How do we tolorance a "match"
How do we tolorance a "match"
- Subject: How do we tolorance a "match"
- From: Ray Maxwell <email@hidden>
- Date: Thu, 02 Nov 2000 10:05:52 -0800
- Organization: CreoSitex Inc.
I am posting this to both the ColorSync Users Group and the "Printing
Color" listserver.
I would like to bring up a subject that rarely gets discussed, but I
think is at the heart of color management. I hear people talk about a
press and proof "match". They usually say it is a good match or it is
not. Sometimes you hear a printer say, "I can see a difference". All
of this is very subjective. I would like to ask for suggestions as to
how we can put a number or figure of merit for a match between a press
sheet and a proof.
Someone is going to suggest that we use delta E. The problem with this
is that you may have a good match of a mid-tone gray that is less than a
delta E of 1 and have other colors on the same proof that are out by a
delta E of 10. How do we evaluate the whole color gamut and assign a
figure of merit that reflects the quality of a color match?
To start off the discussion let me throw out a few ideas.
The technical people at RR Donnelly have chosen to take a subset of the
IT8.7/3 targets (colors near the center of the gamut) and produce a
proof and press sheet. They then do a plot that represents how many
colors in the set are within certain delta E's. The Y axis of their
chart runs from 0 to 100 %. The X axis starts at 0 delta E and goes up
to about 10 delta E. This then shows what percentage of the colors that
were within a delta E of 1, 2, 3, etc. A perfect match would be a
vertical line against the Y axis meaning that all color patches had a
delta E of 0. The curves tell you something about how good the overall
match is.
The only trouble with this is that the mid gray tones have the same
weight as colors out at the edge of the gamut. Most people would want
to give more weight to the mid tones than the colors at the edge of the gamut.
Another idea would be to use a computer file equivalent of the Macbeth
color chart. One would give higher weighting to the flesh tones and mid
tones and lower weight to the solids. The delta E's would be measured
and then put through a weighting function that would result in a number
that represented the quality of the overall color match.
OK, that should get the ball rolling. What are your thoughts?
Ray
--
Ray Maxwell
Color Systems Engineer
CreoScitex Inc.
4225 Kincaid St.
Burnaby, B.C.
Canada V5G 4M1
Phone (604) 451-2700 ext. 2004