QTR-Quadtone RIP

On Dec 26, 2007 6:04 PM, Ernst Dinkla <E.Dinkla@...> wrote:

> ferdinand_paris wrote:
> > --- In QuadtoneRIP@yahoogroups.com, "Roy Harrington" <roy@...> wrote:
> >> The "big dMax loss" is just one symptom that happens to be most
> >> visible. .... How we actually know whether any profile is done
> >> "correctly"?   I admit the dMax issue shows the problem, but other
> >> color profiles "could" be done wrong and we might not know. ....
> >> It's very weird to me that different CMSs with exactly the same
> >> input can produce different results.
> >
> > I am prepared to accept that different CMSs will produce slightly
> > different profile conversions.  While it would be nice to think of
> > conversions as being something that should be mathematically precise,
> > it sounds to me (as a non-expert) that in practice it's a little like
> > the differences between RAW converters - there are matters of
> > judgement involved.  But large differences surely indicate that
> > something is wrong somewhere.
>
> I will check the 3 color engines that I have in Photoshop:
> ACE, ICM and Kodaks's CMM to see whether that gives
> different results.
>

I'm pretty I've tried the various CMMs with Photoshop in the Mac (Adobe Ace,
ColorSync, Apple CMM)
and got the same results.  There's seems to be a lot of lore about that says
different CMS's can
work differently.   How can this go along with printer companies, paper
companies providing
print profiles that people can use with different products?  Do Epson
profiles only work with Photoshop?

Consider how profiles are made -- you print a target WITHOUT using CMS,
right?
Then measure it and create a profile.  The CMS had NO part in the profile
creation so how
could the profile be in any way tied to a CMS?

I just don't get how this all is supposed work if there's no "standard" that
everybody has to meet,
but I suspect there's more to it (there usually is) and I don't know what it
is I don't know about :)
I think B&W is simple enough to have a fixed way to work but color has too
many tradeoffs.
So different products can choose different tradeoffs.

Roy


[Non-text portions of this message have been removed]