[argyllcms] How best to calibrate a wide gamut LCD?
- From: Stephen T <stwebvanuatu@xxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Wed, 25 May 2011 03:00:34 -0700 (PDT)
Hello,
I have recently acquired an NEC P221W LCD display and
MDSVsensor (NEC custom calibrated X-Rite i1Display2 colorimeter). I'm
not running NEC's Spectraview II software because it can't communicate
with the monitor over DDC link. NEC said my graphics adapter (Intel HD
graphics) is not compatible.
My current plan is to calibrate
and profile the NEC P221W with ArgyllCMS. I've been using Argyll CMS for
a couple of years now on CRT, laptop LCD, WLED LCD and standard-gamut
LCD and am pretty happy with it.
OK, so here's the native response of my P221W:
C:\Argyll_V1.3.2\bin>dispcal -yl -r -P 0.5,0.5,2.0
...
Black level = 0.09 cd/m^2
White level = 122.98 cd/m^2 (at brightness 20%)
Aprox. gamma = 2.27
Contrast ratio = 1319:1
White chromaticity coordinates 0.2926, 0.3653
White Correlated Color Temperature = 7174K, DE 2K to locus = 22.3
White Correlated Daylight Temperature = 7144K, DE 2K to locus = 21.5
White Visual Color Temperature = 6102K, DE 2K to locus = 22.1
White Visual Daylight Temperature = 6148K, DE 2K to locus = 21.2
The black level and contrast ratio are great.
The colour temperature is a bit high. I verified this with my old Spyder2
colorimeter:
Black level = 0.23 cd/m^2
White level = 129.00 cd/m^2 (at brightness 20%)
Aprox. gamma = 2.11
Contrast ratio = 549:1
White chromaticity coordinates 0.2923, 0.3691
White Correlated Color Temperature = 7146K, DE 2K to locus = 22.8
White Correlated Daylight Temperature = 7115K, DE 2K to locus = 22.0
White Visual Color Temperature = 6083K, DE 2K to locus
= 22.6
White Visual Daylight Temperature = 6116K, DE 2K to locus = 21.8
Never
mind the difference in black level and CR, the white point chromacity
is nearly identical and I believe my MDSVsensor is fairly accurate. I
believe the Spyder2 should measure whites OK even though it is not
suitable for measuring extremes of the wide gamut.
The problem
above is the large DE 2K to locus = 22.0. I can see a greenish cast in
the display. The DE is still unsatisfactory when switching to sRGB
emulation mode:
Black level = 0.09 cd/m^2
White level = 100.31 cd/m^2 (at brightness 20%, with MDSVsensor)
Aprox. gamma = 2.25
Contrast ratio = 1080:1
White chromaticity coordinates 0.2987, 0.3565
White Correlated Color Temperature = 6973K, DE 2K to locus = 20.0
White Correlated Daylight Temperature = 6951K, DE 2K to locus = 18.6
White Visual Color Temperature = 6049K, DE 2K to locus = 19.7
White Visual Daylight Temperature = 6139K, DE 2K to locus = 18.3
I
noticed that NEC has set different RGB levels for different colour
temperature settings. I feel that hardware adjustment with the 10-bit
LUT should be better than software adjustment in the 8-bit video LUT. I
used dispcal to reduce the green level and shift the white point on to
the daylight locus:
C:\Argyll_V1.3.2\bin>dispcal -v -yl -g2.2 -b120 NECP221W
...
Display type is LCD
Target white = native white point
Target white brightness = 120.000000 cd/m^2
Target black brightness = native brightness
Target advertised gamma = 2.200000
...
Adjust R,G & B gain to desired white point. Press space when done.
Initial Br 120.00, x 0.2929 , y 0.3646 , VDT 6149K DE 2K 21.0
\ Current
Br 119.05, x 0.3026+, y 0.3187- VDT 7205K DE 2K 0.1 R= G= B=
This was achieved with R=100%, G=78.6%, B=100% and I have adjusted brightness
as well.
Only a little more adjustment was required for D65 at R=100%, G=78.6%, B=92.7%.
For comparison, NEC's 5000K setting is R=100%, G=88%, B=69.2%.
Here are the final results:
patch 64 of 64
Brightness error = 1.691718 cd/m^2 (is 121.035284, should be 119.343567)
White point error = 0.233073 deltaE
Maximum neutral error (@ 0.828240) = 0.771539 deltaE
Average neutral error = 0.353793 deltaE
Failed to meet target 0.600000 delta E, got worst case 0.685466
Number of measurements taken = 180
The instrument can be removed from the screen.
Written calibration file 'NECP221W.cal'
The difference between AdobeRGB (wide gamut) and sRGB (standard gamut) is
mainly in the green point. The problem with
lowering the green channel is that I have also reduced the gamut of the
display.
The
colours look quite neutral to me after calibration, which is a good
thing for RAW photo conversion and white-balancing. I think this is more
important than the wide gamut for me. I was hoping for better
(perfection) though.
Now am I doing this right? Should I be
making adjustments in hardware or letting the LUT take care of
everything? Does the wide gamut really come at the expense of the
neutral white point? Are there better recommendations for calibration
and profiling of such a display?
Stephen.
Other related posts:
- » [argyllcms] How best to calibrate a wide gamut LCD? - Stephen T
