RGB vs CMYK

[Stephen Marsh]

Yes agreed Erik, when ink/toner hits paper, the easily demonstrated theory of colour numbers and lab values does not always hold, despite the math saying otherwise.

Stephen Marsh

 

[Stephen Marsh]

Most/all RIPs are CMYK driven, even if the inkset is CMYK+. The gamut of a LF printer is often larger than regular offset press CMYK, so limiting the input to press CMYK may or may not be considered the best print outcome, depending on expectations.

Stephen Marsh

 

[Correct Color]

Man...

That's a lot of info, but honestly, if you ask me, it's all kind of missing the target.

Here's why any device can reproduce CMYK or RGB files:

Because CMYK and RGB in this context only refer to the pixel definitions of the file. That's it. End of story.

Leave it at that and then let color management do the rest. If you do color management correctly, it will work.

Of course you've got to go beyond just RGB and CMYK and ask what RGB and what CMYK, and once you do -- and there's no getting around it -- then every CMYK value and every RGB value in every RGB or CMYK file represent some L*a*b* value, and it's that value that matters, not its numeric representation in whatever CMYK or RGB space the file is in. You can actually think of that representation as just a placeholder in the file for the true L*a*b* (or occasionally XYZ) value.

And, if you get a poor transform from RGB to CMYK, then the fault is in how you're managing the transition. Period. It is not that the final device is incapable of accepting the file.

There is no device that's incapable of accepting RGB files. Anyone who says otherwise is misinformed.

Interestingly, did you ever hear anyone say: "My monitor is RGB, I can't use it to display a CMYK file"?

Makes just about as much sense as saying you can't send any printing device an RGB file.

(Edited to add: Oh, and of course it needs to be said...

RGB has a larger gamut than CMYK...

That is not true. Because by its nature CMYK is reflective and by its nature RGB is transmissive, and since light tends to get more diffused in reflecting than transmitting, it is true that many RGB color spaces have larger gamuts than many CMYK color spaces. This is especially true with most of the standard RGB and CMYK working spaces.

But it's certainly possible to have a small gamut RGB color space (a laptop monitor, say) that is smaller than a large gamut CMYK color space (say an aqueous inkjet on gloss stock, or even a solvent inkjet on cast coated.)




Mike Adams
Correct Color

 

[Correct Color]

Slush11,

Just to answer you directly...

OK so I work in large format and have worked with 4 different brands and types of large format printers now, all which require CMYK. I've spoken to our digital and offset pressmen and all of them say they require CMYK as well, as do our small toner based copiers/printers.

All I do for a living is color management for large and grand format printers worldwide, and I can tell you emphatically that not only is this wrong, it is disastrously wrong. If you want to drop me a line or give me a call, I'll explain why.

We are one of the biggest printers in our city...

Over the years, I've seen some pretty disastrous color management policies at places that would fit that very description.

So...is it just that we don't have profiles set up for this?

In large measure, yes.


Mike Adams
Correct Color

 

[Correct Color]

Just to add...

IMO even if a colour is in gamut, it does not mean that it can be reproduced. In an abstract way of looking at colour spaces, a colour that is in gamut should be reproduceable but with colours made from real pigments, it is my view that all colours within the gamut are not reproduceable.

Actually, it's been my experience that they are.

Keep in mind though that you've got be be really, really sure what's "in gamut." Also you've got to keep in mind the viewing conditions and the rendering intent of the transform.

Let's take for example the gamut for a CMY space. One would think that black and dark grays, which are within the gamut, should be reproduceable but they are not. If one tries to make a black or dark gray from some CMY ink sets, one tends to get muddy browns. When one tries to get a CMY patch to look gray, it usually seems to be a slight bit off the neutral gray.

Well, first I'd point out that you probably need to define what is "black." True, absolute, black would be L*0 a*0 b*0. But, of course that's hard to hit with just about any media-ink combination, let alone with no black at all. So it's pretty unlikely to be "in gamut" to any CMY inkset on any media, so what is the "black" you're actually trying to hit?

I'll just wager here that the problem you're encountering here isn't that in-gamut colors can't be reproduced. Rather it's that you're attempting to reproduce them with a profile that isn't an exact characterization of exactly how the device you're using it to represent actually prints.

Any profile/color space is a characterization of some device reproducing color in a certain state. And that profile is all the RIP knows when it makes the transform from any one space to another. Unless you are 100% sure that the device you're representing with that profile is reproducing color exactly as that profile was made, then any conclusions you draw based on the results of using it are going to be invalid.

Black is no different than any other colour and therefore there can be many other specific colours that can not quite be reproduced too. Anyhow for practical purposes, we use CMYK so that the K can make nice blacks and not mud.

And, of course, the closer any color is to absolute neutral, the more noticeable any discrepancies between the profile and the actual printing characteristics of the device are going to be. Not because our eyes see grey this way or that way or the other way, but simply because since all the colorants have to be near to exactly equal, any discrepancies are magnified for that reason.


Mike Adams
Correct Color

 

[David Milisock]

In an ICC color managed work flow with Ink Jet printers and some toner based digital devices I prefer to send a mixture of RGB/CMYK data.

All images as RGB, spot color converted in the application to RGB (assuming that the application is using Adobe RGB, ECI or a similar wider gamut RGB color space) grayscale as crayscale and CMYK corporate identity or other CMYK data as CMYK.

I use CorelDraw published, native color PDF with transparency flattened. This produces the widest gamut in output for grayscale, spot color, the RGB images with the best chance of reproducing the CMYK data correctly. Other applications can also produce files in RGB, Grayscale, spot color and CMYK.

A t​rue CMYK device will not reproduce many RGB colors well.

 

[Correct Color]

A t​rue CMYK device will not reproduce many RGB colors well.

Really?

Could you name some?

Just the RGB numbers and the actual RGB space would be fine.



Mike Adams
Correct Color

 

[David Milisock]

Sure any of the pure RGB colors will not reproduce well, any RGB color space, sRGB or larger, R255 G0 B0, R0 G255, B0, R0 G0 B255 just to mention 3, the list can go on for quite awhile. If you have access to an application that displays gamut views you can see what will and what won't reproduce on press.

I rarely use Adob​e products for file creation but in CorelDraw set up in RGB mode using North American press default color settings and with soft proofing off ( proofing off which is default) will display colors in the documents color space. Simply create a circle in an RGB color duplicate it and convert the duplicate to CMYK, the color shift that displays (if your display is calibrated) will accurately represent the out of gamut condition.

 

[NeolithPrint]

I have not noticed much difference when color matching my aqueous ink jet printers when printing canvas photo prints or panoramic photo prints. But the difference between the results of my sublimation prints for metal photo prints and stone slate photo prints is very dramatic when I begin with a RGB image versus a CYMK image. For this reason I convert all images that are submitted to my web site http://neolithprint.com to CYMK and validate the conversion versus the original using graphic software prior to sending the print job.

 

[David Milisock]

I have not noticed much difference when color matching my aqueous ink jet printers when printing canvas photo prints or panoramic photo prints. But the difference between the results of my sublimation prints for metal photo prints and stone slate photo prints is very dramatic when I begin with a RGB image versus a CYMK image. For this reason I convert all images that are submitted to my web site http://neolithprint.com to CYMK and validate the conversion versus the original using graphic software prior to sending the print job.

I would assume the AQ printer is close to sRGB or maybe Adobe RGB in it's gamut and the dyesub is not.