[cairo] Spot colors (and CMYK)

Chris Murphy lists at colorremedies.com
Sat Feb 20 16:48:28 PST 2010

On Feb 20, 2010, at 2:04 PM, ecir hana wrote:

> On Sat, Feb 20, 2010 at 9:28 PM, Chris Murphy <lists at colorremedies.com> wrote:
>> I do not follow the logic on this at all.
> It's very simple: sometimes you just don't have a profile.
> I asked you several times, where should I get a profile for Pantone
> duotone. Or glossy varnish. If it was mandatory to always supply a
> profile, what should I do?

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.

>> An XYZ or LAB alternative for a spot color is an obviously good idea if you want to see the spot color on any kind of device that does not print with the exact spot color (ink in a bucket). Depending on how the spot channel is used, determines how it would be defined.
>> If it's just for a logo or something like that, an ICC profile for the spot color isn't necessary, just an XYZ or LAB value so it can be displayed and proofed properly.
> LAB fallback = tagged CMYK fallback, that's what I said.

Not the same thing, at all.

>> If you insist on untagged device values, you can only use spots except where ink in a bucket is present. Bad idea.
>> Wrong. You want a particular result, and specifying it as 100K isn't going to get you there because some downline process *will* convert it the instance K100 isn't available. It's important to define the object type. Black text is unique and needs variable handling. Simply saying it's always 100% K can result in incorrect handling.
> How do you know what I want? I want one solid plate. Don't care where.
> "The correct answer is, it depends.", as you said to Bill Spitzak.

In the context of certain special objects, like black text, what you want doesn't matter, as the content creator. Your job is simply to define it as "black text". Downstream, the decision is made to determine the best way to handle "black text". That is output condition specific. 

>>> That said, I understand that untagged CMYK is difficult to show on the
>>> screen. I'm not against tagged content - I'm against making tagging
>>> mandatory for spots. Perhaps always tagged CMYK is even a good thing
>>> but for DeviceN that means only its fallback. (And I could always
>>> create untagged DeviceN with cyan, magenta, yellow and black). I also
>>> believe DeviceN should have its fallback defined in CMYK only, and the
>>> fallback should compose over CMYK only, as well).
>> This makes zero sense to me.
> That's PDF/X-1a. There is just one tagged CMYK for everything, except
> for spot colors (but their fallback is in that CMYK, too)

PDF/X-1a has limitations - when I referenced it, I'm referring to it as a broad concept. One area where it fails is that spot colors can be defined with CMYK alternates, which is absurd and useless. If it were up to me, it would be mandatory to tag them with an XYZ or LAB value.

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

More information about the cairo mailing list