[cairo] Spot colors (and CMYK)

[Chris Murphy]

On Feb 20, 2010, at 6:52 PM, ecir hana wrote:

> On Sun, Feb 21, 2010 at 1:48 AM, Chris Murphy <lists at colorremedies.com> wrote:
>> 
>> I'm not saying you always need a profile. But such colors do need XYZ or LAB (or spectral if we look forward) metadata associated with them. Further, I propose we could also use ink opacity information.
> 
> You know me, I couldn't care less about alternative definition of a
> spot color. If I was up to me, all would be random - so you could
> easily distinguish between them :). But what I try to do is to
> understand what Cairo needs to ask a user for. For now, for CMYK and
> RGB you need to supply a profile and for spots a LAB triple?

I'll make a supplement to my original statement. We don't actually need XYZ or LAB, but convertible into it. So the alternates could be in a wide gamut RGB space, specifically named (and embedded). I'm not sure exactly how PDF handles this - I'm thinking spot alternates are /devicecmyk or /devicergb so presumably those inherit some kind of space for blending, probably the transparency blending space. I know PDF can also store alternates as LAB, I'm not sure about XYZ.

Anyway, whatever is responsible for rasterizing the PDF (or any file for that matter) would be the one responsible for converting spots to some wide gamut RGB space for compositing/blending purposes.


> And I
> know that in lcms you can create a transform from one profile to
> another but I don't know what to do if I have just a LAB triple. And
> then you say "ink opacity" - where, as a user, am I supposed to get
> such info?

This is an ideal. It's not practicable today because we don't have such data. It's unpublished for Pantone colors as far as I know. But this is something that would improve the predictability and portability of specifying spot colors, as it has a much bigger effect on the result in print than does spot color dot gain, when combining spots. Does spot1 overprint spot2? Or is spot1 translucent enough that spot2 shows through, and how much? I think having a standard way to describe such behavior will help down the road, in making these workflows more automatible, and predictable.



Chris Murphy
Color Remedies (TM)
New York, NY
----------------------------------------------------------------------
Co-author "Real World Color Management, 2nd Ed"