MKD - Mark Kunoff Design

PATH>>> HOME / XTRAS / ICC Rant

My rant about ICC Profiles

An opinion on color profile technology
- by Mark Kunoff

Plain Text Version

Intended audience: Computer graphic artists working with CMYK printing devices who typically edit images in the RBG color space.

Before I go into my rant about profiles, I would say this: The printing market is very competitive right now (even more so these days with many firms going out of business,) and a good printing vendor will be able to match anything you render (within physical limits as I will explain later) and you shouldn't have to think or worry about it. Over the 20 years I've been in the printing trade, I have worked with many "graphic artists" who have no real knowledge of color science as it relates to 4-color process printing, or otherwise thrown around as "CMYK." Most artists I've worked with are not apologetic about it, and expect the service bureau or printer to be the experts in color science, and rightfully so - they know the vendor doesn't want to lose them as a client. These days customers are just too precious. So with that said, a really good printing service that has a knowledgeable pre press (or nowadays called 'production') staff will make this a "no brainer."

In my professional opinion, TIFF (or Tagged Image File Format) is still the best file format to deliver and it is probably the most compatible format for printing services. There have been some that claim EPS is better. Other than overiding a preview flaw with Quark Xpress, there is absolutely no reason to use EPS over TIFF. However, it's not out of the realm of possibilities for a vendor to request a layered photoshop file in it's original format (PSD.) Also, be aware that printing vendors who have been sold on the color profile workflow will defend their reasons for doing so - and with a vengeance. Bottom line is, if they can't deliver the final color you expect, you might consider another vendor. This business is about customer service now more than ever and you should take advantage of it.

The International Color Consortium(R) ( http://www.color.org/ ) explains the "intent" of their technology on the front page of their website.

With no disrespect to the ICC, the intent statement is a bit vague and deserves further explanation. The ICC name also conjures up visions of a neutral and unbiased group of international diplomats, when in fact it was founded by:

Adobe Systems Incorporated
Agfa-Gevaert N.V.
Apple Computer, Inc.
Eastman Kodak Company
Sun Microsystems, Inc.

Although I love the work that all of these companies do and prove that by giving them my hard earned dollars every year, let's face it - these guys have a financial interest in convincing you that this technology is worthy of your investment. And that in itself is not a bad thing, but it does show there are financial motivations behind why this technology has proliferated despite all of it's shortcomings. And you should know that there have been other similar technologies that didn't make it. If I remember correctly, a former Scitex engineer was trying to get the industry to accept his new profile technology called EFI (Quark featured it in version 3, I think. All I remember is we disabled it as soon as we could on all our macs.) It should be said that the technology of color profiling is well intended and in theory it was a potential solution for "guiding" technicians to better and more reliable output.

ICC profiles came into being to combat color shifts that are inherent in any printing workflow, especially between monitor and final printed product.

The most practical application of these profiles these days has to do with devices that print on unusual substrates/materials/papers such as the packaging and signage industries (this would of course include roll fed, large format inkjets.) Furthermore, ink gain (or the absorption qualities of a particular paper) can be managed with properly configured color profiles. For example, many of the newest offset printing presses use their own proprietary profiling methods.

Essentially, ICC profiles were designed to assist technicians deal with conversions from RGB to CMYK, especially when the conversion happens just prior to laying ink on paper.

If only it worked that well. Then, those technicians with less knowledge of color theory would be served by it. In my experience, this methodology is prone to errors. It has not worked as it was designed and continues to befuddle many an operator. Not to mention the tremendous waste of time and materials trying to troubleshoot them, that could be better served by understanding the science of color and it's relationship to raster printing and proofing (including inkjet printers.)

The veteran prepress guys I learned from scoffed at profiles and knew how to accurately reproduce color just by taking density readings from the images, regardless of how they appeared on the monitor. (Are you using the 'info' palette in Photoshop?) These guys also had the $10k hooded Barco monitors but would always rely on the color info readings to make their edits. They also knew paper and how the ink will behave on a certain stock.

The most snobbish of color scientists I know would call the use of profiles, "cheating." The color profile workflow basically relies on "filtering" (it applies a color correction curve) to adjust the image for a particular printing device to (hopefully) avoid any shift in color. But there are situations that no color profile can fix and in fact, make worse. It's ironic that in the real world, those that understand color are really best qualified to write a profile, but any color technician worth a salt will know that not one curve fits all situations. I have heard some sensible arguments why this makes sense for a CMYK device that, for example, uses an alternate cyan and magenta ink. It has also been stated that a profile associated with a particular paper type helps to manage how much ink hits the paper. But as I stated before, no one profile works for all images on a particular stock. Typically, this fact is realized when the final output isn't what the artist or technician expected. Now the technician has to do a correction based on the errant printout controlled by the profile. Now imagine the confusion of trying to color correct for a multitude of inaccurate color profiles - one convoluted mess!

Most experienced artists/technicians have experienced color shifting, at one time or another, on their final printed pieces. It's also a fair bet that these shifts have tended to be bright reds and blues.

Here's a test you can try at home: (Eventually I'll try to convince you Photoshop is the only pixel 'editor' you should use. Use your favorite editor for now.) Start a new small blank image. Create a fill color that has R=255, G=0, B=0 which is the purest red you can achieve in the RGB color gamut. Take a look at the monitor and how brilliantly red it is. Now, convert the color mode of your image to CMYK while keeping a close eye on the monitor. You will see the formerly bright red get darker and duller.

(Note: If you have some really funky gamma settings resulting from your monitor calibration then this test may not be as "apparent." On the other hand, if you are not able to see the shift, then this might raise a red flag whether your monitor is properly calibrated. )

The CMYK color space is a "lake" of possible color combinations compared to the RGB space which is as wide as an ocean.

The reason for the shift is that the CMYK color space physically cannot replicate the brilliant RGB red because it is outside CMYK's own limited gamut. Furthermore, this is because of the science of RGB color which operates on "additive" color theory. Alternatively, CMYK color operates on what is known as "subtractive" color theory. When you are looking at a image on the monitor (regardless of what color space the *file* is in,) what you are seeing is generated from light that is emitted from the monitor into your eyes. Process color (CMYK) works much differently. In printing, what you are seeing is generated by light that is *reflected* off a surface, like a sheet of paper. Your monitor produces a pure white light by combining the red, green and blue channels at full strength. If you saturate the paper with red, green and blue ink this will produce (albeit a muddy) black. You see, this is almost a mirrored effect. Well, not exactly, but it helps to illustrate the point. Of course, in printing you're really dealing with cyan, magenta, and yellow (and black, otherwise we'd only achieve a muddy black) and the reasons for this also have to do with the science of subtractive color.

So beware of those brighter reds and blues (and green too) when you compose your RGB based image. If you have hues outside of the CMYK gamut (Photoshop provides 'alarms' for this,) you will get a shift on your printed piece ***guaranteed*** regardless of whether you have the perfect programmed profile or the world's best technician. However, a knowledgeable operator who knows about RGB to CMYK conversions will be able to get you the best dynamic range possible. Profiles, because of their one-size-fits-all approach, are a losing proposition.

There are ways around the inherent limitations of process color (usually too cost prohibitive) such as custom ink "touch plates", where you can supplement cyan, magenta, yellow and black with a custom ink (like the ones made by Pantone, Inc.) Actually, a 5th color is very commonly used in hi-quality print jobs like automobile brochures. Wired Magazine often uses a 5th color like neon green or orange. "Hexachrome" is a very interesting evolution from 4-color process to 6-color process, but is used by only a few select vendors.

Let's be clear - the 6 and 8 color inkjets are not designed to "enhance" or "go beyond" what we already expect from process color. These extra inks are in place to deal with inherent limitations of ink jet printing devices to accurately reproduce very light highlights, usually this is in the 1-6% range for colors that need it, like cyan and magenta. The most accurate proofers and inkjets like the IRIS only use 4 inks - inks that are high enough quality that they reach highlights with no problem. This is mainly FYI for you the reader. The use of the lighter cyans and magentas is a good thing, especially for roll fed large format inkjets.

Tip: Make the Print Vendor responsible!

Unless the printer is charging you extra for it, they should do the RGB to CMYK conversions for you, and there's no better algorithm for these conversions than what is contained in Adobe Photoshop. I've worked with the world's best dedicated workstations (Creo-Scitex, Hell-Linotype, Barco) and Photoshop's conversion from RGB is well known to still be the best there is. But this is assuming that the technician who is performing the conversion has his/her Photoshop color settings configured correctly. Version 5, for example, of PS ships with some questionable default settings (which created quite an uproar actually) and too many shops didn't realize it. After version 6 this became less of an issue. Ultimately, a good technician will be able to achieve the highest dynamic range possible for a CMYK file from your RGB image using Photoshop and a properly calibrated monitor.

But, if you find the vendor is charging you to correct the color in your images, you might consider doing the conversions yourself and get to know Photoshop.

Ok, let's talk briefly about Adobe's very excellent pixel mashing application. I can honestly say that Photoshop is worth every penny. It is a masterpiece in software engineering and there really aren't a whole lot of programs out there that have that distinction. For color correction, it is unmatched (other than the $100k+ dedicated workstations and it's still a toss up.)

Furthermore, if you have invested in a pro quality monitor that is correctly calibrated, I see no reason to leave these tasks to others who don't know (or care to know) your vision of the piece. You would then deliver your own CMYK TIFF files to your service provider. But this still does not excuse the vendor from the responsibility of giving you the color you expect. The most ideal situation is if you can achieve the results on a printer at home (can be expensive for a decent device) and present an example or 'proof' to the vendor and say, "match that."

Now if you really want to learn more about "color correction" there is no better resource for you than Dan Margulis' "Professional Photoshop" book. (ISBN: 0764536958 ) and you can buy it at Amazon for a reasonable price. (usual retail is $50!) You won't find any cool special effects tutorials inside, but rather the straight skinny on how the science of printed color works and how to make it behave. Much content is centered around the use of the correction 'curve' which can be found in PS under the 'image' pulldown menu and then adjust > curves. This book is written by a veteran of the graphic arts and further expands on the failure of color profile technology and the ICC. It should probably be better titled as, "Real World Photoshop."

The invention of process color is quite magical, for it is an illusion made from the varied size and positioning of printed dots to fool the viewer into "believing" there is color.

Writer's note: Since I originally wrote this article, I have been enlightened by some of the professionals in the trade who specialize in "archival" printing, and this would include prints that are made and sold as fine art. Typically these devices are inkjets which are configured with special archival quality inks for longer life span and often, larger possible color gamut. Under these conditions, use of color profiles makes more sense. My article is based mainly on the commerical printing world.