Re: profiles are odd ducks

[rafeb <rafeb@channel1.com>]

At 03:18 PM 12/29/99 -0500, Dan Culbertson wrote:

>God knows that this Dan (ie *me*) is not predisposed to believe that
>profiles are application independent.  From what I can garner from the ICC
>specification it only mandates the alphabet and maybe a bit of the grammer
>-- which in no way implies that everyone henceforth is going to speak the
>same language.  It is my conclusion that ICC profiles can only *safely* be
>used from within the application where they were developed (or for which
>they were expressly developed).  Which is one of the reasons that I like
>the MatchLock series of profile makers and profile editors -- but then *I*
>don't ever intend to take these profiles outside of Photoshop and use them
>elsewhere.  I do have great pity on the pre-press professionals struggling
>with using ICC profiles as a way to communicate with a print shop!  But
>that isn't my game, at least for now, and maybe things will standardize
>before I ever have to go there -- or maybe get worse?


<snip... of a typically excellent & eloquent Campfire Dan post...>


Goodness.  Are we seeing the first signs of ICC backlash?
Hehe.  So glad to have had a small part in it...
It's nice to see that we're discussing some fundamentals,
though.  Kudos to Richard Moyer also for starting off
that train of thought...

Ah, but seriously folks.  Think about the physics of it,
and what you're asking these profiles to do.  You're not
just transforming one axis to another with a simple curve...
it's a transform in at least three dimensions.  There's
not necessarily a mathematically-correct way of getting
from point A (say, some old monitor) to point B (say, some
kinda Epson printer, running some kinda inks on some kinda 
paper with some kinda driver or RIP at god-knows-what settings.)

Software can do marvelous things, but it cannot change
the laws of physics.  You will need a small miracle to
print a deep purple on an Epson, and that's a fact.

You're taking a rug that was nicely fitted to one N-
dimensional surface, and trying to stretch, shrink,
and re-cut it to cover some other convoluted N-dimensional 
surface.  There will be distortions, you better
believe it!



rafe b.

-
Turn off HTML mail features. Keep quoted material short. Use accurate
subject lines. http://www.leben.com/lists for list instructions.