|Home||About us||Commercial Photography||Print Gallery||Articles/reviews/blog|
Review of Printer Profiling with the X-Rite ColorMunki Photo
The ColorMunki Photo for printer profiling
The X-Rite ColorMunki is fairly new, and while we have a review of all the features, we decided to wait for a few software bugs to be ironed out before doing a more detailed review of some of its functionality.
This review primarily looks at creating ICC profiles for printer profiling, since this is the area that will probably attract most attention from people looking at colour management for the first time.
ColorMunki Printer profiling
Our original ColorMunki review contains information about monitor and projector profiling and the spot colour measurement and palette management capabilities.
Keith's review is based on using an Apple Mac, but most aspects are very similar on a Windows machine.
If you take pictures and get them printed (by yourself or by a lab) the next stage after after sorting out your monitor setup, you'll probably want to address printer colour management.
This effectively means using ICC colour profiles for your printing.
The ColorMunki Photo (Mac and PC) is firmly aimed at the large majority of users who might find some of the kit I've reviewed elsewhere a little daunting for their own requirements.
I often get asked for suggestions about learning more about the nuts and bolts of Colour Management.
My usual suggestion is Bruce Fraser's Real World Colour Management. My own copy is well thumbed. It's my first port of call if I'm asked a question and I feel I don't quite understand an issue well enough to be absolutely sure of an answer.
Check latest price/availability from Amazon.com or Amazon.co.ukRWCM 1st Edition RWCM
RWCM 2nd Edition RWCM
See some other books Keith has on the shelf, on our Books Page
There is more information available from X-Rite.
If you are new to colour management I'd seriously suggest you have a read of our less technical initial review of the ColorMunki before delving too far here, although I've tried to keep things pretty simple...
The ColorMunki kit includes a the all new spectrophotometer from X-Rite.
This USB device is based on the spectrophotometer design found inside the i1 iSis, a much more expensive device I often use for building my own profiles.
The measuring device comes with a soft case and software.
Apart from looking like a large tape measure, it's a solidly built device, which doesn't give the feeling it will quickly fall apart.
That said, treat it as a precision measuring device, and keep it in the soft bag supplied with it.
You measure test charts by sliding the device along the paper.
As such I found it easier to put the printed test chart upside down and scan from right to left with my right hand.
A smooth motion helps get more reliable readings, and it is important to start and end a strip on plain paper.
I'll show the whole profile creation process below, with a selection of screen shots and other images that I hope will give you a clear feel for how the device is used.
Whilst I'll make some comments about functionality of the software on the way.
I'll leave more detailed comments on what I think of the device and just as importantly how this relates to the ColorMunki target market, until the conclusions section at the end.
The principle of making a custom printer profile is pretty easy.
First you print out a test chart with lots of colours on it. Then you measure these colours and use the difference between measured colours and what was sent to the printer to produce a set of corrections. These 'corrections' are essentially what goes into an icc printer profile.
Since I'm just covering printer profiling here, I select the 'Profile my Printer' option to get going.
I'm creating a new profile, and I can choose which printer I'm using.
In this case a Stylus Pro 4880, which is attached to the computer I'm running the software on.
ColorMunki specifically doesn't support remote profiling. There are ways round this by printing to PDFs and the like but regard them as a kludge at best.
Some help is usually available - this is what you get for this initial stage.
A test chart is created and you can print it. The already printed tick box is for if you printed the target earlier, and just want to measure it.
This is one of the key stages in getting a profile. If your test target isn't optimally printed, then your profile won't be the best.
Here's the help for this stage...
Unfortunately that is all the help you get in the application.
If you go to help, an X-Rite web page opens (not connected to the internet? it won't open...)
This is what you get if you look for more info on the stage we're at:
Not particularly illuminating.
This warning appears before you print.
Now all this is very important but it misses any discussion of what are the best media settings for a particular paper.
There are no examples anywhere showing typical printer driver settings (for turning off colour management). I'd have hoped at least for some screen shots of typical driver for Epson/HP/Canon printers for Windows and Mac.
Looking at the Mac 4880 driver, I see a large collection of media settings. Each one of these will change the way ink is laid down on the paper.
If I'm using an Epson paper the the choice should be easy, but what do I pick for a third party paper?
I'm trying a heavy Innova 'Fibaprint' paper and from looking at their web site I see that Velvet fine Art is the suggested setting for Epson x880 series printers.
In previous testing I happen to know that Textured Fine Art paper is actually a better setting for this paper and printer. Hence I'm going to profile with VFA, since if you were just following instruction this is what you'd use.
I have written an article about media settings choices and some of the things you do to help choose.
I make sure to turn off Printer Colour Management (fortunately I knew how to do this).
Just to make sure, I check all printer settings before printing - make a note of these!
Here's my testing setup, with the Epson 4880 connected to my laptop.
After printing, it's important to let prints dry.
ColorMunki suggests ten minutes. I like to leave some stuff several hours or even overnight.
The glossy film I tested takes a couple of days to dry fully with pigment inks. Not a problem for the first sheet, but after measuring it and printing the second, I've got to leave the application running until I'm ready to measure the second.
Call me cynical, but there are aspects of this software where they have gone out of their way to stop you doing any form of remote profiling with this application. That's fine, but say I have three papers that all need long drying times for a particular printer, or want to profile such a paper in several printers...
As ever help is always at hand...
Once done it's time to measure the print, after calibrating the spectrophotometer.
A key feature of the measuring device is the rotating central section.
After plugging it in, I'm told that it isn't set correctly.
I've not had any problems with the rotation but I have heard of some people with limited mobility finding it difficult to rotate.
This is the way I hold it, with my thumb on the roughened surface. This way I also avoid accidentally pressing the measurement button (the dished bit with three raised dots on it).
You move it to the calibration setting.
Click on calibrate or press the button.
Note too, the option for an explanatory video.
Then move to the measurement position.
You are now set to measure the chart.
This aspect of the procedure is well documented.
The device really does slide along easily.
As you measure strips, the display is updated to show your progress.
With few exceptions, the strips were all read first time (see the conclusions for more on this though).
A second test chart is now created, based on the colours just measured.
This is printed, as before.
Notice the polite "if you are unable to recall the setting for the 1st test chart, you should start over and be sure to record these settings" :-)
This is important, so all the warnings are quite appropriate.
The second chart now needs to dry.
During this interval I got in some piano practice and felt hungry...
Then measured the second chart.
Here are both charts after I've measured the second one.
You pick a name for your profile, and after a minute or two, your profile is ready.
That's it, no settings, no preferences, no adjustments - it's done.
You can then configure various applications to use this profile by default.
Well that's what it says. I've not seen any significant change to my Photoshop CS3 settings whether the tick box was enabled or not.
And in detail...
This is the setting from the 'ColorMunki Tray'.
It allows me to pick any ColorMunki profile.
FYI: I run a LinkedIn group (~7500 members) for people interested in any aspects of modern Black and White photography: Digital Black and White
A further feature is that you can get the software to measure all the colours in an arbitrary image (skin tones for example) and create a further set of test patches for you to print and read.
After printing and reading these additional patches, the software creates an updated profile aimed at improving rendition of the colours in the image you used.
Of course you don't need to know how it all works... in fact, the ColorMunki is very much aimed at people who don't want to know what goes on under the hood ;-)You can refine any profile that you've already made with the ColorMunki. The profiles contain the previous measurements, so you are adding more measurements.
Usually more measurements will make a better profile, but they need to be accurate measurements made from targets printed at exactly the same print settings (you did write them down?).
You select an image to base the optimisation on.
In this case I'm using my standard black and white test image, to try and improve the greyscale performance.
The image is loaded and analysed.
Unfortunately there are no guidelines as to what is a good image to use for a particular purpose. I'm assuming the software just analyses the amount of pixels at different colours in the image?
I could probably have used an image of a grey ramp here just as well.
A new chart is prepared for you to print.
The picture below gives a better idea of the range of patches created.
The chart is measured as before.
Afterwards a new profile is created.
Remember that profiling is a classic example of 'Garbage In, Garbage Out' so if your refinement sheet isn't quite perfect (a blocked nozzle for example) then it will throw out the whole profile.
Unfortunately if you make a mistake and the software doesn't notice it, there is no facility to go back and re-scan a row. Whilst I wasn't able to duplicate this condition, I have read a report of someone doing it.
I tried quite a lot of papers and the results were good. The picture below shows some test patches drying. You can see some of the gloss differential that you get with pigment inks too.
Two sheets of glossy film and part of a 17 inch roll of canvas.
The area printed on the canvas indicates a noticeable failing in the software (although I suspect intentional).
There is no provision for changing the paper size in the software. Almost every other Mac application I use has the standard 'page setup' options. This doesn't.
You can't alter the test chart in any way. If you wanted to profile a small dye sub printer or even a printable CD, then it's no-go.
The world seems to use US Letter paper according to ColorMunki - er sorry chaps we don't!
Look at the A4 sheet printed on a 4880 below. Not one sheet has ever printed that whole line of text.
Just as well there's no way of putting custom text on the target...
With the canvas I just printed the 'normal' target, and then trimmed off an A4 sheet for printing the second. You can probably see why I didn't bother testing too many 44" width rolls of paper on my 9600 ;-)
Both of the (A4) sheets below were printed on the 4880.
The glossy film took several hours to dry properly. Whilst you can print several first sheets, like above, you can't do the same with second sheets. This really messes things up if you wanted to get all your profiling done in one go. I've heard of printing to PDFs and other 'fixes' for this but you are fighting against the software and far more likely to get errors.
I've got lots of different test images I use [page with links to most of them]. I've loaded one from Scott Martin (see below) that's quite nice for checking out a variety of features. It also shows some of the areas that can catch the unwary when evaluating test prints.
The bright red bits are actually where I've switched on the 'out of gamut' warning in Photoshop (I've set the warning to red to show a bit easier).
I'm soft proofing the image with a profile I've just made. This gives an indication of how a print will look.
The greyscale ramp shows any banding or false colours quite nicely. This was fixed with my grey optimisation stage to some extent, but nothing about the ColorMunki package gave me the impression that people doing serious black and white printing were part of its target audience. To get decent black and white prints, your printer needs to have several different black or grey inks. If you are using a printer with just CMYK inks then don't expect good black and white results. If by chance you do get a reasonable B/W print with basic coloured inks, then expect it to change as soon as you change the room lighting.
Note that I've already written an article about using the ColorMunki spectro for improving black and white printing - this uses the ColorPicker application to read patches. The QTR ICC profiles created are nothing to do with the ColorMunki software I'm looking at here.
As you can see, strong saturated colours are beyond the range (gamut) the printer can reach - or at least with this profile, using the particular media settings I've used.
People often worry about out of gamut colours when looking at profiles, but it's important to realise that unless you are printing pie charts, these colours rarely occur in real life images.
Look for prominent banding or flat colours in these areas as an indication of profile condition, but remember this is just part of what you're looking at.
Of course, dealing with out of gamut colours is one of the reasons you'd try printing with different rendering intents, but that's information you'll just have to find elsewhere. There is no supplementary info supplied with the ColorMunki to help ease users into this subject.
Once again I tried the AppSet application - it did nothing on my machine, so I can't comment on just what it fiddled with in my Photoshop settings.
The AppSet functionality is enabled by a plug-in, called "ProfileSetterCS3.plugin" for my copy of Photoshop CS3. Every other plugin installed gives its version info in a small window from the Photoshop>About Plug-in menu. This one doesn't. This is basic plug-in coding practice - it didn't inspire confidence.
Whilst I can see the 'keep it simple' approach to managing settings at work here, I do wonder just how much trouble will occur in multi user environments when one application starts messing with the settings of another in an arbitrary way.
Of course the AppSetter seems to assume that I've only two profiles on my system.
The print dialogue shows otherwise...
These are mostly installed by the Epson printer drivers and with the quality of OEM profiles these days (for the large new printers), I'd almost certainly choose many of them over a ColorMunki profile, particularly if you've used something like Epson ColorBase to calibrate your printer.
Not knowing what's being changed (or not) I always had a good look at the full printer driver settings before pressing the 'Print' button.
That test image has some good tests of shadow detail (the bits of electronics) which showed up some slight issues in some profiles.
The skin tone for the model is one quick test that I've found more people look at for a quick evaluation. If she looks a bit green, then it really doesn't matter how good other parts of the profile are, particularly if you are printing pictures of people.
Do remember that if you use an image for refining a profile, it's best to use one that you know is correct. If I was refining skin tones, then I'd probably chop out images of people from test images and use these small parts for the refinement. The software doesn't know that you want to optimise for skin tones just because there is a small face on the picture somewhere.
But how good are the profiles!
All in all I tested 5 different papers on the Epson 4880, two on my Epson 7880 and one on the Epson 9600. I've also tested two papers in my HP K80 (where the CM created passable CMYK profiles due to the postscript driver I'm using) and a Canon glossy photo paper in my old Epson 1160 (the one used in my media settings article).
Papers ranged from a glossy plastic film, though OEM papers to heavy matt art papers and even plain copier paper and a glossy canvas.
But... if I understand the market the ColorMunki is primarily aimed at, users probably require more and clearer examples of how to print from common applications using the profiles - there just isn't the supporting information aimed at helping people make use of profiles.
Applications like Aperture and Lightroom are actually mentioned (in the on-line help) but the brief notes on using your ColorMunki profiles with them really do need expanding to include much more detail. There is more info on the ColorMunki site, but you do have to find it first.
For the market the ColorMunki Photo is aimed at, I have to say that I was generally pleased with the results.
The profiles are Version 4 ICC profiles. While I didn't have any problems with this during my testing, I have heard of a few people mention some old versions of applications having difficulties with this newer format - nothing I was able to test though.
However, without exception, all the RGB profiles were just a tad too dark in the shadow areas (for my tastes). If you print with perceptual rendering from Photoshop then many people probably wouldn't even notice.
But why would I choose this rendering intent? The AppSetter did nothing to my (Mac OSX 10.4.11) system, and nowhere in the ColorMunki help files have I seen any information that even suggests the user should use these settings. If you were to experiment and try say, Relative Colorimetric, the shadow detail was often (but not always) compromised.
When I created the profiles for the Innova paper, I also printed a 1728 patch A3 sized target and created a profile using an i1 iSis and ProfileMaker Pro (PMP5) for the TFA media setting. That's over £5000 worth of kit, so I would expect it to do well ;-)
In a blind test (carried out on unsuspecting visitors to Northlight :-) with half a dozen prints of the the test image above, everyone picked the PMP5 print as best (not by any great margin, but unanimous) Those that expressed an opinion all pointed to the skin tones. How much this reflects a more pleasing profile as opposed to a more accurate one is another matter.
Greyscale printing with all the basic CM profiles showed slight steps in the bull's-eye test on my B/W test print but not obviously noticeable in real world images ... unless of course a smooth gradient - say the sky, just happens to cross one of the bumps.
Unfortunately there were often slight green or magenta bands visible in a greyscale ramp. These were fixed to some extent when I used a greyscale image to refine the profile. If you are serious about black and white printing, then do note that the ColorMunki software does not currently pay black and white printing any particular attention.
Some printers (I don't have one) are probably just not going to work well, but I also suspect they are the sorts of printers that are difficult to profile with any package - If you have one of these, then it's a matter of "if it works,it works - if it doesn't, it doesn't" ... time for a new printer? or at least try a different paper ;-)
Anyone who's done much profiling will appreciate some of the nifty software that has gone into making quite respectable profiles from 100 patches. This does to some extent rely on newer printers being a lot more linear and predictable than in the past. Whilst you can always add more test charts from images, this may well give diminishing returns after a while.
My one try at this, suggested that a greyscale and a good varied colour photo (200 patches in all) improved the profile quite a bit over the original. Since the software can work out where some of the weaknesses in the profile are, it should easily be possible to have it automatically generate extra pages for B/W or colour optimisation.
I suspect that some of the issues I saw were due to less than optimal media selections in the drivers.
The only media selections you can really trust are for OEM papers used with OEM inks. Once you try a third party paper, you really do need to experiment. Don't assume either that the paper manufacturers will always be of help (most times they are). It's just not feasible to expect paper manufacturers (or I should say, the people who's names appear on the boxes ;-) to have one of every printer commonly available. All too often, the creation of profiles you can get from manufacturers will have been contracted out, and you can't be sure that the people who made them were prepared to test lots of settings with the rigour one might desire.
For one of the excellent Innova 'FibaPrint' papers I tested (IFA09), the suggested media setting was 'Velvet Fine Art' (VFA). I tried many of the settings available on the 4880 and found that 'Textured Fine Art' (TFA) was better on the 4880 (note that this might not be true with other printer models).
While the ColorMunki Profile made with the TFA setting was a bit better, my suspicion is that not many would notice - even so, the better your initial media choice, the better your profile should be (TFA and VFA are similar papers).
I have seen photo papers where the best setting on a particular printer was for 'plain paper'. Using VFA produced denser more blocked up profiles for the Innova paper, but that's what I tried, since if I'm following the fairly minimal instructions with the ColorMunki all I do is:
I can see why this simplistic (but accurate) advice is given, but there really was no great effort required to add genuinely helpful 'help' info for those with a bit more knowledge who might be looking for more clues and guidance. I know that it's assumed there is a demand in the market for this 'simplified' approach, but I suspect it would grate with for those eager to progress their knowledge. I know that there are further resources on the ColorMunki web site, but I just don't compare visiting a web site to useful hints and tips that I can just call up on my own machine (and no, not every machine I work on is always connected to the internet).
The help pages on the ColorMunki web site feel rather 'light weight' and not checked very carefully, given the number of extraneous characters:
The software installs an awful lot on your system without the good grace of leaving you a "what's installed where" log file. I've only tested the Apple Mac version, but it could do with some general interface improvements, such as having preferences available from both ColorPicker -and- ColorMunki Photo. ColorPicker has a Page Setup menu option... how about one in ColorMunki Photo for those of us without piles of 'US Letter' paper? I shouldn't need to cut sheets of A3 paper in half to print two targets either. Some of these things can be changed in the printer driver, but every other bit of software I use to print has a Page Setup menu item...
When installing the software, it says that you are limited to three 'activations' of the software. Since there is no 'uninstaller' this might suggest problems if you change machines. Fortunately the software seemed to 'activate' on any of my machines with no problem.
2009 UPDATE - the 3 Activations limit is dropped
Given you need a ColorMunki spectrophotometer to use the software, this system seems a little convoluted, unless perhaps it signifies add on (paid for) functionality that might be available in future incarnations of the equipment. At the moment though, X-rite are sticking to the line that the ColorMunki is not an upgradeable system...
The software driving the measuring device needs to see sufficient variation between patches to identify them. If your printer/driver/paper/settings produces fairly good prints, then there isn't a problem. I had slight problems with the glossy film where a few rows took several attempts to measure.
The point is, I know why this is happening and I can look at printed patches and guess where there are problems. I've already seen reports from people using other printers with particular papers where they have wasted ink and paper without getting successful readings. Whilst I might just give up and decide that a particular paper just isn't going to work, there are no hints or advice available. There is nothing wrong with saying that some printers are rubbish or that some papers/inks/papers just won't work. It's not admitting a failure in a product to suggest that profiling can only do so much.
In spite of this, I'd recommend the ColorMunki for people to try - it can, with care, produce rather good profiles and I hope that comments like mine (and similar ones I've heard elsewhere) will persuade X-Rite to improve the software, and in particular the support environment for those who want to do a bit better.
Comments/Questions - feel free to ask below
A colour management solution aimed at a much larger market than other X-Rite products.
Buying the ColorMunki
We make a specific point of not selling hardware, but if you found the review of help please consider buying the ColorMunki, or any other items at all, via our links with Amazon or B&H
It won't cost any more (nor less we're afraid) but will contribute towards the running costs of our site.
Excellent measuring device, and it can produce very acceptable profiles, but is let down to some extent by an interface which doesn't address any of the issues that make a difference between an OK profile and a good profile.
To get the best results you will still need to learn a bit more about what it is you are doing, unfortunately there is little help in this respect in the accompanying documentation.
X-Rite ColorMunki Photo - Minimum System Requirements
X-Rite ColorMunki Photo Package Contents
Note that ColorMunki solutions are not designed to be upgraded. They are available only in single site or limited multiple seat licenses – i.e., those with a maximum of three computers.
The X-Rite range of products (see below for our reviews of just about all of them...)
Disclosure of Interest
Northlight Images has been involved in testing the ColorMunki, but has no commercial connection with X-Rite. Note that we do not sell hardware or software at all, and do not offer a custom profiling service.
Thank you for supporting our site.
The views in this article represent those of Keith Cooper.
Keith is always happy to discuss matters raised in his articles. You can Email Us
Northlight Images prides itself on its independence when giving advice. We do not sell hardware or software and have no direct commercial links with any of the software or hardware vendors that may be mentioned here. See our Review Policy for more information.
You can search all the many hundreds of articles and reviews on the site for more information
Have you found an article on the site useful or helpful?
If so, please consider sharing a link to the article or mentioning it on a forum or blog - Thanks to everyone who's helped the site become better known.
Explore our site... Digital Black and White photography and printing - some of Keiths thoughts, techniques and tips for those interested in a digital approach to black and white. There are many hundreds of entirely free articles and reviews on the site. New site content appears on the News, articles and reviews page.Thanks to the visitors who've made Amazon purchases (any kinds of items whatosever)
via: Amazon UK/Amazon France/Amazon Germany/Amazon USA/Amazon Canada
It won't save extra money we're afraid, but it does help in the running of the site, and we really appreciate it...
Commercial Photography - UK Architectural Photography services - Industrial - Print Gallery - Landscape Photographer