended stopping below a certain value (5 dE
average? not sure) because "further iteration would result in actually worse
match".>
Actually, during the initial printer calibration GMG does recommend after the first or second iteration to increase the delta E limit from "0" to ".5-1" to prevent recalculating an already ideal value. One could do the same during color profile creation. You can also observe during iteration of the color profile the max values, those already as close as they will get, bounce around a slight amount as calculations try to dial it in closer. I would say the value changes .5 -1 delta E between iterations, so it doesn't actually get too much "worse".
<Nothing prevents any manufacturer of ICC profiling software to build
iteration into its product; actually it's easy! but really, tightening the
match of an arbitrary testchart, when you don't have any control of the
interpolation (because it's a CMM made by someone else, working in Lab or
XYZ) could actually harm the overall performance of the profile.>
Fair enough, but this doesn't describe GMG as they control the interpolation.
<Tightening -by iteration- the match of your profile's output gridpoints,
however, should provide nothing but benefits. I don't know why they don't do it!>
So are you saying that iteration in some circumstances has merit. Could you elaborate (i.e.dumb down ;-) ).
regards,
mike
_______________________________________________
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