X-Rite ColorMunki Spectrophotometer - a review

[Edward Crim]

Here it is, folks, the complete scoop.

http://web.mac.com/edwardcrim/Edward_Crim/ColorMunki_Review.html

 

[Garbz]

The thing about wideband colour spectrometers is that there's a good reason many of them are prohibitively expensive.

I haven't heard of one that can out perform a decent photodiode based colorometer (like the ones mostly used for screens) at even double the price. But since I'm looking for a cheaper spectrometer for uni based work I'm definitely happy to be proven wrong

If you still have the unit could you possibly note the CIE values of a white screen with both units at once? See if they actually differ or if the software just fouled up the ICC profile creation.

 

[Helen B]

Here's an informative and fairly comprehensive review of the ColorMunki: Northlight Images.

Best,
Helen

 

[Edward Crim]

Keith has some good photos to show how the X-Rite ColorMunki works, but all he really does is reiterate what the X-Rite promotional literature says. I didn't get any useful information from his review.

 

[Bifurcator]

I like the way you started off the review. "Color management seems to me to be a lot like voodoo; you pay your money, do the long and complicated ritual, and maybe your enemies (bad colors, in this case) roll over and die." with a huge stress on the word "maybe".

 

[Helen B]

Keith has some good photos to show how the X-Rite ColorMunki works, but all he really does is reiterate what the X-Rite promotional literature says. I didn't get any useful information from his review.

We're going to have to differ on that one. I read a lot of information in the three parts to Keith's review (intro, colour profiling and B&W) that was not in your review or the X-Rite literature. There's a lot about exactly what the processes are, how the device functions, the output resolution of the device, the format of the output data files, how the software works, how it can be used with QTR now and how QTR will provide direct support for it int he future, among other things. In short he gives an excellent introduction to the way in which the system is used and what else it can do beyond the bundled application, and what potential it has in the future.

Your review is complementary to that information, but I felt that it left more questions than it gave answers. If I was looking for a budget spectro or an abridged spectro (I use the i1 with PM5 and a bunch of other programs) what I would be interested in seeing, after reading your review, is what you think is the cause of your problems. If your review is aimed at beginners, it might be worth explaining exactly why you had to make pdf files. I'd also be interested in a quantitative and qualitative comparison between the ColorMunki, the PrintFix Pro and the bottom-end i1 bundles. At the price of the ColorMunki it could be considered as good value solely on its performance as a print profiler.

Best,
Helen

 

[Helen B]

More thoughts about useful information that could be added to a review of the ColorMunki:

Can it be used as a colour meter when hooked up to a laptop? (The i1 can, for example) This could be useful info for a photographer considering the purchase of a colour meter.

Can/Will SpectraShop, BabelColor etc read directly from it?

The output resolution is in 10 nm bands. Does it oversample or undersample - eg the i1 reads in 3.5 nm bands and reports in 10 nm bands (oversampling); the SP-68 measures 20 nm bands and reports in 10 nm bands (undersampling).

Best,
Helen

 

[Edward Crim]

Helen,

Your erudite exposition of technical terms not withstanding, I suspect that most users who would look at a device such as the ColorMunki regard it as a means to the end (accurate color in the workflow) rather than the end itself.

My only question is does it work? The answer I got was a firm "sort of".

Caveat Emptor

 

[pcummins]

Helen: The ColorMunki takes 135 different nM band samples, so depending on how much is not considered at the start/end of the samples it looks to be on par with the i1. You can use it to capture a sample of colour using the Photo ColorPicker program by placing the ColorMunki on an object and then clicking the capture button. You can't however easily take a measurement of ambient colour or projected colour (at this time). I expect improvements in the SDK (from X-Rite) will open this possibility up for other programs to use the ColorMunki directly, but at this time it's 1st party only. Note that I've been watching the USB data directly from the ColorMunki, so I'm not sure if the above information (about nM bands and sample size) is publically written down anywhere from X-Rite.

I have found that the profiles that the ColorMunki creates is very good. Printer profiles need a few iterative passes before the entire paper gamut is accomplished, so it's worth passing a few good pictures you like (in different colour ranges) to optimise the profile properly.

 

[Alpha]

Pro Photographer had a somewhat scathing review of it in their most recent issue. They said it was fine for calibration but was a pain in the ass to figure out how to use and the instructions weren't of much help.

 

[Helen B]

Helen: The ColorMunki takes 135 different nM band samples, so depending on how much is not considered at the start/end of the samples it looks to be on par with the i1.
...
Note that I've been watching the USB data directly from the ColorMunki, so I'm not sure if the above information (about nM bands and sample size) is publically written down anywhere from X-Rite.

Here is what X-Rite say in their feature comparison table:

" i1Photo and ColorMunki Photo feature very high resolution due to 3.5 nm physical sampling interval and 10 nm optical resolution, capturing more detail on paper and provides precise monitor and other emissive measurements such as ambient and digital projectors."

That should work out at 100 bands of 3.5 nm, between 380 nm and 730 nm. I wonder where the other 35 are coming from.

Best,
Helen

 

[Edward Crim]

Here is an e-mail I got Wednesday:

Edward,

My name is Stephen Rankin and I am the product manager for ColorMunki at X-Rite. I read your recent review of ColorMunki Photo and was surprised to hear about your poor experience with the display profiling functionality. I’d be happy to speak with you when you have time and perhaps build a display profile with you over the phone if you’re interested.

I was glad to see the printer profiling provided results that agreed with your expectations. This kind of positive experience is what most ColorMunki user are reporting with both the display profiling and the printing profiling.

Anyway, I just thought I’d drop you this email and introduce myself. If you’d like to talk about the product, please feel free to contact me.

Best,

Steve.

My Response:

Dear Stephen,

Thanks for the note! I would love to have you talk me through the issues I had with ColorMunki! After all, I spent almost $500 on the darn thing! I looked all over the X-rite site for someone to talk to about the issues, but could not find any specifics about whom to contact or any e-mail addresses (You might consider changing that to make your company more approachable).

Sincerely,

Edward Crim

The reviews I am finding on-line aren't as favorable as Stephen seems to think, though: Check out this one (though I'm fairly certain I know what he did wrong).
http://www.thinbits.com/2008/05/colormunki-pretty-much-doesnt.html

 

[Edward Crim]

I also tried the Color Munki with my Epson 2200 using the OS X printer driver and wasn't impressed. It seemed identical to what I got with Ilford's profile.

 

[pcummins]

That should work out at 100 bands of 3.5 nm, between 380 nm and 730 nm. I wonder where the other 35 are coming from.

Ah, good find. I'd imagine a bit of data is discarded (or processed differently) if it reaches into the IR or UV ranges that we normally wouldn't see with our eyes (ie, 17 and 18 on either end discarded). The hardware certainly seems very good, what I'm seeing is a "black" calibration sample, then the actual spot sample sent afterwards so it can minimise any noise on the sensor, I haven't had time to match the raw data with the captured post-processed data, this would tell me which parts are discarded (if any). At some point I'd like to get a Pantone matching system to confirm what the ColorMunki is reading is correct (we hope!) or chart for photography/scanning to compare.

In any case I recommend the ColorMunki, I am surprised people like Edward are having issues with CRT/display calibration as I find that part was pretty straight forward; getting printer drivers to output sans calibration/profiles was pretty non-obvious, and many people who have had issues I believe are not quite getting the settings right or consistent across prints. (For example, my Canon driver on Mac OS X likes to switch colour calibration back on by itself).

Admittedly, a full run through of the driver settings and warnings/cautions for various popular drivers (ie, Canon, Epson, HP, Brother) from X-Rite to go along with the ColorMunki will go a long way towards mitigating a lot of the dissatisfaction with the product when generating profiles. Obviously, if you need pro level calibration you'll need to spend big bucks on the software to do it, but for me, the ColorMunki seems adequate (then again, I like tinkering behind the scenes).

Edward: Re: Epson 2200 + Ilford profile, what paper/ink were you using + settings?

 

[Helen B]

If it is using the same measurement unit as the more expensive instruments then it would make sense that data is being sent from the near UV, even though it is a UV-cut instrument (as far as I know). It does appear to have greater potential than the PrintFix Pro, which is a much simpler, less versatile instrument.

Making good printer profiles that are better than many of those available online may not be as easy a task as some people imagine. No matter how sophisticated or simple the instrument and software are, the user will always have to do some testing to determine the optimum printer settings for their purpose (finding the right ink settings for the paper - usually called paper settings - to balance density range against separation, for example). With less precise printers (ie those that are not perfectly consistent) the imprecision of the printer will set a limit to the degree of perfection that is possible in the profile. I've had five Epson 2200s and used them with a variety of ink sets and RIPs, while using a variety of instruments for linearization and profile creation. There's a limit to what you can expect from a 2200 in comparison to the pro printers.

There will also be papers that would only be optimized with ink settings that are outside the simple set of settings offered by basic printer software.

In short, I don't think that there is any such thing as a really easy way of making good printer profiles, and anyone who wishes to make their own profiles should be willing to spend some time learning how to make good profiles, and then spend time making good profiles. You can't expect to buy an instrument and software, then start churning out good profiles. (I'm not directing that at any of the contributors to this thread, but just making a general comment about the usefulness of things like the ColorMunki to 'beginners', and the promises of marketing people)

Best,
Helen