• Please use real names.

    Greetings to all who have registered to OPF and those guests taking a look around. Please use real names. Registrations with fictitious names will not be processed. REAL NAMES ONLY will be processed

    Firstname Lastname

    Register

    We are a courteous and supportive community. No need to hide behind an alia. If you have a genuine need for privacy/secrecy then let me know!
  • Welcome to the new site. Here's a thread about the update where you can post your feedback, ask questions or spot those nasty bugs!

CMY(K) coordinates

Doug Kerr

Well-known member
From time to time, we see here discussion of a particular image in which the commenter mentions that a certain spot has a certain color, described in CMYK terms (and thus, perhaps, some different postprocessing would have been desirable).

This always seems to me to be inappropriate. For openers, our theory of color perception treats color as three-dimensional - three "coordinates" (under some color model) are needed to define any color - not four.

My understanding of the CMYK color description scheme (I hesitate to call it a coordinate system, since bona fide color coordinate systems have only three coordinates) can represent any specific color with an unlimited number of "quadruplets" (to use a mathematician's terms for suite of four numerical values). Said another way, a given color does not have a single CMYK description.

For example, any color within the gamut of a particular scheme can be represented with a CMYK quadruplet having a K (black) value of 0. Or with a CMYK quadruplet having various non-zero values of K - essentially arbitrarily.

The CMYK system is really intended to express a "recipe" for rendering a particular color following the principles of four color printing. There, conceptually, any color within the gamut can be produced by the appropriate densities of cyan, magenta, and yellow ink. But in reality, for one thing, true blacks are hard to actually achieve this way, and so for this reason, and others, black ink is used as well.

The appropriate black density to be used for rendering any given color (for example, a reflective color we can describe in three coordinates under the L*ab coordinate system) does not have a unique value. As I understand it, various algorithms are used in the printing process to determine this, based on balancing several criteria (and of course depending on the actual process to be used). (I think that even the relative costs of the different kinds of ink actually figures into it.)

Thus, while our image editing and manipulation software may take any pixel (perhaps represented in memory in RGB coordinates, or CMY coordinates) and report to us a CMYK "recipe", that is arbitrary.

Now, any defined three-primary "subtractive" color model (CMY) does have a unique "triplet" for any color.

But it seems that stating a color (not a recipe at the point where we are ready to lay down ink) in CMYK form is misleading.

So my thought is that, when the use of the "subtractive" type of color model is handsome because of the understanding it gives, we should describe colors of interest in terms of their C,M,Y coordinates, not their C, M,Y, and K values.

Of course, I don't really work with these concepts, so I may be barking up a wrong tree here. Perhaps real workers here in the realm of CMY and CMYK representations can let me know where I might have gone wrong.

Thanks for any help you can give.
 

Doug Kerr

Well-known member
Another interesting wrinkle is this. The "serious" CMYK models (such as those employed in Photoshop) are based on various specific four-color printing contexts, and take into account the properties of specific inks and media and various decisions about black ink "strategies" (UCR, GCR, UCA, etc.), as well as the esoteric area of halftone "dot gain".

As such they aren't really suitable for precise "color description" in the general image processing sense. (For general palaver about skin beauty and the like, of course there is no precision involved, so that is no problem.)

There is also an "simplistic" CMY model, which is just "RGB looked at from the other end" (subtractive rather than additive outlook). Its CMYK form really leaves up to the user to decide what to do about the black value.

For example, in Picture Publisher 10, whose CMYK coordinate system is of that form, the transformation from RGB to CMYK is as follows (all values on a 0-1 scale for conciseness):

C = 1-R
M = 1-G
Y = 1-B
K = 0

If the user thinks that some black in the description would be good, he can set it in, and the program will then change C, M, and Y accordingly (to still represent the same color, under the "simplistic" model).

Best regards,

Doug
 

Jack_Flesher

New member
CMYk is simply a color output method using 4 colors; it isn't a color model. It has been bastardized into behaving as a color model so we can translate data from a theoretical color model directly over to its output regime. As such, a CMYk "profile" is developed for the particular output device and media, but in reality that is only a translation that renders RGB coordinate system colors reasonably close using those four colors (C, M, Y, and k) of ink.

Cheers,
 

Andy brown

Well-known member
Doug,
I see the world through "red/green' colourblind eyes and really, any colour can only be seen through the eye of the beholder.
I see colours extraordinarily differently to most and am deprived of some colours yet see unbelievable colours where others see merely bland ones.

I don't think colours can be mapped.
Maybe from a purely physical, mathematical perspective but not from any human base reference point.
 

Doug Kerr

Well-known member
Hi, Jack,

CMYk is simply a color output method using 4 colors; it isn't a color model. It has been bastardized into behaving as a color model so we can translate data from a theoretical color model directly over to its output regime. As such, a CMYk "profile" is developed for the particular output device and media, but in reality that is only a translation that renders RGB coordinate system colors reasonably close using those four colors (C, M, Y, and k) of ink.
Well said.

Thanks.

Best regards,

Doug
 

Doug Kerr

Well-known member
Hi, Andy,
Doug,
I see the world through "red/green' colourblind eyes and really, any colour can only be seen through the eye of the beholder.
I see colours extraordinarily differently to most and am deprived of some colours yet see unbelievable colours where others see merely bland ones.

I don't think colours can be mapped.
Maybe from a purely physical, mathematical perspective but not from any human base reference point.
A point well taken.

Thanks.

Best regards,

Doug
 

Andrew Rodney

New member
I'm in full agreement with both Doug and Jack.

CMYK is a goofy color space that requires Black due to the messiness and impurity of the other three inks inability to produce a good black. Unless you are in print/prepress, best to ignore it as it can make your head hurt (one has to consider the huge influence of UCR and GCR on the CMYK recipe and the results of ink on paper!).

CMYK is an highly device dependent output color space.
 
Top