The root of my question was not a practical matter, however, rather
a technical inquiry: is it a technically sound approach from a
colour management perspective to profile a printer's contract
proofer for both proofing and separations for press, assuming the
printer can match their own proof?
I'd have to say that profiling a proofing system as opposed to a press
is ALWAYS preferable even if it does compromise the color match
somewhat, but with the following caveats:
* The "best" thing to profile would be a dot proofer like a Kodak
Approval, Spectrum or Fuji FinalProof HOWEVER if it's color-managed
using only curves (no ICC device links) you're going to have some
issues between the "perfect" wet trap characteristics of these proofs
vs. the imperfect nature of a press's wet ink trap. This will
compromise the color match you'll get on pure secondary colors.
* In general, I would not recommend profiling the output of an inkjet
printer that's simulating a press condition. Better to simply grab
either the ICC profile or the characterization data used as the
"source" in the inkjet proofing system, that way you're using 1st
generation data rather than a 2nd generation simulation of that data.
Plus, I think there's all kinds of weird spectral things you may get
into when profiling inkjet output. This is not an example of inkjet
output exactly, but I can remember trying to profile the "color-
managed" output of a Fuji Pictro some time ago and marveling at the
complete wackiness of the shadow gray balance, which could not have
been anything like the press condition it was intended to emulate. Had
this profile been used for proofing or, worse yet, for separations it
would've been a Bad Thing.
Of course, I think the best of all worlds is using *standardized*
characterization data sets that's been taken from actual press runs
like the current "G7" SWOP/GRACoL data sets or any of the multitude of
Fogra data sets.
In my own work, I stepped away from profiling customer's presses a
couple of years ago and instead adjusted tone curves to match a
standard press condition such as SWOP or GRACoL and then set the
proofer up using the equivalent standard data set/profile. This has
worked extremely well for some time. Lately however, I've gotten back
into custom profiling of presses for the purpose of using device link
profiles in the workflow either instead of plate curves or in
conjunction with plate curves. Custom profiling of presses is an ugly
business if process control isn't front-and-center in the minds of
those in the pressroom.
