Revisiting A Color Myth

From ColorWiki

Revision as of 00:10, 9 November 2018 by Patrick (Talk | contribs)
(diff) ←Older revision | view current revision (diff) | Newer revision→ (diff)
Jump to: navigation, search
Reserved Article

This page is a
Reserved Article.
For more details see
Reserved ColorWiki Articles

ColorNews

This reserved article originally appeared in CHROMiX ColorNews Issue 65 on October 10th, 2018.

Click here to see the original in its original context.
Email
colornews(at)chromix.com to subscribe to the ColorNews newsletter.


Do ICC profiles correct the color back to what it’s supposed to be?


Long-time readers of the CHROMiX ColorNews have enjoyed the benefit of several years of excellent news, advice, and suggestions on how to make the world of color a better place. In particular, I have long held that being familiar with the “myths” of color management can go a long way toward helping newcomers avoid pitfalls and embarrassing mistakes. Some of these myths keep popping up, like a whack-a-mole game at the county fair. This article harkens back to Myth #2, from way back in 2003. As this issue continues to resurface, it’s worthwhile to revisit it:

Myth #2: There is some internal Lab / color reference that the output of printers is compared to when profiling.

http://www.colorwiki.com/wiki/Color_Management_Myths_1-5

Another variation of this myth is: “ICC profiles are sort of like filters that correct the color back to what it is supposed to be.”

The answer to this myth is: No, there is no “ideal” reference that your printer is compared to. Rather, the unique behavior of your printer/paper/ink combination is mapped out, and the image going to your printer is rendered to take advantage of how the printer prints.

Groovy! More colors man!

Or, as Steve Upton originally put it: “Unlike a strict, conformist military academy, profiling a printer does not find out how a printer performs and then force it to conform to a certain behavior. It’s much more like a hippie commune.”


This remains as true today as it was then. Let’s dive into the explanation a bit deeper. When we make a profile, we will send the printer a profiling target image - a sampling of all the colors under the sun (actually, all the colors the printer can print). This image will have a reference file associated with it, often in the form of a .txt file.

If you take a look at a CMYK reference file in a simple text editor, you’ll see that the bulk of it only contains numbers from 0 to 100 (0 - 255 for RGB). These are machine numbers (what we call “device” values) that by themselves don’t relate to any Lab values (which define colors.). http://www.colorwiki.com/wiki/Lab_is_warped They are literally just numbers that tell a machine how much ink to lay down. These numbers - by themselves - don’t define a specific “color”. Note that these values are part of the target image itself. In Photoshop, you can use the eyedropper to sample a colored patch in a profiling target and confirm that the patch has the same RGB or CMYK numbers called for in the reference file. Now, of course, those numbers would result in very different colors if sent to different kinds of printers / papers / inks (think fine-art inkjet paper vs. newsprint).

Notice that with these device values in the target reference, there is nothing there which defines actual colors. So, even if the system WERE set up to compare to some magical standard, there is nothing in the system capable of identifying actual colors to which it would compare.

Device values can be likened to the gas pedal on a car. The position of a car’s gas pedal does not tell you how fast the car is traveling; it really only determines how much gas/air goes to the engine. Other factors are involved in determining speed. You could be going slowly uphill or coasting downhill. Or, like me, you could have forgotten to release the emergency brake. It will take something else, like a speedometer or cruise control, to make determinations about speed. Like a gas pedal does not determine speed, device values do not define color.

Gamut boundaries

Images that you send to the printer might include colors that are in gamut or out of gamut. And yet, you’ll still want the printer to be able to print ALL colors that are sent to it as best as it can, whether the colors are in gamut or not. You will want them to be moved into gamut so that they will be printed as close to the original as they can be, or at least be printed in a pleasing way. This “make the most of what you’ve got” behavior is incorporated into what the profile does.

If you’re looking to clothe your lanky teenager, you don’t compare him with some size standard he’s “supposed to be.” That would not provide you with any useful information for your task. Rather, you measure him and then buy him clothes that fit, because you want him to be properly clothed, regardless of what unique shape he is growing into. In fact, measuring him accurately can provide very handy information. (You can use this to figure if he’s tall enough to wipe down the dust on top of the cupboards. What a useful purpose for those long arms!)

You get your choice of rendering intents man!

In the same way, we want to know about the color coming through your printer. Whatever odd gamut it may have, we can know about that and use it as necessary to print similarly situated colors in your image. If you have out-of-gamut elements in your image, a well-designed profile will move those colors into gamut in an intelligent way, hopefully retaining the same hue and as much of the “look” of the original color as possible. Like a hippie commune, the in-gamut colors are allowed to “be themselves.” Only when no path is available to produce the out of gamut color is coercion invoked to move a color into a printable space. Even then, the changing of color is done according to one’s choice of rendering intent. Groovy. Freedom of choice, man!

In the end, a well-made profile will handle everything you throw at it with aplomb, and we at CHROMiX celebrate and honor its own irreplaceable, unrepeatable, indispensable, wonderful uniqueness!


Thanks for reading,

Patrick Herold
Tech Support
CHROMiX, Inc.



Personal tools
Namespaces
Variants
Actions
Navigation
Toolbox