RGB workflow with paper cast
RGB workflow with paper cast
- Subject: RGB workflow with paper cast
- From: email@hidden
- Date: Wed, 27 Jun 2001 10:31:10 +0100
Hi folks!
Suppose you have lots of cover images from books of a publisher which
uses a yellowish paper for the covers (sort of a corporate identity
thing). These cover images, not the original artwork, are the
"reference appearance", since from them, ads are derived for
magazines, back-lit displays, or the web. The yellow paper cast should
always be preserved. It is approx. CIELab 96/0/8.
While using the CMYK production data and a corresponding profile works
for now, one might prefer to keep the dynamic range of the original
(whatever this is) for reuse. So we'd like to set up a RGB workflow
between service bureaus, publisher, marketing agencies, and book
printers. But in no case we'd want to go beyond this yellow tint.
Accidentally, this is also the first time for me to dance the RGB
with customers, so I'm trying to give all the details, and I'm
probably boring you a little. Please bear with me...
Printing the books comes first, retargeting can be dealt with later.
So I have to deal with only three profiles: scanner profile, RGB
working space, and yellow CMYK (the press result on the yellow paper);
and with two conversions: scanner RGB to working space, and working
space to yellow CMYK.
The scanner profile is whatever the service bureau uses.
Regardless of the working space, the CMYK profile has to keep the
paper tint in the tables and not in the white point tag, so that one
can use the perceptual rendering intent for separating and still
account for the yellow. Using ColorBlind with "Absolute Gray" turned
on, I can confirm with the color picker in PS6 that indeed CMYK
0/0/0/0 maps to CIELab 100/0/8 (perceptual or relative colorimetric
intent). RGB images are correctly separated with less yellow (or more
blue) for the yellow CMYK than for neutral CMYK profiles. So this part
seems ok.
Now for the working space. Since the yellow cast should be always
present, I selected a standard RGB profile large enough to capture the
gamut of the print and, to some extent, of the original. For testing,
I used ECI-RGB (from the web site www.eci-org; profile version 1.0).
ECI-RGB has D50, gamma 1.8, and NTSC primaries. I then patched the
wtpt information (to CIELab 96/0/8, or 100/0/8). This shifted the RGB
primaries a little (since they are relative XYZ values), but the gamut
is still ok to me.
It seems that to actually see the yellow tint in PS6, I have to view
RGB data (with this yellow RGB assigned) through the soft-proofing
mechanism set to simulate the same yellow RGB space with absolute
colorimetric intent.
The problem remains to convert from a neutral RGB (scanner or
matrix-based D50 or D65) to the yellow RGB, which should counteract
the cast by pushing blue. Absolute colorimetric conversions work of
course, but out-of-gamut colors (neutral whites) are being cut off.
I'd like to use perceptual conversion, but this ignores the difference
in white points, since this time I cannot keep the tint in the RGB
A2B and B2A profile tables as done with CMYK.
Then I tried RGB-to-RGB device links made by ColorBlind Edit. In the
first AppleScript conversion, they haven't been accepted (invalid
profile, but this may be due to CS 2.6.1). I also haven't investigated
if such a perceptual link profile accounts for the cast.
So how should it be done? Scan the image, convert to working space,
correct tint manually until screen image pleases? I'd expect something
better, but I'm in the dark...
Hoping for enlightenment,
Johannes