ColorMunki Photo / Printer Profiles

[Scott Martin]

To tweak the gray axis for perfect neutrality.

[Kevin Sholder]

Thanks, that works.  In running my test tonight at the bottom of the step when it asks you to load an image it states:  HINT - To maximize profile accuracy, you may repeat the Optimization process using different images.

So it sounds like:

A.  It does provide a better sampling since it can create the chart from more images with varying colors / tonalities.  So it would be interesting to see how those profiles update internally from one to the next.

B.  X-Rite wants you to use the ColorMunki a lot and it tops out quickly, but without being able to see inside the profile one would not know what is going on.  So in a rhetorical fashion, how many do you do...  :-)

Kevin

[Scott Martin]

So in a rhetorical fashion, how many do you do...  :-)

Not very many. One at most, if you're picky.

[Kevin Sholder]

If I'm asking too much, say so, but for experiment sake, could someone look at the two attached profiles and see if there how much of a difference there is between them?  It would be of interest to know how much of a change is taking place when optimizing a profile using the ColorMunki.  Or tell me what I can use (without spending a ton of money) to view these myself?  I work in a Microsoft Windows environment.

The v1 is just that and the v2 is optimized using the black & white image from:  www.on-sight.com/download/Onsight_BW_Evaluation.zip, but the file had to be converted to a TIF as the ColorMunki software only uses JPG or TIF files for input.

If it would be of interest the paper used for this profile is Moab Somerset Museum Rag on an Epson R3000.

Thank you for taking the time to run the experiment and let other users as well as myself know how much of a change we would expect.

Kevin

[Scott Martin]

Before we start, your internal names of these profiles are the same so you won't be able to tell them apart from profile menu's in Photoshop and the like. I used the ColorSync utility to rename them with the "v1" and "v2" that you've entered in the file name.

Well without writing a short book, here are two quick ways that anyone with Photoshop can see if there are differences between these profiles:

1) Print with each profile (and potentially each intent) and compare the print results under high quality lighting. This is always the most important results that should be valued above others.

2) Open an evaluation image Photoshop and convert it to one of these profiles. Then Assign the other profile to it. If you see a visual change when assigning the second profile, these changes represent the differences between the two profiles. It will look worse when assigning the second profile but don't get caught up in this or conclude that one is better than the other here. This is not the test to determine which profile is superior - only if they are different and by how much.

There are other more geeky ways of comparing profiles with a variety of color geek applications, but the above two are simple that anyone can do. With this, you can see that your two profiles are actually quite different! I'm guessing that if you perform test 1 with Perceptual and look at the prints in daylight that the v2 profile is better and worth the extra effort.

[Kevin Sholder]

Scott,

Yes, they are the same as I simply over wrote the previous one, but always save backup copies of my profiles to go back to if needed.  I think I'll change my methodology at this point for naming so I can get multiple versions around.

Based on #1, I should see the same results then when printing from Lightroom correct?  Lightroom is my primary printing source in my workflow.

Glad to see that there was a big difference, that actually makes me feel good about the extra time and effort, as I'm looking to get the most out of the equipment that I have currently.  Then when I'm ready for a larger printer, I should have a good handle on it.

So what are some of the more geeky ways to look at these?  Being an IT Pro, geek is of interest to me as it will help me to get the next level for my presentation.

Thanks,
Kevin

[Scott Martin]

Based on #1, I should see the same results then when printing from Lightroom correct? 

You should get different results from these two profiles. You haven't printed with these yet?

[Kevin Sholder]

I've printed with v1 last weekend and the prints look very nice.  Created v2 last night following our discussion and will print tonight after work using the same images so I should see the difference.

Kevin

[digitaldog]

If I'm asking too much, say so, but for experiment sake, could someone look at the two attached profiles and see if there how much of a difference there is between them?

Here you go. Using a 988 patch (iStar) target in Adobe RGB (1998) for these results:
--------------------------------------------------

dE Report

Number of Samples: 988

Delta-E Formula dE2000

Overall - (988 colors)
--------------------------------------------------
Average dE:   0.31
    Max dE:   3.03
    Min dE:   0.00
 StdDev dE:   0.35

Best 90% - (888 colors)
--------------------------------------------------
Average dE:   0.22
    Max dE:   0.61
    Min dE:   0.00
 StdDev dE:   0.13

Worst 10% - (100 colors)
--------------------------------------------------
Average dE:   1.14
    Max dE:   3.03
    Min dE:   0.61
 StdDev dE:   0.55

--------------------------------------------------
--------------------------------------------------

Worst dE patches sorted in order, you can see you improved mostly grays:

Worst (RGB 127) shows a difference (note, not necessarily improvement) in aStar and bStar.

[digitaldog]

Well without writing a short book, here are two quick ways that anyone with Photoshop can see if there are differences between these profiles:

Here is a 3rd that can use Photoshop and shows where in the image the differences appear in a different way:
http://digitaldog.net/files/Apply_Image.pdf

[Kevin Sholder]

Andrew,

THANK YOU so much for sharing this!!  The improvement in the gray areas makes sense since that was the resource file type I used.  Is there a resource I can obtain to learn more about how to understand what is happening here then?  And have the basics of color management changed in the last 8+ years or is the foundation pretty solid?

Kevin

[digitaldog]

Is it an improvement? It is different. Some patches are shown twice, keep that in mind as this target is used for detecting positional color differences on a sheet thus they fall in different places when measured. I see a small range of colors in source RGB areas that are highly affected in both a and bStar and wonder why other colors in the range remain unaffected! RGB 127-133 and 37-45 are high in color difference. How do they compare with the 'before' print in those areas? Part of that is of course subjective and subject to viewing conditions.

[digitaldog]

And have the basics of color management changed in the last 8+ years or is the foundation pretty solid?

No and no.

[Kevin Sholder]

Got it, once a make another set of prints tonight or tomorrow I'll should be able to see the difference and follow up on it then.

Thanks again!!

[Kevin Sholder]

I would like to thank both "The Digital Dog" and Scott Martin for their guidance on this forum in helping me improve my printing abilities.

Well, after creating multiple profiles using the optimization feature within the software provided with the ColorMunki, I've come to the following conclusion.

Optimization is beneficial to a point.  That point for me was after two phases of optimization.  I actually made five optimization updates from my base profile, so there were a total of six versions of my printer / paper profile that I printed and reviewed.

The first optimization was built using the image found at:  http://outbackprint.outbackphoto.com/printinginsights/pi049/essay.html.  I liked this image the best because of the bottom patch ramp which was most helpful in determining how much detail there was in both highlight and showdown areas.

The second round (v3 of the profile) of optimization built, used the image found at:  http://www.on-sight.com/download/ being the Onsight B&W Evaluation Image.  With this image however, you will need to convert to either TIF or JPG as the ColorMunki software does not read PSD files.  I converted mine to TIF.

Further optimizations were performed using the Granger (v4) and Gamut Charts (v5) found at:  http://www.luminous-landscape.com/essays/test-charts.shtml

Then for v6 of the profile I added an additional B/W image found at:  http://www.northlight-images.co.uk/article_pages/black_and_white_test.html the direct link to the image is:  http://www.northlight-images.co.uk/downloadable_2/bwtest_2.zip

Changes from v4 and v6 showed minor improvements, but I did find that the blacks on the bottom patch ramp started to block up sooner by v6 of my testing.

So in my opinion and with the testing I was able to do over the past week, the next time I create a printer / paper profile using the ColorMunki I will do the following:

1.  Run the software to create its first color patches, print, dry and measure to obtain the second color patch set.

2.  Print the second patch set, let it dry, measure it and build the profile.

3.  Then I’ll run the optimize routine using the image found at:  http://www.jirvana.com/printer_tests/PrinterEvaluationImage_V002.zip.  Print the resulting patch set, let it dry, measure it and build v2 of that profile.

4.  Then I’ll run the optimize routine a second time this time using image found at:  http://www.on-sight.com/download/Onsight_BW_Evaluation.zip. Print the resulting patch set, let it dry, measure it and build v3 of that profile.

5.  Ready to make some beautiful images!!

Please note that I let each patch set dry for 8 to 12 hours before I reading it using the ColorMunki, not the 10 minutes that x-rite recommends.  This is what I have found will work for my equipment and am extremely happy with the profiles and the resulting images from my Epson R3000 printer.

Thanks again!!  And happy PROFILING.

[Scott Martin]

Well, after creating multiple profiles using the optimization feature within the software provided with the ColorMunki, I've come to the following conclusion. Optimization is beneficial to a point. 

Perfect - I agree! Tip of the hat to you for doing these tests and coming to this conclusion - good work.

[gbillett]

Interesting and helpful discussions.  Have just upgraded from 3800 to ipf6400 ( awaiting delivery ) and am now considering making my own profiles of the paper stock I use, mostly Ilford Galerie Pearl,  Gold Fibre Silk and Canson Photographique.  The manufacturer's profiles have always seemed pretty good - is there any noticeable advantage in spending £300 on ColorMunki?  Prints are for exhibition purposes.

Thanks
Geoff

[Kevin Sholder]

gbillett,

I did not start with a manufacturers profile, so I do not know.  I may try that in the future.  I bought the ColorMunki as my color management system, so it works for me.  If you are happy with your current profile I would continue to use it, I was just trying to see what I could get out of my system.

Kevin

[gbillett]

Hi Kevin
Some papers ( eg fine art cotton papers ) are reputedly difficult to profile, depending on variables such as climate etc,  and manufacturers profiles are conservatively produced.  I live 20 miles away from St Cuthberts Mill,  makers of Somerset Enhanced paper,  and would like to explore their and similar products eventually.  Just wondering aloud though with the baryta type just how much of an advantage profiling papers would be.  Only way to find out is to progress along the ( expensive ) learning curve of colour management,  purchase ColorMunki and see :-).   
Cheers
Geoff

[Alan Goldhammer]

  Have just upgraded from 3800 to ipf6400 ( awaiting delivery ) and am now considering making my own profiles of the paper stock I use, mostly Ilford Galerie Pearl,  Gold Fibre Silk and Canson Photographique.  The manufacturer's profiles have always seemed pretty good - is there any noticeable advantage in spending £300 on ColorMunki?  Prints are for exhibition purposes.

Some manufacturer's profiles are quite good and it may be difficult to improve upon them.  Remember that the profiles are also based on the consistency of the manufactured printer so that each unit is within specification.  they are profiling for the 'average' printer and not yours.  The original Ilford Gold Fibre Sile profiles for the Epson 3880 were really quite dreadful and it was very easy to improve upon those doing your own profiling.  In the end it all amounts to a cost trade off - does one invest in the hardware and software to do profiling and invest the time to optimize results.  I started doing my own profiling four years ago first with the ColorMunki and then moving up to an i1Pro using ArgyllCMS software (which is free but has a significant learning curve).  My profiles are better than the manufacturers for the papers I print on but not by very much.  One option to consider is to have someone expert in the field do the profiling work if you have settled on a small number of papers to print on.

My two cents worth on the topic.

Alan