Re: Posterisation & burnt out highlights
Re: Posterisation & burnt out highlights
- Subject: Re: Posterisation & burnt out highlights
- From: Steve Upton <email@hidden>
- Date: Tue, 18 Sep 2001 15:03:59 -0700
At 4:31 PM +0000 9/18/01, email@hidden wrote:
1] When making a cmyk separation from a known good RGB source file, what would
make the destination profile cause posterisation and / or burnt out
highlights?
Is this a problem that can be edited out?
The device behavior or the profile could cause the problem. A linear
device is a good thing and the first thing I would test. Then I might
look at patch numbers in the targets - there may be too few - and
then perhaps measurement problems. Device first though.
2] When a softproof is not a softproof! Three years into the game
and one thing
you think you have totally tied up is printing an acurate representation of
what is on the screen, even if it's poor! However, not today :-P I've
encountered three instances, one of which I can explain, and two
which I can't.
So Here goes.....
2.1] I was a little disturbed when my printed blues & greens were far "better"
than the screen representation. However, a simple gamut analysis tool made it
clear that I was actually producing printed Blues and Greens outside of the
gamut of my monitor (didn't actually expect that!) thus, the "softproof" is
totally inaccurate. That's the one I figured out!
Yes - a moment of realization isn't it?
2.2] What could possibly cause a gradient to separate out (via a profile)
smoothly on screen, and yet print with a solid band from say 70% to 100% dot?
A couple of possible things. If you are viewing the CMYK through the
CMYK profile on screen - so you have generated the CMYK already -
then you are at the mercy of the proofing transform of the profile.
It may or may not be describing the device behavior well. The problem
between 70% and 100% could be caused by relative colorimetric tables
without black point compensation, or some other rendering
intent-based problem. Still, the disconnect between screen and print
is the issue here isn't it.... hmm...
2.3] Why would a printed output print extremely "washed out" and yet separate
out to the correct ink weights when softproofing. This happened whilst testing
some new profiling software and a DTP41. Just to add confusion, the same
package with a different instrument separated correctly on screen (as did the
DTP41 profile) and yet it printed as displayed!? I know it isn't the DTP41 at
fault as I use this successfully day by day as I have done since.
There are several things here which may contribute to the confusion.
Surface reflection can cause an instrument to read very light
shadows. This can build a profile that proofs with those light
shadows. Now the rendering side of the profile (Lab->CMYK) might
still render out to the full 0-100% range and give good color, it
just doesn't proof right. Says the profile "no matter how much I
throw at this printer, I know I still won't get much better than
about L:25 for dark". This might be the exact inverse of what you are
dealing with (as I review your note above) but I think it is along
the lines of the problem... it is bit of a mystery.
If you get a chance, please forward the profile, Lab measurement
file, and CMYK reference file to me. I am documenting exactly this
style of testing in the new version of ColorThink and would
appreciate the chance to try it on a mysterious problem.
Regards,
Steve Upton
+--------------------------------------------------+
CHROMiX / Profile Central
www.chromix.com www.profilecentral.com
+--------------------------------------------------+
--