Color management systems such as Apple’s ColorSync and Microsoft’s ICM 2.0 are touted by their creators as easy, automatic tools for matching colors from your original, through your scanner, to your monitor, to your printer. In practice, they’ve proven to be complex and often unwieldy, but one thing is clear: color management lives or dies by the accuracy of the device profiles it uses.
If you’ve been struggling with color management and failing, it’s quite likely that a major part of the problem is inaccurate profiles, particularly if you’ve been relying on the generic profiles shipped with many low-cost color printers, scanners, and monitors. Custom profiles made for the actual equipment you use can produce much better results than you’ll ever get from a generic profile, but how can you obtain custom profiles to replace the generic ones? (See sidebar A Profiling Primer for an explanation of the profiling process.)
Not so long ago, getting consistent color from your printer required a substantial investment of time, money, and intellect. Building a printer profile that ensures predictable color from your printer has been a matter of buying a $1,000 software package, a $2,000 handheld measurement device (up to $2,500 for an automated spectrophotometer), plus a good chunk of time and a fair bit of rocket science. It’s hardly surprising that most normal mortals were reluctant to invest the time in the learning curve involved and to spend upwards of $2,000 to profile a $500 inkjet printer. (See sidebar “Fast, Easy, and Cheap: Color Management’s New Generation.”)
There had to be an easier -– not to mention cheaper — way. Monaco Systems pioneered the concept of using an inexpensive everyday flatbed scanner instead of a costly dedicated spectrophotometer to take the measurements needed to build a printer profile — it wasn’t the first company to announce such a product, but it was the first to ship one that worked. But since Monaco released its $299 MonacoEZcolor package, the company has faced competition from two other vendors with scanner-based profiling packages: Horses LLC’s $229 MatchLock Profiler II and Praxisoft’s $79 WiziWYG.
But do these things actually work, and if so, how well? And just how easy are they to use? To find out, we used each package to build profiles for an Epson Photo Stylus EX inkjet printer. For comparison we also built a profile for the same printer/ink/paper combination, using a high-end measuring instrument, the GretagMacBeth Spectrolino/SpectroScan, and a high-end profiling package, GretagMacbeth’s ProfileMaker Pro 3.0.
This article was last modified on January 18, 2023
This article was first published on May 10, 2000
Commenting is easier and faster when you're logged in!
Leave a Reply
same
Here are a few comments from someone who has been watching the debate about low end profilers from the sidelines. My perspective is color photographic printing (Nikon LS 2000, Epson Photo EX).
First, I am a big fan of Bruce Fraser’s. Blatner and Fraser’s Real World Photoshop 5 is the bible for people like me. I read his articles avidly and have no disagreement with this one.
Second, I agree that color accuracy is important, that having your monitor reasonably calibrated is key, etc. But, I have found that there is an even more fundamental requirement: perceptually smooth mappings near the gamut boundary.
I have been plagued by discontinuities in the Epson ICCs that show up as posterization in faces and exaggeration of noise in shadows and in certain browns and oranges.The inability to soft proof RGB in Photoshop caused me to use Epson’s RIP so I could softproof the CMYK before wasting ink. Using the LS 2000 with multiple scan averaging and a 12 bit/color workflow has helped enormously. (Please, Adobe, let’s have all the tools and filters work at 12 bits in the next Photoshop release – and RGB soft proofing!!) Nonetheless, continued problems motivated me to look closely at the ICCs and at the ICC Standard Document.
From Andrew Rooney at http://www.digitaldog.net I finally learned how to set up soft profiling ICMs for RGB. Then using Steve Upton’s “Gamutvue” techniques (www.chromix.com) I learned to display the Lab space in a series of luminosity slices. Softproofing the Epson ICCs in this way exposes the discontinuities dramatically as vertical/horizontal bands in slices around L=.5-.7. I then downloaded Cone’s free ICCs for Epson OEM inks and papers. These show up as perceptually smooth in soft proofs both of the Lab slices and on previously problematic images.
A look at the ICCs themselves suggests what is going on. For the BtoA0 tag (Lab to device, perceptual intent), Cone has 33 grid points, Epson 17. In the 3 colors this means Cone has 7 times more data. In principle, this allows for a smoother mapping at the gamut boundary. In practice, the extra data may be only part of the story; more sensible smoothing algorithms may be involved.
I suspect a problem is the ICC standard requires a single grid spacing across the whole space. Perhaps more grid points at the boundary would make the perceptual transition smoother?
Finally, it is troubling that something as important as ICCs are black boxes that require heroics to understand what they are doing. I am sure there are better tools associated with high end profiling activities. What we need in the desktop world are tools that display how an image is being mapped along any axis through the 3D Lab space. This would show up where the discontinuities are and allow us to reject bad ICCs and make corrections using Photoshop tools. It would also be useful for such tools to show where a particular point in an image sits along such arbitrary axes, to do histograms along the axes, etc. Some basic 3d visualization techniques would be extremely valuable and this is another wish for Adobe in the next Photoshop release.
As Bruce has pointed out, there is no defined way to handle Perceptual rendering. We need the tools to monitor and control such rendering.
It’s tough to please everyone when sharing results from scientific testing of any kind. For years, I was a black and white film sensitometrist – studying and testing photographic materials. My experience there tells me that sometimes results are what we want to see, not what they actually are. I believe in a measured approach to testing materials – part scientific and part empirical. That is to say, after following the scientific methods, play with the stuff in real world environments.
Using this approach, we dove into color management, first testing Colorblind, then Kodak colorflow, then Matchlock. Since our goal was producing pleasing photographic images – we did the science, then we just played for awhile.
Our results? We like Matchlock. The interface was easy and intuative, the resulting profiles were “tweakable” – and the process a snap.
Since we have been testing, we have been able to almost instantly generate profiles with Matchlock, and print on exotic papers with several inks. It works for us, and we like it.
Hugh Milstein –
CEO Digital Fusion LLC
I cannot make a comment on EZColor as I have not used it, but Profiler II made substantially better profiles than WiziWyg did – on any paper.
It cost more than Wizi, but if you value quality profiles it is definately worth the increased price.
The colors were more accurate, much better shadow detail. On all papers that I use the Profiler prints were easily better than Wizi’s. Customer support from Horses is also #1. It goes without saying though, that you get what you pay for!
I must say that I have mostly printed photos though and cannot comment on Bruce’s statement of vector based images.
Murray Z.
Attempting to write an objective review of something as fundamentally subjective as color image reproduction is a process fraught with danger, but I knew that going in.
I’m not particularly surprised that others have had very different experience from mine with these products: there are at least two huge uncontrolled variables-the scanner, and the printer/paper/ink combination. That said, I stand by the results I obtained with this particular scanner (a UMAX Powerlook 1100) and printer (Epson Photo Stylus with claimed-archival inks and paper from MISsupply).
I note that no-one who has complained about my treatment of Matchlock Profiler II (which in case it isn’t abundantly clear from the review I consider to be a fine product) made any mention of using the profile to reproduce anything other than photographic images, which it handles well. I use color management on images a great deal, but I also use it on vector art, logos, tint builds and other page elements, and in my (continuing) experience, it doesn’t do nearly as good a job on these types of material.
I made a point of sharing the results of the objective test with each vendor (identifying only that vendor’s profile by name in each case) and none of them voiced any concerns about its accuracy. The Perceptual renderings are by definition a much more subjective case.
I gave Monaco EZColor a very slightly higher score than the other two products because it was the only one that gave good results with both perceptual and colorimetric renderings, emphasizing in the text that if all you care about is image reproduction, the objective test doesn’t matter that much. I gave WisiWYG the same score as Profiler II because, for less than a third of the cost of Profiler II, it produces a scanner profile as well as a printer profile, and because it did a substantially better job of colorimetric rendering than did Profiler.
For the record, I created about a dozen printer profiles with each package, and I’m quite confident that were I to repeat the tests today, I’d get the same results with the same equipment. Your milage may, of course, vary.
On a couple of specific points:
Bernard Moritz complained that neither WisiWyg nor EZColor allows you to control total ink limit. After working with six or seven different RIPs for low-end inkjets over the past couple years, I’ve come to the conclusion that, unless you’re prepared to discard the printer’s proprietary screening and use conventional halftones, you cannot control these printers at the ink level (and with the six-color photo printers, not even if you discard the printer’s proprietary screening). The real ink limits are set by your choice of paper type and dither. Putting ink limits in the profile will certainly influence the outcome, but it simply isn’t possible to make these devices produce a pure cyan patch, for example. In that light, I’m tempted to consider the lack of ink controls a feature rather than a shortcoming.
Finally, Robert Snow and ncm both faulted Monaco’s technical support. Unfortunately, I’ve talked so much with all three companies over the years that the person who answers the phone usually recognizes my voice, so I was simply unable to make any honest evaluation of the tech support that civilians would normally receive. I regret this.
I haven’t used EZcolor or WiziWYG, so I can’t comment on them, but I’m surprised that Mr. Fraser got rather mediocre results from MatchLock Profiler. I’ve used Profiler since January, on 2 different Epson printers with 2 different ink sets, and on 6 or 7 different papers. I’ve found it gives remarkably good, consistent results, with excellent shadow detail and rich colours, nothing like the dull and muddy examples shown in the article.
Having read much of what Bruce Fraser has to say on Color Management and being very appreciative of his invention of the “Bruce RGB” color space, I was more than a little surprised to read his review of EZColor, WiziWYG, and Matchlock.
I’ve used all three products and while each has it’s merits, I have consistently obtained the best results with Matchlock. EZColors’ 1.5 release has jumped them ahead of Wizi IMHO and is very respectable, but it still doesn’t give me the kind of subtle shadow detail I get with Matchlock–even unadjusted.
I _did_ have one experience that I would call similar to Bruce’s. One profile I made for the Epson 1270 caused a noticeable flattening of tonal range and a loss of color. I re-made the profile after giving the test target overnight to “settle out” and it came out spot on. I don’t know why the process failed that once or whether letting the test print age awhile made any difference, but the problem was easily corrected. Since Bruce’s tests are very carefully documented and based on measurable output I assume it’d be easy for Horses (or anyone else) to see if they come up with similar numbers or if there is a possibility of something having gone awry.
In any case I’m always interested in advances in color management, so thanks for printing a review of these products.
–Dave Cardinal
Cardinal Photo
https://www.cardinalphoto.com
Having just finished 15 days of printing test for a user group program on Printer, Papers, and Profiles..I disagree with Bruce Fraser’s evaluation of the profiling products. Ez color never worked well for me although I don’t have the latest version and its profiles are almost 1mb in size. Matchlock profiles are 70 to 108 k in size. WIZIWYG worked resaonably well but I did get some “transition” problems” especially in the darker grays. I did not work at all well with my Sienna Fotoprinter with photographic chemical prints. I have always had a problem with saturated yellows on the Fotoprinter where, at a certain point, they fall off a cliff to a gold band.Matchlock was a winner all the way!! It even solved the yellow banding. Smooth as silk. So incredibly simple and it works.
For 32 types of media from canvas to art felt to glossy photo film I only hand tweaked 1 profile..to add some saturation. All with a $99 scanner and no targets to fade or get lost.
I found it hard to believe but it worked. It was also the only package with the controls to fine tune the profile built in..a big plus! Finally the CMYK profiles worked extremely well for simulating professional proofer output on the inkjets. I have used the profiles thru a PowerRIP 2000 set up and been very happy.
I also would reverse the rating which Bruce gave the products. Matchlock 90, EZcolor and WIZIWYG 75.
After trying all three of these packages, I must disagree with the findings. Matchlock gives me truer color, more open shadows, and one can tweak the profiles. Normally, the profiles are dead-on the first time unless I am profiling non-coated art papers. For a low-end profiling system, I just don’t think Matchlock can be beaten.
Finally, try for technical support from either of the other two companies! It is practically non-existent. With Horses, rapid and helpful response is almost immediate. For me, at least, this is of paramount importance. They bust their buttons to solve any problems.
bob snow
Andrew Carpenter wrote:
“Bruce’s essay is informative and well-
organized. However, his “objective”
evaluation differs greatly from my
“subjective” comparison of Praxisoft’s
Wiziwyg and Horse’s Profiler II.”
I think I was fairly explicit about the
differences between the objective and
subjective evaluations. I stated flat-out
that the objective eval says nothing
about the profile’s ability to reproduce
images. Absolute Colorimetric performance is
important if you’re trying to reproduce
spot colors, or you’re using the printer as
a proofer, but it’s only one component.
Evaluating perceptual renderings is necessarily
a subjective effort, which is why I chose to show
results I feel are representative of each package,
and allow readers to draw their own conclusions.
There’s also a huge uncontrolled variable,
which is the scanner. Obviously, different
scanners (and different scanner settings)
can have a huge impact on the final result.
Finally, one of the things factored into the
rating was price…
I think this type of article is useful for all the people now getting into digital image making and printing for the first time. But an important issue – especially with something as complicated as colour management – is not addressed. This is comparative user support from the makers of the products reviewed.
I don’t know if my experience is typical but it has led me to abandon Monaco EZColor totally. If you cannot use the product due to a serious problem (in my case creating a profile for my Nikon LS-2000 filmscanner) that prevents the program from working as required, and get no feedback after repeated requests for technical assistance…. then the product is useless. Monaco tech support is literally non-existent. At least that has been my experience. On the other hand Horses matchlock has excellent support. I turned to that product – highly recommended on the Epson Ink-jet email list among other places – and found it created better printer profiles than Monaco EZColor and all my questions regarding the product were answered very quickly by its tech support – the very person who happens to have written the code for Matchlock. I have decided that I don’t need to profile my filmscanner in order to make accurate prints. And although Monaco worked for creating a profile for my monitor I have since purchased a more complete calibrating/profiling solution, OptiCal and so don’t use Monaco at all anymore.
It is too bad when a decent product suffers from the insolent attitude of its manufacturer who obviously saves it’s support for the high-end products and leaves us “low-enders” out in the cold.
Bruce’s essay is informative and well-organized. However, his “objective” evaluation differs greatly from my “subjective” comparison of Praxisoft’s Wiziwyg and Horse’s Profiler II.
In general, to my eyes, Horses provdies a more lifelike and detailed print — it is much less “congested” and “muddy” than my prints made with Wiziwyg, which tend to lack shadow detail and have a lot more shadow posterization.
For example, here are some differences in two prints of the photodisc test target. Both were printed on my Epson 750 using Generations ink (beta black) on Media Street’s Glacier Matte paper – my curent favorite combination for “budget archival” prints.
THE FACES:
Shadow transitions in faces: As I’ve experienced in other images, Horses opens up the shadows a lot. The Wizi transitions are more abrupt and contain less shadow detail.
The light-skinned faces: Horses version definitely more “smooth” and “natural”, Wizi version more “discontinuous” and some “ink jet artifacts” apparent — false “baby eczema”, unnatural transitions from the “shiny spot” on the third baby’s head.
I, do, however, prefer the Wizi skin tones, esp. for the first face (which Wizi prints as a beautiful “chocolate brown”). In all other respects, however, the Horses profiles wins here hands down.
GEARS: Definitely more detail and tonal range in the Horses version, esp. in the top gear. The Wizi print gives less detail in the dark spots of the lower gear — the shadows there do seem to be “blocked up” in the manner that David described.
FLOWERS ON TOP: The leftmost flower is a little orange-y in the Wizi print, more yellowish (and more natural looking to my eyes) in the Horses print. No contest for the other two flowers: the Horses print has more detail in the petals and looks more lifelike.
PURPLE STONES IN FISHTANK: Same story here: the “darker” Wizi print blocks detail and looks less lifelike. The purple color is a lot more evident in the Horses print. Even though shadow posterization was the reason I switched from Wizi to Horses, I don’t see any in either print.
COLOR OF BACK WALL: Subtle but definite greenish color cast in Wizi print; Horses print more netural.
What about the CYMK capabilities of the packages? Hint: Matchlock is the only solution of the 3 that allows you to set essential parameters such as total ink limit.
What printer / paper / ink was used for the evaluation? Matchlock has been reported to work well with certain combinations that EZ couldn’t handle (WiziWYG is so far behind that I don’t think it’s even worth mentioning).
WiziWYG produces a noticeable green cast in most images and Matchlock yields better shadow detail than either WiziWYG or EZ Color. Not mentioned in the review.
I don’t own a hardware spectrometer (just a DTP-92 for monitor calibration). But the fact that I get more neutral grays (I’m working exclusively in CYMK mode through PressReady) from Matchlock Profiler CYMK than I get with EZ Color 1.5.3 tells me that the delta e measurements in the review should taken with a big grain of salt.
IMHO the review is too shallow and the final result doesn’t match with my experiences. I would rate the products
1. Matchlock = 90
2. EZ Color = 85 (I don’t have the 1.6 version yet – it’s not available as of 00/05/13; again, this is not mentioned in the review). For CYMK work I would rate EZ = 75 because some really important controls just are not there (e.g. TIL).
3. WiziWYG = 60
Moritz
It is a nice article in covering the complicated subject area and the illustrations of the workflow in comparing among the 3 products are very helpful. As an avid photographer/ graphic artist, I have used both the high end product (>$6000) and the new simple one as described in the article. I agree with Bruce that the simple product would probably be sufficient for most users and is certaintly a lower learning curve as compared to a high end color management system. I would also highly recommend that a new user should consider doing some basic background reading on the subject of color management beforehand. In general the software documentation are usually poor and confusing especially in a WIN/PC enviroment.