X-Rite ColorMunki Photo Color-Management System

Get the Best Digital Photos You Can by Calibrating Your System

By Ctein Back to


Color management is a serious— if not downright painful— business. Without good color management, the hues you see on your computer screen won’t be anything like the ones you thought were in the photograph, and what winds up on your print will look even worse. If you’re sharing files with someone else, they won’t see the photo you do.

X-Rite’s long been a serious player in the photographic control business, from traditional densitometers and sensitometers through computerized control. They now own Pantone, long a source of “off icial” colors. The fruit of their joint labors is ColorMunki Photo (Figure 1). It’s compact! It’s cute! (It doesn’t look anything like a simian, though.)

The $499 package includes the ColorMunki puck, case, USB cable, and installation CD. It runs under Windows (XP or later) or Mac OS X (10.4 or later, G5 or Intel). ColorMunki makes profiles for people who loathe color-management theory (nothing wrong with that) and may not even know what a profile is good for.

Until you start using some form of color management, you can’t reliably produce good images on your computer monitor or good prints from your printer. That’s true whether you’re doing color or black-and-white. In fact, it’s probably easier to understand in black-and-white terms.

Monitors and printers aren’t perfect, but they get better with each successive generation. Unfortunately, better means different and different means that the photographs won’t look or print the same way on a new system as they did on your old one. Let’s suppose you’ve got a photograph that looks beautiful on your current monitor. Suppose you get a monitor next year that’s twice as bright and has twice the dynamic range. If you display the same photograph exactly the same way, will it look the same? No, because the whites will be twice as bright as they were before, but the blacks stay just as black and everything in between will get stretched. Your brand-new monitor shows a brighter, contrastier photo than your old one did. Similarly, a photograph that prints beautifully now can look way too dark if you get a printer that can print deeper blacks and shadow tones than your old one.

A profile translates the photograph so that it looks the same no matter where it’s displayed or printed. It takes each gray (or color) value in your photograph and converts it into output data that will look correct. The translation is specific to the output device. That set of conversion instructions is called a profile. It ensures that your output device, no matter how idiosyncratic, produces the most standardized colors it can.

Manufacturers usually provide profiles with the products, but they can’t take into account slight product variations or exactly how you’ll use them. They also can’t take into account what happens when you decide to try a new art paper that just came out or what happens as your monitor ages and its rendition shifts. Dealing with all of that requires custom profiles: profiles that are made for your particular computer system and are updated periodically to account for changing conditions.

Getting up and running

ColorMunki has veered perhaps too far in the direction of “simple.” The substantial looking 48-page manual contains only the most minimal instructions and it doesn’t explain what you’re going to be doing with it or why. Installation was trouble-free, I’m happy to say, although the Mac version installs a menu-bar icon without asking and fails to provide a choice of install locations. Poor form for professional software. I should mention that using ColorMunki requires “activation” and you are allowed to use it on no more than three machines. Activation does not involve any offsite connections or intrusive measures. One mouse click and it’s over, instantly and painlessly. Should you wish to move an activation to a different machine, you can deactivate an old one. That part is nicely done, X-Rite.

After I launched ColorMunki, I clicked on the Help menu item. Did that bring up any online help? No; it took me back to the X-Rite Web site’s online documentation. It’s some of the poorest Web design I’ve seen. There is no coherent or downloadable manual, no global table of contents, no index. An extensive series of pages and sub- pages each address a single isolated topic. There are no navigation buttons on each page to take you back to the previous page or the home page. Some text has embedded links that are not called out in any way. The only way I discovered they were links was if I happened to scroll my cursor over them. I felt like I was playing a text adventure game. The text is gray-on- gray, which is ergonomically poor for reading. I designed better Web sites a dozen years ago.

At this point I was starting to feel moderately grumpy. Fortunately, and in marked contrast, the in-program instructions get an A+. The opening screen gives very clear choices. There are short animated movies for each choice and each written instruction that are concise and helpful. There is visual on-screen feedback—for example, when the software wants you to switch the ColorMunki to the “calibrate” setting, a diagram shows you the current setting of the ColorMunki control dial and indicates how you should turn it. When you turn it, the picture changes in synchrony. The software will not let you make a mistake, and it holds your hand in a way that is not annoying. It more than made up for the unpleasant online experience.

Monitor profiling

Monitor profiling with ColorMunki really is simple. The easy mode requires no decisions whatsoever and calibrates to D65 (6500°K color temperature, a common industry standard). An advanced mode lets you adjust the contrast and brightness of your monitor (if it has such adjustments), take ambient lighting into account when optimizing your display, and choose D50 or D65 color temperature. D65 may be industry standard for graphic arts purposes, but D50 is better for art photographers like me. Prints are very rarely viewed under D65 conditions. D50 is a closer match to normal daylight viewing, and incandescent light is actually down around D30. Photos that look good on a D65 screen look too warm in prints viewed by daylight and are way off in incandescent-light viewing. D50 is a good match to daylight and substantially reduces the visual difference between the screen and in an incandescent-lit print.

ColorMunki runs the show. It instructs you when and how to set the puck to calibrate itself, put it in its padded case and position it on your computer monitor. The case strap is weighted to help keep the puck in position (although it needs a few more ounces in the strap). ColorMunki displays a series of predefined color patches on the screen, measures them, and calculates the adjustments needed to produce correct color.

At this point you have a finished profile, and ColorMunki becomes a little too simple. You can compare the “before” and “after” results of profiling, but if you accept the results, it automatically saves the profile and makes it the system profile. If you want to try calibrating a couple of different ways and then compare them to each other, you won’t find an easy way to do this, because when ColorMunki creates a new profile it overwrites the old one. I know how to edit profile names. The folks that are most likely to buy ColorMunki won’t.

By comparison, Profile Mechanic from Digital Light and Color (http://dl-c.com/content/ view/48/76/), reviewed previously in these pages (January/February 2005), lets you assign a unique name to a profile when you save it. It also provides information on the color gamut that the monitor displays and the corrections that are being applied if you’re so interested (Figure 2).

Figure 2. This chart shows the color gamut of the MacBook Pro laptop display (white triangle) as compared to the standard D65 display gamut (black triangle). The laptop display is incapable of displaying most saturated colors; its color range is severely limited, making it difficult to calibrate. This plot was produced by Profile Mechanic; ColorMunki lacks this capability.


Figure 3. ColorMunki did a better job of profiling my Apple 23-inch Cinema Display than Profile Mechanic did. The contrast and saturation in the gray scales shown here have been greatly increased to make visible the differences between individual steps (which only differ by a single value unit). The top scale is the standard Apple profile. The middle scale was rendered with Profile Mechanic’s profile. Some of the steps are nearly indistinguishable and most of them deviate from good neutrality. The profile produced by ColorMunki (bottom) has evenly spaced and neutral steps, with very little residual color cast.


Certainly not everyone is going to need all of these features, but most people will need or want some of them. There are no additional information or control options in ColorMunki; the simple (and inflexible) way is the only way. I’d like a bit more of its smarts to be accessible.

ColorMunki lacks a CRT monitor choice, but setting it to LCD works just fine. I used ColorMunki to profile my RasterOps 21-inch CRT monitor on the Windows machine and my 23-inch Apple Cinema and 15-inch MacBook Pro LCDs. ColorMunki and Profile Mechanic were evenly matched on the CRT, creating high-quality profiles at both D50 and D65.

Profile Mechanic didn’t do well with my 23-inch Apple Cinema Display, even though that’s a full- fledged 24-bit display. There were problems both with grayscale neutrality and some excessively sharp breaks between tones and colors in what should be smooth gradients (Figure 3). ColorMunki had no such problems. There was a slight difference in overall color cast; ColorMunki produced a screen that was 3–4 CC more red-magenta than Profile Mechanic. I didn’t care for what seemed to me a slightly pinkish cast, but it’s not a big difference and the overall rendition of colors and tones was better. The D50 Color- Munki calibration looked ruddy to my eye, but it seemed to match the prints well, and that’s the whole point of it.

Profiling a laptop display is something of a fool’s errand. With the possible exception of a few recently announced machines (and I emphasize “possible,” because I’ve not inspected those displays), laptop displays can’t provide accurate color or tone. The screen gamma varies with vertical viewing angle. Even with your head in a fixed position, there will be a substantial variation in tonal placement from the top to the bottom of the screen. Laptop screens have a poor color gamut (Figure 2). Most saturated colors fall outside the screen’s range. Further, laptop screens (and inexpensive desktop LCDs) do not display full 24-bit color; they fake it via a combination of display tricks including dithering.

Figure 4. Laptop displays are inherently bad at displaying accurate tone and color, but a good profile makes it better. On the left is the standard Mac OS X profile for a MacBook Pro display. In the center is the display after profiling with Profile Mechanic; on the right, after profiling with ColorMunki. Both custom profiles produce more accurate colors and reduce harsh transitions.


Figure 5. Even ancient printers can be improved by profiling. The figure on the left shows the unprofiled output of my HP DeskJet 970. Colors are dark and poorly distinguished, the grayscale is uneven, and there are sharp transitions in the gradients that shouldn’t be there. Profiles from Cathy’s Profiles (center) and ColorMunki (right) are immensely better. They don’t look the same but they are equally good.

Figure 4 shows that the Mac OS X standard profile (left) does a really bad job of handling color. It produces an extremely neutral grayscale from one end to the other, but individual colors are not accurately mapped and there are harsh transitions as colors get more saturated. By comparison, both Profile Mechanic and ColorMunki do a much better job overall with color. Subjectively, I would say ColorMunki is superior; there were fewer visible artifacts in real photographs. Although the color and tone rendition weren’t close to being accurate, it would be usable for rough work.

Where both third-party profiles failed was grayscale rendition. They went significantly pink in highlights. This is not noticeable with color photographs, but it is extremely distracting when working with black- and-white photographs or simply looking at ordinary display windows. The Mac OS profile is smart enough to keep the grayscale neutral, but these two programs aren’t. Fortunately, it’s easy to switch between display profiles, so I’ll keep my laptop set to the standard Apple profile except when I’m viewing or working on color photographs.

Print profiling

Printer profiling is literally a hands-on operation. The ColorMunki prints out a color test target (Figure 1). The ColorMunki puck works as a handheld scanner: place it at the end of each row of color chevrons in the test chart and slide it down the paper. Don’t move it too fast; three to four seconds to scan a row is about right. The software reports if the scan was successful and moves on to the next row; if it wasn’t successful, you try again. I’d usually get four out of five of them on the f irst try. It never took me more than a minute or two to scan a page.

The software analyzes the scans and prints out a second color target to fine-tune the data collected from the f irst target. After you scan that one, the software presents you with a finished profile. Unlike Monitor profiling, you can name this prof ile with a custom name of your choice. My convention is to use printer model, paper type, and quality setting (e.g., “R2400 Harman Gloss FB Al PhotoRPM”). My profiles then appear in selection menus grouped first by printer, then by paper type, and then by print quality.

ColorMunki gives you the option of assigning any profile you create as the default profile for your graphic arts applications. I didn’t take advantage of that feature because I’m always switching printers and papers. If most of your work is done with one combination it’s a nice touch.

A most unusual and valuable feature is that you can take an existing profile and further improve it. Select the optimization option, tell ColorMunki which profile you want to improve, and direct it to a TIFF or JPEG image that you’d like it to analyze. The software looks for the important colors in the image and creates a new custom target that focuses in on those portions of the color space. Print and scan the target, and ColorMunki uses that data to reduce residual errors in the initial prof ile. You can use multiple images with successive optimization runs.

While home-monitor profiling is a necessity, I’ve tended to be skeptical of home-printer profiling. The low-cost kits out there rarely do anywhere near as good a job as professional-profiling hardware and software. They’re better than nothing, but there are better, more cost-effective options, unless you have an awful lot of printers and papers to profile.

I regularly make use of Cathy’s Profiles (www.cathysprofiles.com ). Cathy charges only $35 per profile. If three or more profiles are purchased at the same time, the first two profiles cost $35 and all the additional profiles are only $30 each. Furthermore, she provides three different mappings of each profile and will help you troubleshoot your workflow. Check out the details on her Web site. I’ve been recommending her for years, and I’ve never heard of a dissatisfied customer.

I set ColorMunki loose on four printers—the Epson Stylus Photo R800, R2400, and R2880, and an ancient HP DeskJet 970. Why the DeskJet? Because it’s hard to prof ile. A decade ago it was a fine machine, but the dye set is kind of weird and has a very peculiar gamut in the warm colors. I had Cathy do a profile for me several years ago as an experiment, and it greatly improved the quality of output.

I chose the R800 because it’s not strictly a CMYK printer; it has red and blue links as well. The Epson premium profile for this printer is lousy. It was going to be interesting to see how ColorMunki fared.

The R2400 and 2880 are standard fine-art printers for 13×19-inch and smaller prints. I’ve got the Epson premium profiles for the R2400, which are so-so, along with much better profiles from Cathy.

Figure 6. This is a difficult photograph to print because the important sky colors reflected in the waterfall at the limits, or even outside, of the gamut of CMYK printers. All three prints here were made with the Epson R800 printer. The top print was made using Epson’s “premium” profile, downloadable from its Web site. This profile consistently produces results that are too cold/magenta in the green-blue part of the spectrum. I have not been at all happy with it. The middle print was made using the profile that ColorMunki generated from its standard test targets. It’s pretty good (it also looks very similar to the print made with the Epson R2400 using the Epson premium profile). The bottom print was made on the R800 with an optimized profile created by ColorMunki. This is superb—as good as or better than the custom profiles I’ve had made for my printers. The results for the R2400 and 2880 printers were similar.


I went at the DeskJet first, just because I thought it would be fun to see how badly things broke. Figure 5, left, shows the unprofiled output from this printer. There are many obvious problems: Greens that are way too dark, weird orange contamination in the yellow part of the spectrum, no lavenders at all, and abrupt transitions between tones and color. Cathy’s profile for this printer really improves things. Lavenders appeared, greens got a lot lighter, and most of the sharp breaks went away, except for an unusually sharp transition from red to yellow. Overall, a huge improvement.

ColorMunki did as well as my custom profile. The results were different: purples, lavenders, reds, and yellows were substantially better in the ColorMunki profile, but greens and cyans fared worse. There were also more abrupt changes to the darker colors. The two profiles have very different “flavors.” If you were printing people pictures, ColorMunki would do a better job; if you were printing landscapes, Cathy would be the way to go.

On to the R800. As I mentioned, this printer has an unusual ink set (as do the Epson R1800, 1900, etc.). The Epson premium profile that one can download from their Web site is very unsatisfactory; see the top print in Figure 6. (The standard profile that ships with the printer is even worse.) Cyans shift strongly to the magenta, coming out cold blue at best, and greens lack definition. ColorMunki created a much better prof ile from its standard test targets; the middle figure has much better overall color and much more believable sky hues, although they are still too blue.

Optimizing the profile using this photograph (and two other similar landscapes for my portfolio) got me to the print at the bottom of the figure. The sky color is now excellent and subtleties in the blades of grass are rendered more accurately and clearly. This optimized profile is producing more accurate color and tones on the R800 printer that I’ve seen with any other profile.

The results with the R2400 and R2880 printers were nowhere as dramatic, because the Epson-provided premium profiles aren’t as bad, just so-so. The premium profile from Epson produces prints that look very much like the middle print in Figure 6; the profile from Cathy produces much better prints. But I have to say that the multiply optimized profile I made myself using ColorMunki was superior even to hers for my particular work. Understand that I am really splitting hairs here; they’re both excellent, and there’s no objective way to choose between them.

What I find remarkable is that I can produce prints that equal or better the ones I made using prof iles created with a professional profiling setup by an expert. A gadget that sells for less than $500 that can do this in my own home is a remarkable achievement. ColorMunki is an outstanding product.

About the Author

Ctein is a technical writer and expert printmaker. He is also the author of Digital Restoration and Post Exposure—Advanced Techniques for the Photographic Printer.