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: Roberto Michelena <email@hidden>
- Date: Thu, 25 Nov 2004 18:02:49 -0500
On 11/25/04 11:39 AM, "Roger Breton" <email@hidden> wrote:
> Actually, this idea of an underlying model is what surely permeates the
> better profiling package (me speculating, here). Case in point, PrintOpen.
> The latest version I use to build press profiles from has a unique Correct
> Measurement Data feature which I have not seen in any other profiling
> package to date. It does wonders on targets that have suffered uneven
> inking, you know, because of density variation across the sheet.
>
> This feature may not be construed as iterative. It is not, in fact. But I
> thought it had some bearing on the discussion as we touch on different
> algorithmic approach to profiling.
There is ALWAYS an underlying model. In some cases it's very simplistic:
"lets read a lot of patches and use trilinear interpolation between them"
In some cases it aims to really model the process and then rely on less
measurements. The first of such beasts was "Profile/80" derived from work
done at RIT, which as its name says claimed to profile presses with only 80
patches. And indeed, for presses it was not bad. But for inkjet it sucked.
Then came some profiler whose name I don't recall, delta-something, which
was only good for toner based devices, alas with very few patches.
And the latest in this game is Best's (now EFI) ColorProof Premium profiler
option, which uses also a very small number of patches to create the
reference (supposedly an offset press) profile. For the proofer profile it
still uses IT8.
Aurelon's spectral profiler also had an underlying model, in this case
predicting how the individual inks mixed together; so that you only measured
ramps of pure inks and 2c overprints, and it could build a profile.
But it had been developed on the basis of standard dye inks, so when
Ultrachrome & co. came along, it became useless. Ultimately they phased it
out and relied on ICC profiles instead.
You need two models, by the way: one for the reference (press) behaviour, so
you can properly interpolate between the patches that you used to build your
profile. And another one for the proofer, whose behaviour might be totally
different.
But CMMs don't offer choices of interpolation methods, as far as I know.
They in some way "assume" the behaviour of any device can be interpolated in
the same way. Which could be true if our profile grids were really dense,
but 33 points in each axis does not look so dense to me, specially when
dealing with proofers which have more than 4 colors and whose inks react in
such complex ways with the media.
What's the largest grid seen on an ICC profile, by the way? besides size and
manageability, are there limits built-in in the specs or in the various
software packages that deal with ICC profiles?
-- Roberto Michelena
Lima, Peru
_______________________________________________
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