Jeff,
experimet we did, indeed, but time has been limited and permutations are
infinite. Some responses to your commments:
> 1) In the Hex-FS inkset, the Pc=C and Pm=M -- basically yielding a 4-
> ink inkset. I was wondering why the curves for the Pc-C and Pm-M
> were so different? I think I can see the advantage of starting them
> off in a slightly staggered fashion to smooth the transisition
> (minimize dots), but not later. Did any one try using only 4 curves
> (actually 6 curves, but with 2 doubled up)?
Essentially you answered the question yourself. Staggering minimizes dots. I
have a hex profile with 3 grays and a black and it's OK but the dots are more
apparent (on a 7000-era printer). Staggering would probably matter less on a
1280 / 2200 generation printer.
I did the 4-ink proflle on a 6-ink printer because I wanted to use two different
inksets: Sepia and Warmtone Piezotones with Museum Black. In that
scenario, each grayscale is made up of 3 grays and a black. The Yellow and
Magenta are Sepias and the Light Magenta and Light Cyan are the
equivalent of Yellow and Magenta for WarmTones. The Cyan and the Black
are common for the two scales (but very little C is used, and it is WT). So I was
able to make a nice warm (sepia) print on German Etching and turn around
and do a cooler Warmtone print on EAM or PRag. It would have been more
dramatic to have Seleniums and Sepias (and blend interestingly between
them). Perhaps next time.....
>In other words,
> conduct more of the linearization directly in the curves rather than
> apply what I assume (jump on me if I am wrong) to be a single
> transfer function to the composite profile.
You may want to look up an earlier message where I explained that
Linearization is but an option in IJC. You can throw a quick-and-dirty set of
curves together, read them in and let IJC "bring them in", via linearization.
Or, you can tweak the ink curves until you are dead on. Each approach has
it benefits and problems.
The general caveat is that Linearization won't correct gross error - so you
have to do _some_ work with the ink tweaks. Also, correcting the shadow
portion of the curve with Linearization alone can get to be dicey. I like to do
all the work in the Ink Tweaks, until I hit the target within +-0.02 units of
density. Then, use Linearization in future uses of that profile, if needed, to
counteract drift, paper variances and the like. This isn't as much work as it
sounds if you start from an existing set of curves and have some experience
doing it.
But of course, there are a hundred ways to approach a profile and the nice
thing about IJC is that it doesn't lock you into any one of them. I hope once
more people start using it, we will be able to exchange profiles not just to
make prints but to see everyone else's approach to profilemaking.
Antonis