Dan Margulis Applied Color Theory

The RGBK Definition

Subj: [colortheory] RGBK
Date: Friday, September 22, 2006 5:22:03 AM
From: Ronald Greenberg

Dear Dan,

One of the useful things (there are so many) that I found in the most recent version of Professional Photoshop is your RGBK profile for working in CMYK. I took to heart your dictum to learn how to color correct in CMYK, even though my final output to my desktop printer is RGB. I was bothered that some colors became dull when going into CMYK because my printer is an Epson 1800 and has a gamut wider than traditional CMYK for a press. Now I set my working CMYK space to RGBK and, amazingly, nothing is lost when I later go into Lab or back to AdobeRGB.

First, a question. Are there any tips or tricks, or anything else to be aware of, when using RGBK? It seems to just work great and, unless you indicate otherwise, I assume everything you write about in the first seven chapters re CMYK, such as appropriate numbers for flesh tones, neutrals, highlights and shadows etc. still applies when using RGBK. Also, I assume that when using RGBK for this purpose, setting dot gain or GCR generation type is irrelevant, unless I wanted to make a false black plate to apply a curve to or some other specialized purpose.

Second, a suggestion. In the next edition of your book, you might want to mention RGBK much earlier in the book, perhaps just a little box in an early chapter, mentioning how it can help for those who want to work in CMYK but don't need it as a final output space.

If others in this group use RGBK, I'd like to hear about it. And if no one else does, I'll just consider it a personal gift to me, and thank you for it.

Best regards,

Ronald Greenberg
___________________________________________________________________________

Subj: [colortheory] Re: RGBK
Date: Friday, September 22, 2006 7:00:29 AM
From: Stephen Marsh

Ronald Greenberg wrote:

First, a question. Are there any tips or tricks, or anything else to
be aware of, when using RGBK? It seems to just work great and,
unless you indicate otherwise, I assume everything you write about
in the first seven chapters re CMYK, such as appropriate numbers for
flesh tones, neutrals, highlights and shadows etc. still applies
when using RGBK.

I am not Dan, but Dan knows his numbers...so I would not presume that he would say that the WGCMYK values are close to what most common CMYK users are dealing with.

Ronald, I think that this is the profile made by listmember Mike Russell?:

http://www.curvemeister.com
http://www.curvemeister.com/downloads/index.html

For free download, he has on offer the light GCR profile.

As a feature of his Curvemeister plug-in, he has WG CMYK (as well as RGB, LAB and HSB curves). The WG CMYK GCR option can use no, light, medium, heavy, maximum and working CMYK black generation.

Personal opinion follows which may or may not differ from Dan's and Mike's:

It is OK for many images (if the image does not visually suffer and the edits in WG CMYK make the move worthwhile), but there will be images that have a wider gamut than this wide gamut CMYK profile and visual detail will be lost when converting to it from say ProPhoto RGB.

RGB, LAB and CMYK each have their good and bad points. So does WGCMYK.

It is useful for sharpening, plate blending and other things. It is useful for colour correction, but in a visual way more so than going by the numbers - unless you wish to  learn a whole new set of numbers. Simply convert an RGB grayscale gradient and step patches and skintone sample to SWOP CMYK and evaluate the numbers...then do the same for WG CMYK. Try editing to standard SWOP type aimpoints in WG CMYK. I think that you will find that the two profiles describe two very different "devices".

A search of the list archives will turn up more than a few hits, using keywords such as "RGBK", "WGCMYK" or "Wide Gamut CMYK".

Regards,

Stephen Marsh.
___________________________________________________________________________

Subj: [colortheory] Re: RGBK
Date: Friday, September 22, 2006 7:52:36 PM
From: Ronald Greenberg

Dear Stephen,

Thank you very much for your thoughtful reply. I have Mike Russell's WGCMYK as well as his Curvemeister plugin. Dan says WGCMYK is very similar to his RGBK.

Dan said, basically, that all he did to make RGBK was to change the values of the primaries so they were more intense. So, the values in the file will be different, in the sense that the same values will now represent different colors to expand the gamut. But I think the values could be used in the same way as before. For example, Dan gives formulas for judging whether numbers for skin tones are correct. These formulas are in the form of ratios, for example the cyan should be about a fifth to a third of the magenta, the yellow should always be equal to or more than the magenta, etc. Now, here's my question: I should be able to still use these formulae with the numbers I see on my info palette in the same way I would if I were in normal CMYK, can't I? Dan always urges to correct by the numbers, and in this sense aren't the numbers still "the same", even though they now represent slightly different colors, because everything has been shifted proportionally? For example, shouldn't I be able to use the same formula for judging neutrality (magenta equals yellow, with cyan slightly higher), and therefore can't I still use the same tests when looking for color casts? Also, since CMYK values are expressed as percentages, it seems that the numbers shouldn't be different, or rather the proportional relationships between the numbers shouldn't be different at all.

You are right, I should do some of my own tests to answer my own questions. But some of this can be subtle, and I am interetsed in comments from people like you who are smarter and more experienced than I am.

Thanks,
Ronald Greenberg
___________________________________________________________________________

Subj: [colortheory] Re: RGBK
Date: Friday, September 22, 2006 11:12:47 PM
From: Stephen Marsh

Ronald Greenberg wrote:

looking for color casts? Also, since CMYK values are expressed as
percentages, it seems that the numbers shouldn't be different, or
rather the proportional relationships between the numbers shouldn't be
different at all.

OK, so the RGBK profile is different to the WGCMYK one. I have no experience with this profile and can't comment further on it directly.

Create a new CMYK blank file in RGBK or WGCMYK, as a simple test fill the document with 50c40m40y and then inspect LAB values for the colour - are the AB values at zero? This can and should be repeated along the tonal scale from shadow to highlight and compared with standard CMYK aimpoints for neutrals.

Then do the evaluations with skintones, foliage etc.

My point being that standard aimpoints were taken from how real life press, stock and ink interact. You are now working in a synthetic, idealised space - that may or may not be close in behaviour. This is what comparing standard SWOP aimpoint values for known colours to what the synthetic spaces provide will indicate - how close is this space to the standard CMYK that most users know.

Stephen Marsh.
___________________________________________________________________________

Subj: Re: [colortheory] RGBK
Date: Saturday, September 23, 2006 4:18:03 PM
From: Dan Margulis

Ronald writes,

I was bothered that some colors became dull when
going into CMYK because my printer is an Epson 1800 and has a gamut
wider than traditional CMYK for a press. Now I set my working CMYK
space to RGBK and, amazingly, nothing is lost when I later go into
Lab or back to AdobeRGB.

OK, that should work for *you*, because apparently you never have to give a CMYK file to anybody else and because you are only interested in RGB output. If either of those statements were not true, using a false profile as your CMYK definition would be dangerous.

First, a question. Are there any tips or tricks, or anything else to
be aware of, when using RGBK? It seems to just work great and,
unless you indicate otherwise, I assume everything you write about
in the first seven chapters re CMYK, such as appropriate numbers for
flesh tones, neutrals, highlights and shadows etc. still applies
when using RGBK.

Fleshtones would not be the same; you'd need more cyan to make up for the assumed intensity of the magenta and yellow "inks".

Also, I assume that when using RGBK for this
purpose, setting dot gain or GCR generation type is irrelevant,
unless I wanted to make a false black plate to apply a curve to or
some other specialized purpose.

That's right. Since you're converting back to RGB anyway, the structure will be lost, so unless you have some purpose for it, you can ignore it.

Second, a suggestion. In the next edition of your book, you might
want to mention RGBK much earlier in the book, perhaps just a little
box in an early chapter, mentioning how it can help for those who
want to work in CMYK but don't need it as a final output space.

Actually, I've taken RGBK out, because I still include Mike Russell's Wide Gamut CMYK, which is close to the same thing.

I don't emphasize this workflow, not because it isn't a valid one, but because it isn't for everyone. It's really for people with RGB workflows who are simply more comfortable with a CMYK structure. Anybody who's going to a real CMYK later anyway doesn't need it, and anybody who is comfortable with RGB maneuvering only needs to go round-trip to CMYK in a small percentage of images. Also, a lot of people who have PCs use Curvemeister, Mike's product.

Me, I don't use it, because on the unusual occasions when I go to CMYK and then come back to RGB, I'm aware of the pitfalls and take  make sure I don't lose anything. I *could* use it, but I'm just afraid that it would confuse me and that at some important time I would forget the CMYK that I was in. On the other hand, I know that a significant number of readers, like yourself, are making use of it. It sounds like your workflow is effective.

Thanks for the kind remarks about the book.

Dan Margulis
___________________________________________________________________________

Subj: Re: [colortheory] RGBK
Date: Saturday, September 30, 2006 10:33:09 PM
From: Don Schaefer

Dan, I use RGBK very successfully in my RGB in/out workflow, when it is called for. I appreciate the understanding I've gained from your Professional Photoshop series and LAB book. It took a while to understand, but the effort was well worth it. I'm a photographer who captures and prints RGB, but I am comfortable in all the spaces, thanks to you. Some of my pro colleagues are somewhat fearful of "all the data loss going into LAB", and consider you somewhat the "crackpot", but I just quietly go on with my work. Others from that group find you enormously helpful, as I do.

I know you are scaling down your many activities, but please keep your hand in with us here. We need you.  

best,

don schaefer

___________________________________________________________________________

Subj: Re: [colortheory] RGBK
Date: Sunday, October 1, 2006 10:14:24 AM
From: Andrew Webb

Watch out for skies and water. They'll change on you sometimes.

Cheers,

_andrew webb
___________________________________________________________________________