Luminous Landscape Forum

Raw & Post Processing, Printing => Colour Management => Topic started by: VitOne on January 11, 2011, 10:29:25 am

Title: ColorMunki VS i1Pro - Monitor profile validation
Post by: VitOne on January 11, 2011, 10:29:25 am
I am using the SpectraView Reference 271 (Europen verison) monitor and the SpectraView Profiler 4.1.24 software. My i1Pro is HP-branded rev D.

I could not find a "diagnostic software" for rthe ColorMunki. This is the i1Pro i1Diagnostic log:


General Information
Date and Time: 11/01/2011 11:37:15
Application version: Version 2.5.1
SDK version i1: Version 3.4.0 Build 131
SDK version iO: Version 1.1.2 Build 100
SDK version iSis: Version 1.0.6 Build 72
Platform: Microsoft Windows ???,

Device Information
Device type: i1Pro (NO filter) [ Ambient Light Emission Reflectance Reflectance Scan Flash ]
Serial number: 304522
Firmware version: 502
CPLD version: 999
Dark measurement count: 2248
Lamp burning time: 838.724 seconds
Spot measurement count: 174
Scan measurement count: 220

Dongle Test
Module "Scanner" is licensed
Module "Monitor" is licensed
Module "Printer" is not licensed
Module "Printer RGB only" is not licensed
Module "Printer CMYK only" is licensed
Module "Beamer" is licensed
Module "Printer Easy RGB" is licensed
Module "Printer Easy CMYK" is licensed
Module "Digicam" is not licensed
Module "Editor" is licensed

Button Test
Button is working normally

Calibration
Emission calibration: Successful

Emission Test
White measurement: Successful (X=106.112 Y=112.678 Z=120.152)
Gray measurement: Successful (X=35.516 Y=37.532 Z=41.326)
Black measurement: Successful (X=1.194 Y=1.226 Z=1.360)
Red measurement: Successful (X=60.519 Y=27.365 Z=1.367)
Green measurement: Successful (X=24.601 Y=76.363 Z=10.753)
Blue measurement: Successful (X=20.212 Y=8.389 Z=105.740)
Emission Test: Pass

Dark measurement
Dark measurement: Pass (-185.414)

Noise Measurement
Noise measurement: Pass (2.645)

Lamp Test
Reflectance calibration: Successful
Reflectance measurement: Successful
Check measurement result: Successful

Reflectance Test
Reflectance calibration: Successful
Reflectance spot test: Successful (LOT=839.364 seconds, CIT=0.01244 seconds, LIT=0.01336 seconds)
Scanning calibration: Successful
Reflectance scan test: Successful
Measurement 1: Successful (L=96.012 a=-0.496 b=2.430)
Measurement 2: Successful (L=95.987 a=-0.492 b=2.454)
Measurement 3: Successful (L=95.969 a=-0.493 b=2.459)
Measurement 4: Successful (L=95.950 a=-0.489 b=2.462)
Measurement 5: Successful (L=95.948 a=-0.490 b=2.470)
Measurement 6: Successful (L=95.942 a=-0.486 b=2.477)
Measurement 7: Successful (L=95.941 a=-0.485 b=2.472)
Measurement 8: Successful (L=95.941 a=-0.488 b=2.484)
Measurement 9: Successful (L=95.939 a=-0.485 b=2.487)
Measurement 10: Successful (L=95.928 a=-0.493 b=2.492)
Reflectance drift test: Pass (0.025)

*** PASS ***


I did this:

1) Calibration, caratterization and profile making with one hwadrware
2) Profile validation with the same hardware.
3) Profile validation with the other hardware.

Test settings:
D65; L*; 110 cd/mq; Black Minimum Neutral.

All values are DeltaE 2000.
Both devices where positioned in the center of the screen.
Before I made the meusurements I left the device on the screen for a reasonable time (more than 30 minutes) and the screen was turned on about 60 minutes before the beginning of the test.

First test

Profile made with ColorMunki.

CM validation without re-calibrating the instrument:
max 0.6 med 0.2
contrast 633:1
black 0.17 cd/mq

CM validation 1 after re-calibrating the instrument:
max 1.93 med 0.42
contrast 1025:1
black 0.11 cd/mq

CM validation 2 after re-calibrating the instrument:
max 1.28 med 0.40
contrast 654:1
black 0.17 cd/mq

CM validation 3 after re-calibrating the instrument:
max 1.13 med 0.29
contrast 746:1
black 0.13 cd/mq

i1Pro validation 1 for the CM-generated profile:
max 7.29 med 2.07
contrast 672:1
black 0.17 cd/mq

i1Pro validation 2 for the CM-generated profile:
max 8.03 med 2.08
contrast 596:1
black 0.19 cd/mq

i1Pro validation 2 for the CM-generated profile:
max 6.17 medi 1.76
contrast 646:1
black 0.17 cd/mq

Second test

i1Pro generated profile.

Validation without re-calibrating the instrument:
max 0.66 med 0.28
contrast 948:1
black 0.12 cd/mq

Validation 1 re-calibrating the instrument:
max 1.22 med 0.4
contrast 768:1
black 0.14 cd/mq

Validation 2 re-calibrating the instrument:
max 0.78 med 0.38
contrast 824:1
black 0.13 cd/mq

Validation 3 re-calibrating the instrument:
max 0.65 med 0.36
contrast 822:1
black 0.13 cd/mq

Validation with CM:
max 6.62 med 1.73
contrast 732:1
black 0.15 cd/mq

Validation 2 with CM:
max 5.87 med 1.64
contrast 580:1
black 0.18 cd/mq

Generl conclusion:
It seems that there is a max DE 2000 of 6 and a med DE 2000 of 2 between this instruments; using them to validate the other-generated profile.

You can find both profiles here for DL (one .zip, size about 200k): http://www.megaupload.com/?d=G5JR23W0

i1Pro-generated profile with CM validation:
(http://www.pctunerup.com/up/results/_201101/20110111155138_colorMunki01.JPG)

I know that this test does not show if one hardware is better the the other, what is seems to me is that this hardware are different. What do you think?
Title: Re: ColorMunki VS i1Pro - Monitor profile validation
Post by: sarangiman on January 11, 2011, 10:03:31 pm
Wow. This is a really cool test/post.

You used SpectraView for all tests?

This sort of result worries me. I two spectrophotometers are giving different results, then who's to say either one profiles your monitor correctly?! I thought spectrophotometers remove the variability from the equation by not using color filters like colorimeters do. People complain that if a colorimeter is not paired to the primaries of the monitor (er, a little shaky with LCDs no?), you're not guaranteed that the colorimeter accurately profiles your device. Enter spectrophotometers with their gratings.

But your test suggests that even spectrophotometers are not infallible. Which is disconcerting.

This brings up another question I've had for some time. Just because you validate your profile using your own colorimeter/software (and say you get a low average deltaE), doesn't mean that you're getting accurate colors, does it? Because what if your colorimeter's reading is clipped in one or more channels b/c the monitor's saturation of some colors is too high (e.g. a wide gamut monitor)? When you go to validate, with the same colorimeter, the profile you built with this colorimeter, you wouldn't be able to account for hardware failure. In fact, isn't colorimeter variability even within the same model/make a major problem? I know I saw this documented somewhere...

As I understood it, the validator can only tell you how good your profile is *given* that you have good/accurate hardware. And I don't know of any way for me to test that. I assume X-Rite or whatever manufacturer has a way to test/calibrate a colorimeter/spectrophotometer.

All of this became a major concern to me when I got a new wide-gamut Dell U2711 monitor. My i1 Display 2 creates terrible profiles for it. Although it itself says that the max deltaE is like 3.97, I can see that reds are completely shifted to magenta (blue-shifted) in color-managed applications. Which causes me to compensate by adding orange back to my reds to shift them back to a reasonable hue. But then when I print or view on any other monitor, my so-called 'reds' are now orange.

I recently bought a Color Munki to see if this is a hardware-related problem, but haven't had a chance to test yet. Your results make me quite uncertain about this whole field of color management & device profiling in general!

Rishi
Title: Re: ColorMunki VS i1Pro - Monitor profile validation
Post by: sarangiman on January 11, 2011, 11:40:56 pm
Wait, I'm sorry. I need to correct my previous statement. After profiling my U2711 with my i1 Display 2, reds go to a desaturated orange in color-managed applications. I tried to compensate by blue-shifting & saturating my reds, these same reds appear magenta & over-saturated in other output devices (my MacBook Pro, my Dell 2005FPW, an Epson 9900 print... consistently across all these output devices, buttressing my opinion that it's my U2711 that is incorrectly profiled).
Title: Re: ColorMunki VS i1Pro - Monitor profile validation
Post by: Czornyj on January 12, 2011, 04:17:41 am
Generl conclusion:
It seems that there is a max DE 2000 of 6 and a med DE 2000 of 2 between this instruments; using them to validate the other-generated profile.
This is weird... Could you try to create matrix instead of type LUT type profiles, and cross-validate them with both spectros? I had decent inter-instrumental agreement between the i1pro and CM spectros that I had played with, but I didn't try to compare LUT profiles...
Title: Re: ColorMunki VS i1Pro - Monitor profile validation
Post by: VitOne on January 12, 2011, 02:23:21 pm
I used SpectraView Profiler (basICColor software) to make all the measurements.
Here I would like to talk about spectrophotometers, there are a lot of discussions about colorimeters but I would like to don’t speak about them in this post, I hope you can understand why.
The validation process, for what I know about colour management, should not be used to check how good a profile is. You can find some information here about validation process in ArgyllCMS: http://www.argyllcms.com/doc/profcheck.html
I had both instruments and I just wanted to see if they behavior was similar or not, because I read many posts where people is saying that i1Pro and ColorMunki are share the same basic hardware and they differ only for the light source (A vs LED).
Now I really don’t know why the validation process shows different results (I don’t know enough about how the spectrophotometers work) but, for what I understand, my ColorMunki and my i1Pro D are different.
Using both profiles I can see some differences using some test images, but with prints I can see no big differences, probably because the patches where I have the bigger DeltaEs are outside the gamut of the printer I am using (Epson 7900 with Harman Gloss Baryta paper).
(http://www.pctunerup.com/up/results/_201101/20110112201939_Prova.JPG)
 
This is weird... Could you try to create matrix instead of type LUT type profiles, and cross-validate them with both spectros?
Right now I am only using table-based profiles (LUT) but I can make this test if you want.
Title: Re: ColorMunki VS i1Pro - Monitor profile validation
Post by: stefohl on January 15, 2011, 05:24:32 am

The validation process, for what I know about colour management, should not be used to check how good a profile is.


A better way to test the quality of the profile and the monitor is U-Dact from UGRA. You will get a report that show if the colours in the FOGRA test wedge is shown correct on the monitor.

I've done a test of almost 50 calibrators, both spectrophotometers and colorimeters. I did it by first testing my EyeOne Pro against a high-end Minolta CA-210, and they showed pretty much the same results. I then calibrated two monitors, one Eizo CG211 which has a colour gamut a bit larger than sRGB, and a Eizo CG 221 which has a gamut about the same as Adobe RGB. Then I did a validation with my calibrator. When I did the validation with the other calibrators, I could compare the validation reports. They ought to be about the same.

The results showed that about 40% of the colorimeters showed a result that deviated more than Delta E 5 on the average from the validation report created by my EyeOne Pro or a peak DeltaE above 10. This was with EyeOne Displays and Spyder III:s. We didn't test that many Optix, but they showed a better result.
Title: Re: ColorMunki VS i1Pro - Monitor profile validation
Post by: VitOne on January 17, 2011, 04:31:23 am
Thanks for the answer. I am not searching for the right way to test the monitor (I already used the UGRA software, the screen is from this report: http://www.megaupload.com/?d=RT6ROI95). I used the same hardware (i1Pro RevD) to calibrate and to run the UGRA test.
I was asking myself and others how we can trust ColorMunki and i1Pro for monitor calibration if they have this difference in their behaviour. Also I was asking myself if it is correct when we speak about “monitor validation with low DeltaE”.
I read your post there: http://www.luminous-landscape.com/forum/index.php?topic=47162 and the link http://lists.apple.com/archives/colorsync-users/2009/Nov/msg00175.html
Did anybody compare two or more i1Pro rev together?
Title: Re: ColorMunki VS i1Pro - Monitor profile validation
Post by: probep on January 17, 2011, 08:22:02 am
Did anybody compare two or more i1Pro rev together?
I didn't compare different revisions, but I compared four i1Pros Rev.D (with and without UV-filter) on NEC 2690WUXi2. The max color difference between them was 0.55 [deltaE CIE2000].
Max deltaE [CIE2000] between these i1Pros and a ColorMunki Photo was 1.24