Testing a spectrophotometer


(Public Enemy Nº 2) #1

Hi! I’m considering buying a Colormunki Design spectrophotometer.
I found a used device at a reasonable price, but since it’s used I wonder whether there is a reliable way to diagnose and test it so I know it works properly before buying.

I have two colorimeters already (one quite old with aged filters and a colorhug that never worked quite well) and this would be my first spectro

Can anyone give me some pointers? Maybe @gwgill can give me some advice :slight_smile:

(Mica) #2

Check if your library has a Pantone book. Or see if someone in your local photo club has a color target.

(Public Enemy Nº 2) #3

I have pantone books and IT8 targets. I suspected that measuring known values and checking the readouts expecting a little DeltaE could be a resonable test, but I wanted to know if there was a more comprehensive way to diagnose the device and make sure everything is in place.

(Graeme W. Gill) #4

This rather depends on what you intend to use the instrument for, and what resources you have available to you.

Obvious things are to check that it calibrates and goes through the motions.

If you have a ColorChecker available, you could do a sanity check by measuring
some of the squares and comparing to the known references to make sure it is in the ballpark.

Another sanity check is measuring an old fashioned incandescent lamp (using ambient mode), and checking that the spectrum looks like it should, and that the color is close to the black body radius (x,y about 0.448 0.407), and is something like 2800K CCT.

Of course the best way would be to compare various readings to a known good spectrometer.


Good evening, @gwgill,

Would Argyll be capable of producing such a spectrum graph
or what software and steps are needed?

Claes in Lund, Sweden

(Public Enemy Nº 2) #6

Thank you very much @gwgill!
I intend to use the instrument for anything useful for my design practice, being spot readings, as an aid for device calibration and profiling (hopefully making my Colorhug1 finally usable or putting my iDisplay2 back to working state), producing decent print profiles, helping in photography, etc.
I’d say I’ll cover all the practical uses a Colormunki spectro has, but I’m fine as long as the level of accuracy I get is enough for what I do.
Colour is of course important in my work, but as my clients don’t have accuracy requirements it’s not a critical aspect, so getting a reliable colour workflow and reasonable consistency is enough. It’s not that I have any client who would reject a job if a deltaE is a bit off.

So my main concern is that the device is not b0rked and it works reasonably well, and I feel that the recommendations you just gave me will give me an indication of that. Thanks again.

One last question: I know every hardware ages, but as far as I know a spectro is quite different to a colorimeter (specially the cheap ones with plastic filters). Is there something in a spectro that could indicate aging or breakage so I pay special attention to it?

(Public Enemy Nº 2) #7

@Claes Check http://www.argyllcms.com/pro/ It looks fantastic, I’m considering to buy it :slight_smile:
Otherwise, the Argyll you already have in your machine has all the needed stuff, but it takes some more learning. I’m not by any means an expert so don’t quote me, but the documentation of commands ilumread and specplot seems like a good place to start.


I don’t know much about spectros. If accuracy is your concern, would it be possible to have it checked and / or calibrated by the manufacturer or a service that does that? However, depending on what you are talking about, this option might not be available or affordable.

(Public Enemy Nº 2) #9

That’s unfortunately out of the question. It’s not available here, and I’m buying a used one because our customs taxes (Argentina) are quite abusive, so the whole idea of sending the device abroad for diagnose and paying for it isn’t viable at all.
Buying a new one and even paying the import taxes would end up being more convenient.


It is like that here too. Good luck.

(Graeme W. Gill) #11

Sure - use spotread -S.

(Graeme W. Gill) #12

For spectrometers like the i1Pro that uses an incandescent lamp as a light source, then the age of the lamp can be an issue, as they do wear out, and there will come a point where it won’t calibrate for reflective work, or where the lamp could have burned out. Generally this isn’t an issue unless the instrument is quite old and/or has been heavily used. The X-Rite driver does keep count of the usage, although the ArgyllCMS driver doesn’t (I didn’t want to risk writing to the instrument EEProm). I think spotread -v does return this information.

This is not so applicable to the ColorMunki, since it uses a LED light source.

The main danger for most spectrometers is being subject to physical shock, since this can displace the optics, shifting the wavelength calibration. The i1Pro2 has a green wavelength calibration LED built in to calibrate and/or detect this type of degradation. The technical way of detecting this is by measuring a wavelength reference light source. The practical approach is to check it’s measurements on some more saturated color references (hence my suggestion about using a ColorChecker.)


(Andrew Keech) #13

Been using Argyll Pro a couple years for little jobs here and there (photometry for lighting, comparing white paints in sample books, etc), it’s fucking awesome.


Oh dear: you mean that I have not read the Manual properly? :frowning: Ah well, luckily there is an acronym to cover this deficiency: It was an RTBFM error. Sorry about that!

Have fun!
Claes in Lund, Sweden

(Public Enemy Nº 2) #15

I received the used Colormunki Design today and tested it quite thoroughly with pantone swatches, comparing the reads with the swatches lab values, and everything seems pretty darn close, so I’m assuming the device is right.
I’ll perform other tests, but so far it’s picking up spot readings as expected and it did a fine job calibrating and profiling several screens (finally colour in my two-monitor set matches!)

I tried DisplayCal on windows but it didn’t pick up the instrument, while DisplayCal on Linux did without any issues. I’m trying to figure out what’s missing. Any pointers?

I’m also planning to produce a calibration matrix for my Colorhug1 which is close to useless without correction (the infamous red shift). I sort of get the overall idea (i.e.: reading the same screen with the spectro and the colorimeter and producing a matrix transform to counter the shift), but I’m not sure about the process. I’d appreciate some guidance or resource to learn more about the topic.

Thanks in advance

(Mica) #16

I’d assume on Windows that the color munki needs a driver.



Have fun!
Colourfully Yours,
Claes in Lund, Sweden

(Public Enemy Nº 2) #18

Thanks @paperdigits and @Claes,
I have the official driver installed, of course. It’s “the windows way”. :smile:
The instrument is not detected by DisplayCal (it is detected by the included software).

Anyway, reading the DisplayCal documentation again, which I apparently didn’t read carefully enough yesterday, I found that some devices need specific Argyll Drivers. The first time I read that sentence I took it as there was no need to download specific drivers for the Colormunki, but it turns out that that was true for the Colormunki colorimeter, but probably not true for the Colormunki spectro. Same brand, different devices.
I suspect that those drivers are already included on linux, and that’s why the spectro worked with DisplayCal flawlessly there.

I’ll try again and keep you guys posted.


Er… I have a faint memory of something having to be done in/around Argyll… I’ll scout through my links & return to you in a few minutes (or so).

BTW: You’re on Win 10 64-bit Pro or what?





Good luck!