I was wondering if PA246Q User Mode is still buggy, so I decided to test it by myself. As you may be interested in results and there are some questions regarding them I present results below.
The test was performed with the monitor connected via DP to Radeon HD6600. Resolution was set to 1920x1200 and colour depth to 32bpp.
Argyll CMS 1.3.4 and dispcalGUI 0.7.3.7 were used. Huey Pro colorimeter along with correction matrix for the LM240WU4-SLB1 panel (derived by using ccmxmake utility from Argyll suite, Color Munki Design spectrophotometer and NEC SV241 monitor with the mentioned panel).
Applied CCMX correction matrix:
0.84308 -0.012484 0.028477
0.023148 0.83371 -3.9284e-003
-0.013409 0.025818 0.94938
Tested PA246Q was produced in mid-April ("ISO week 14" according to EDID info).
Calibration target was: sRGB, whitepoint 6500K, whitelevel 140cd/m2, native black point, ambient light level 136.6lux, profile type XYZ LUT + matrix.
At first User Mode was selected and hardware calibration was performed. Contrast was left at it's default value (80), and brightness and RGB gains were adjusted to reach desired whitepoint and white level.
Target Br 140.00, x 0.3128 , y 0.3291
Current Br 140.37, x 0.3130-, y 0.3286+ DE 0.5 R- G++ B-
Black = XYZ 0.19 0.20 0.29
Red = XYZ 95.17 43.74 1.55
Green = XYZ 20.26 65.53 8.72
Blue = XYZ 27.68 10.18 146.95
White = XYZ 131.36 138.14 150.93
Sum = XYZ 1.042 1.060 0.8984 (normalized)
Normalized sum of RY+GY+BY is close to 1 - 1.06. Normalized sum of RZ+GZ+BZ isn't close - 0.8984. Is it OK - I'm not sure.
After calibration profiling was performed and profile quality was checked. Verification has shown average dE*76=0.61 and maximum dE*76=2.81 for verify_extended.ti1 (take a look at attached: verify_2011-09-05_18-10.html). For verify.ti1 it has shown 0.3 and 0.86 (take a look at attached: verify_2011-09-05_18-08.html).
On one hand normalized sum of RZ+GZ+BZ is not close to 1, and on the other profile verification shows that the monitor's characteristics was captured correctly as if the user mode worked fine...
Next colour accuracy was verified against sRGB IEC61966-2.1 simulation profile. Results were bit surprising - average dE*76=1.82 and maximum dE*76=10.9 (take a look at: verify_2011-09-05_18-13.html)! Why maximum dE*76 was so big?
Then the monitor was calibrated and profiled from the Standard Mode. In the mode only white level could have been set to desired value.
Target 140.00
Current 141.52 -
Black = XYZ 0.19 0.20 0.30
Red = XYZ 83.84 38.07 1.38
Green = XYZ 28.81 93.11 12.19
Blue = XYZ 26.85 9.65 142.82
White = XYZ 139.41 140.63 155.97
Sum = XYZ 1.045 1.057 0.8985
After calibration profiling was performed and profile quality was verified. Verification has shown average dE*76=0.60 and maximum dE*76=2.80 for verify_extended.ti1 (take a look at attached: verify_2011-09-05_19-03.html).
Once again normalized sum of RZ+GZ+BZ is not close to 1, but profile verification shows that the monitor's characteristics was captured correctly.
Next colour accuracy was verified against sRGB IEC61966-2.1 simulation profile. Results were even more surprising - average dE*76=1.84 and maximum dE*76 as high as 6.36 (take a look at: verify_2011-09-05_19-05.html)! Why maximum dE*76 was so big?
Prad.de calibration and profiling from Standard Mode was successful, and this one was not. Something seems to went wrong. But what?
Is it the case of the panel chromacity spectrum discreepency between two different LM240WU4-SLB1 units causing CCMX correction matrix being inadequate for the particular panel used in the PA246Q tested?
Revealed errors couldn't be results of some colours falling out of a gamut as all sRGB colours fall inside of the panel gamut.
For more detailed results and log files look here .