[argyllcms] Calibration issue with colormunki
- From: Emmanuel Michonneau <emmanuel.michonneau@xxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Sun, 21 Nov 2010 15:05:21 +0100
Hello everybody,
I recently bought a x-rite colormunki photo to replace my old spyder 2
colorimeter.
After a week of testing it I am very disapointed because I am not able to
get any right profile of my display.
When I make a profile with the colormunki, dark tones are too bright and
lose a lot of details and contrast.
I have investigated on this problem but I don't understand why the
colormunki results are so bad.
Here are some tests made to compare results from colormunki with old of
spyder2 (which give me good results excepts on the green tones that are
oversatured due to my extended gammut display)
First, the "check all" results (option 5) of the discal prompt (the display
is working with factory settings except for luminosity):
Colormunki :
Doing check measurements
Black = XYZ 0.11 0.12 0.18
Grey = XYZ 16.64 18.55 27.83
White = XYZ 84.09 91.31 106.21
1% = XYZ 0.73 0.82 1.11
Target Brightness = 90.00, Current = 91.31, error = 1.5%
Target 50% Level = 19.87, Current = 18.55, error = -1.5%
Target Near Black = 0.91, Current = 0.95, error = 0.0%
Target white = x 0.3128, y 0.3291, Current = x 0.2986, y 0.3242, error =
7.51 DE
Target black = x 0.3128, y 0.3291, Current = x 0.2748, y 0.3085, error =
14.63 DE
Spyder2:
Doing check measurements
Black = XYZ 0.27 0.24 0.35
Grey = XYZ 18.73 20.43 29.86
White = XYZ 81.79 86.07 101.71
1% = XYZ 1.19 1.28 1.59
Target Brightness = 90.00, Current = 86.07, error = -4.4%
Target 50% Level = 18.73, Current = 20.43, error = 1.9%
Target Near Black = 0.86, Current = 1.35, error = 0.5%
Target white = x 0.3128, y 0.3291, Current = x 0.3034, y 0.3193, error =
4.94 DE
Target black = x 0.3128, y 0.3291, Current = x 0.2930, y 0.3152, error =
8.55 DE
=> There is a big difference in black and near black reading between two
instruments. How the colormunki can "see" less ligth than the spyder2 in
this case?
Second results are the calibration curves generated with dispcal (dispcal -v
-qm -yl -t6500 -b90 -g2.2)
http://emmanuel.michonneau.free.fr/pub/Curves_colormunki.png
http://emmanuel.michonneau.free.fr/pub/Curves_spyder.png
We can see clearly that the colormunki pull up the dark tones whereas the
spyder 2 does not.
Another thing that i don't understand is why black point of the colormunki
curve does not have a zero response.
I have tried to use adaptive and hires modes of the colormunki. This
improves a little the results in calibration but the generated profiles are
still too brigth in darks tones.
Where is the mistake?
Do I miss something in the use of colormunki?
Is my own colormunki broken?
Is my display not compatible with the munki? (It's an HPw2408h with a TN
panel and wide gammut)
Thanks for your help.
Other related posts:
