I’ve been trying to figure out color management (mostly in InDesign, but this question is more general) in a print production workflow in more detail recently.
What I am trying to achieve is printing a color that is as close as possible to what I define in InDesign, without having “proof colors” turned on.
I want to use a very obvious but not quite realistic example (and assume that the screen of my computer renders colors correctly):
I defined a color that needs to be printed exactly as viewed on screen (it’s inside the screen’s and target profile’s gamut) and this color equals, totally by chance, the color of the media white point of the target color profile of my document. The target profile is
PSO Uncoated ISO12647 (ECI), so I get the media white point from the ICC file—values in the XYZ color space:
X = 0.846,
Y = 0.877 and
Z = 0.747.
I can then translate these to the Lab color space:
L = 95.03,
a = 0.07 and
b = -2.03.
And these values again I can add as a color swatch in my InDesign document and convert them to a CMYK swatch:
C = 5,
M = 3,
Y = 1 and
K = 2.
I add two rectangles to the artboard, one in the Lab and one in the CMYK color.
In preview mode it looks like this and we can see that the two colors match:
(We are now back in the real world where I can’t print a proof of this document but need to rely on the soft proofing function of InDesign.)
Now I turn on “proof colors” to see how this will look when printed. To get as close as possible to reality, I make a custom proof setup and enable paper color and black ink simulation (of course for the document’s already assigned color profile).
What I see now and what I think is about correct is this:
The colors will be printed on the paper that has a color of its own—and by that get “multiplied”.
But since I actually want to see my originally defined color printed, wouldn’t it be a better solution to print “nothing”, because my desired color is already there (in the paper)?
So I subtract the media white point from my CMYK color by hand, because these values are already waiting on the sheet… and get
C = 0,
M = 0,
Y = 0 and
K = 0—what a surprise. The output looks the way I imagined it now:
Of course this doesn’t make sense with multi-color graphics like photographs because we need a real white point, but that’s what the “perceptual” rendering intent is there for when converting images to a different color space and / or profile.
But when trying to print a certain, defined color, wouldn’t compensating the media white point be the way to go? Or is this something that can be (or already is?) handled later in the process, in a color profile for the print machine for instance?
The idea is the same for any color but most obvious by using the paper color directly. You could subtract the paper color from
C = 100,
M = 75,
Y = 50,
K = 25 as well—that would yield in
C = 95,
M = 72,
Y = 49,
K = 23 and should give a color more accurate to the one you have seen on screen than simply printing the original values.
Maybe this is something very obvious that is already dealt with but I’m lacking the knowledge and words for it.
(The paper color used by the color proofing functionality of InDesign doesn’t match my defined paper color exactly, but I will ignore this for the sake of the example.)