- This topic has 15 replies, 2 voices, and was last updated 6 years, 8 months ago by
.
-
Posts
-
Hello Florian,
I have two different ColorMunki Colorimeters (Display/smile) and both show inconsistencies measuring gamma.
I found that If either of the ColorMunki devices haven’t been used for a little while (gone cold) and I do a uncalibrated reading in DisplayCal, the results I get in further readings are inconsistent.
I did some tests below using the ColorMunki Display, using xrites w-led correction (monitor is Acer x34a) and these are the results :
1.9 monitor setting
1st attempt
18:22:24,105 Uncalibrated response:
18:22:24,105 Black level = 0.2489 cd/m^2
18:22:24,105 50% level = 82.44 cd/m^2
18:22:24,105 White level = 276.29 cd/m^2
18:22:24,107 Aprox. gamma = 1.74
18:22:24,107 Contrast ratio = 1110:1
18:22:24,107 White chromaticity coordinates 0.3078, 0.3215
18:22:24,107 White Correlated Color Temperature = 6852K, DE 2K to locus = 2.8
18:22:24,108 White Correlated Daylight Temperature = 6857K, DE 2K to locus = 2.1
18:22:24,108 White Visual Color Temperature = 6739K, DE 2K to locus = 2.6
18:22:24,108 White Visual Daylight Temperature = 6947K, DE 2K to locus = 2.0
18:22:24,108 Effective Video LUT entry depth seems to be 8 bits
18:22:24,108 Black drift was 0.000000 DE
18:22:24,108 White drift was 0.000000 DE
18:22:24,108 The instrument can be removed from the screen.
18:22:24,325 DisplayCAL: Reached EOF (OK)
18:22:24,3762nd attempt :
18:58:10,459 Uncalibrated response:
18:58:10,459 Black level = 0.2558 cd/m^2
18:58:10,460 50% level = 84.84 cd/m^2
18:58:10,460 White level = 277.36 cd/m^2
18:58:10,460 Aprox. gamma = 1.71
18:58:10,460 Contrast ratio = 1084:1
18:58:10,460 White chromaticity coordinates 0.3054, 0.3180
18:58:10,460 White Correlated Color Temperature = 7044K, DE 2K to locus = 2.0
18:58:10,460 White Correlated Daylight Temperature = 7050K, DE 2K to locus = 2.8
18:58:10,461 White Visual Color Temperature = 6954K, DE 2K to locus = 1.9
18:58:10,461 White Visual Daylight Temperature = 7183K, DE 2K to locus = 2.7
18:58:10,461 Effective Video LUT entry depth seems to be 8 bits
18:58:10,461 Black drift was 0.000000 DE
18:58:10,463 White drift was 0.000000 DE2.2 monitor setting
1st reading :
18:15:47,746 Uncalibrated response:
18:15:47,746 Black level = 0.2467 cd/m^2
18:15:47,746 50% level = 67.45 cd/m^2
18:15:47,746 White level = 275.97 cd/m^2
18:15:47,746 Aprox. gamma = 2.03
18:15:47,746 Contrast ratio = 1118:1
18:15:47,746 White chromaticity coordinates 0.3088, 0.3233
18:15:47,746 White Correlated Color Temperature = 6773K, DE 2K to locus = 3.4
18:15:47,747 White Correlated Daylight Temperature = 6778K, DE 2K to locus = 1.5
18:15:47,747 White Visual Color Temperature = 6639K, DE 2K to locus = 3.2
18:15:47,747 White Visual Daylight Temperature = 6836K, DE 2K to locus = 1.4
18:15:47,749 Effective Video LUT entry depth seems to be 8 bits
18:15:47,749 Black drift was 0.000000 DE
18:15:47,749 White drift was 0.000000 DE
18:15:47,749 The instrument can be removed from the screen.
18:15:47,915 DisplayCAL: Reached EOF (OK)2nd reading :
19:00:27,736 Uncalibrated response:
19:00:27,737 Black level = 0.2560 cd/m^2
19:00:27,737 50% level = 70.14 cd/m^2
19:00:27,737 White level = 277.33 cd/m^2
19:00:27,737 Aprox. gamma = 1.98
19:00:27,737 Contrast ratio = 1083:1
19:00:27,739 White chromaticity coordinates 0.3053, 0.3180
19:00:27,739 White Correlated Color Temperature = 7050K, DE 2K to locus = 2.1
19:00:27,739 White Correlated Daylight Temperature = 7056K, DE 2K to locus = 2.8
19:00:27,740 White Visual Color Temperature = 6959K, DE 2K to locus = 1.9
19:00:27,740 White Visual Daylight Temperature = 7189K, DE 2K to locus = 2.7
19:00:27,740 Effective Video LUT entry depth seems to be 8 bits
19:00:27,740 Black drift was 0.000000 DE
19:00:27,740 White drift was 0.000000 DE
19:00:27,740 The instrument can be removed from the screen.
19:00:27,957 DisplayCAL: Reached EOF (OK)2.5 monitor setting
1st reading :
18:52:10,430 Uncalibrated response:
18:52:10,430 Black level = 0.2554 cd/m^2
18:52:10,430 50% level = 57.74 cd/m^2
18:52:10,431 White level = 276.56 cd/m^2
18:52:10,431 Aprox. gamma = 2.26
18:52:10,431 Contrast ratio = 1083:1
18:52:10,431 White chromaticity coordinates 0.3057, 0.3184
18:52:10,433 White Correlated Color Temperature = 7016K, DE 2K to locus = 2.1
18:52:10,433 White Correlated Daylight Temperature = 7022K, DE 2K to locus = 2.8
18:52:10,433 White Visual Color Temperature = 6925K, DE 2K to locus = 2.0
18:52:10,434 White Visual Daylight Temperature = 7152K, DE 2K to locus = 2.7
18:52:10,434 Effective Video LUT entry depth seems to be 8 bits
18:52:10,434 Black drift was 0.000000 DE
18:52:10,434 White drift was 0.000000 DE
18:52:10,434 The instrument can be removed from the screen.
18:52:10,601 DisplayCAL: Reached EOF (OK)
18:52:10,6432nd reading :
19:01:55,437 Uncalibrated response:
19:01:55,437 Black level = 0.2561 cd/m^2
19:01:55,437 50% level = 57.99 cd/m^2
19:01:55,437 White level = 276.61 cd/m^2
19:01:55,437 Aprox. gamma = 2.25
19:01:55,438 Contrast ratio = 1080:1
19:01:55,438 White chromaticity coordinates 0.3055, 0.3181
19:01:55,438 White Correlated Color Temperature = 7035K, DE 2K to locus = 2.0
19:01:55,440 White Correlated Daylight Temperature = 7041K, DE 2K to locus = 2.9
19:01:55,440 White Visual Color Temperature = 6947K, DE 2K to locus = 1.9
19:01:55,440 White Visual Daylight Temperature = 7176K, DE 2K to locus = 2.7
19:01:55,440 Effective Video LUT entry depth seems to be 8 bits
19:01:55,440 Black drift was 0.000000 DE
19:01:55,440 White drift was 0.000000 DE
19:01:55,440 The instrument can be removed from the screen.
19:01:55,607 DisplayCAL: Reached EOF (OK)Gaming mode monitor setting
1st reading :
18:53:42,188 Uncalibrated response:
18:53:42,188 Black level = 0.2556 cd/m^2
18:53:42,188 50% level = 48.07 cd/m^2
18:53:42,188 White level = 276.57 cd/m^2
18:53:42,190 Aprox. gamma = 2.52
18:53:42,190 Contrast ratio = 1082:1
18:53:42,190 White chromaticity coordinates 0.3057, 0.3184
18:53:42,190 White Correlated Color Temperature = 7019K, DE 2K to locus = 2.0
18:53:42,190 White Correlated Daylight Temperature = 7025K, DE 2K to locus = 2.8
18:53:42,190 White Visual Color Temperature = 6930K, DE 2K to locus = 1.9
18:53:42,190 White Visual Daylight Temperature = 7157K, DE 2K to locus = 2.7
18:53:42,191 Effective Video LUT entry depth seems to be 8 bits
18:53:42,191 Black drift was 0.000000 DE
18:53:42,191 White drift was 0.000000 DE
18:53:42,191 The instrument can be removed from the screen.
18:53:42,358 DisplayCAL: Reached EOF (OK)
18:53:42,4012nd reading
19:03:18,381 Uncalibrated response:
19:03:18,381 Black level = 0.2561 cd/m^2
19:03:18,381 50% level = 48.21 cd/m^2
19:03:18,381 White level = 276.58 cd/m^2
19:03:18,381 Aprox. gamma = 2.52
19:03:18,381 Contrast ratio = 1080:1
19:03:18,382 White chromaticity coordinates 0.3054, 0.3180
19:03:18,382 White Correlated Color Temperature = 7041K, DE 2K to locus = 2.0
19:03:18,382 White Correlated Daylight Temperature = 7048K, DE 2K to locus = 2.9
19:03:18,382 White Visual Color Temperature = 6953K, DE 2K to locus = 1.9
19:03:18,384 White Visual Daylight Temperature = 7183K, DE 2K to locus = 2.7
19:03:18,384 Effective Video LUT entry depth seems to be 8 bits
19:03:18,384 Black drift was 0.000000 DE
19:03:18,384 White drift was 0.000000 DE
19:03:18,384 The instrument can be removed from the screen.
19:03:18,552 DisplayCAL: Reached EOF (OK)I did the tests in order shown, just as I reached the 2nd reading of the “gaming gamma mode” (a couple minutes from a cold reading), the 1st and 2nd readings are almost identical, this leads me to think that this colorimeter needs to warm up each time to report accurately.
Surely this can lead to inaccuracies in calibration, especially if people use their first “uncalibrated report readings” as target.
Calibrite Display SL on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.My bad, my before/after times are a little off, I said the 2nd reading of the gaming mode was 2minutes apart, it was 10 minutes. I tried to edit again, but i reached the maximum edits.
I see no inconsistency whatsoever in your results (gamma is also a very poor indicator of measurement consistency), all measurements seem well within the expected repeatabilty with a typical LCD display.
Oh ok, I was seeing the gamma measurement change and I was getting a little too worried :/
Is there a way to see the gamma curve in DisplayCal on the fly or do I have to do a measurement report to see it each time?
I’ve a little issue with green gamma more than anything, if i change the value ever so slightly, say -1 or +1 (testing with Nvidia’s gamma setting) , I get terrible banding in the dark green shades, and grayscale, even if using a single curve profile – this is also reflected when a profile is applied. I have tried to tweak the grayscale beforehand using hcfr (in srgb with same w-led correction), in hope that is would help tweak before lut changes are applied – and usually can it pretty flat yet but once I fire up DisplayCals interactive display, white balance is reported as way off. I’m assuming this is because the settings in hcfr and DisplayCal are not matched?
I would love to be able to add a note to profiles in DisplayCal so that I can refer to the values changed on the monitor. Would that be possible?
I’ll be donating again soon (payday) as you always seem to put the effort in helping us. Thanks again!
Oh ok, I was seeing the gamma measurement change and I was getting a little too worried :/
A gamma fluctuation of +-0.05 peak is insignificant, so no need to worry.
Is there a way to see the gamma curve in DisplayCal on the fly or do I have to do a measurement report to see it each time?
You can use the profile information to view the tone response curves.
testing with Nvidia’s gamma setting
Note that nVidia’s control panel AFAIk changes the gamma (and other color related settings) in software, via the videoLUT (thus it cannot be used for calibration), and potentially limited to 8-bit.
I would love to be able to add a note to profiles in DisplayCal so that I can refer to the values changed on the monitor. Would that be possible?
You actually can store that information in the profile metadata when uploading it to the online database (menu “File”, “Upload profile”). If you don’t want to upload the profile, just click “Save”. You can retrieve the it again by looking in the metadata section when viewing profile information.
I’ll be donating again soon (payday) as you always seem to put the effort in helping us
Thank you, appreciated 🙂
Hi Florian,
I can’t achieve a grayscale without red/green/yellow discoloration. The default ColorMunki Display (edit – Display, not smile) app results in discoloration in the Grayscale, as well as DisplayCal (using single curve/xyz defaults)
DisplayCal reads my monitors 2.2 gamma setting as 2.0, and 2.4 as 2.27. Regardless of which setting I use and even if I target the gamma from uncalibrated report, I still see coloration in a grayscale gradient
Are there any steps I should take to achieve a neutral grayscale, even at a loss of color accuracy? I’m not too bothered about banding, just the red/green/yellow tints.
I have tried upping the neutral patches to 256, which helps just slightly but I still see a color cast.
My monitors grayscale looks clean (no color) when using a linear curve. Is my Nvidia 1080ti to blame perhaps? From what I’ve gathered from your forum is I’m to leave the contrast setting alone, which I have done.
It’s really bugging me as i’m spending way too much time on this.
Monitor set to defaults, only the gamma changed to 2.5 (to get as close to 2.2 as possible).
I appreciate all the help I can get. Once I achieve a neutral grayscale profile I can rest.
Profile looks fine, report as well, I see no room for improvement with such (already very good) results. How do you evaluate grayscale?
Heya, thanks for replying so quick 🙂
I usually open a grayscale gradient using an external image viewer, such as Windows Photo viewer or Irfanview. I observe the grayscale using a linear curve (no profile applied) and then apply the profile via DisplayCal profile loader to see any differences.
I have also seen the grayscale being affected whilst gaming, where dark shades of gray would appear green/reddish. The color cast of the gray closely resembles that I see when observing a Grayscale gradient on the dekstop (with 1d profile applied)
Most of the discoloration I see in the grayscale is “near black”. The mid tones/lighter tones (near white) are pretty smooth.
I think green may be running out of color as I see heavy banding in a green gradient – after doing a calibration (with both ColorMunki software/DisplayCal). Blue and Red isn’t so bad. Is there anything I can do to help prevent this?
The lesson to be learned here is that processing in 8 bit may introduce quantization when no dithering is applied or no noise is present.
The lesson to be learned here is that processing in 8 bit may introduce quantization when no dithering is applied or no noise is present.
What I don’t quite understand is, with all these colorimeters being sold, there isn’t much to be found on the web that relates to my issue. Surely many people with calibrators will be using 8bit monitors with non quadro/ati cards and be pulling their hair out when they see color greys looking green/pink (which I have been!)
Typical usage doesn’t involve looking at synthetic grayscale gradients all day long. If you’re bothered by the colorization, just add a tiny amount of noise to the gradient. Voila, the problem goes away.
Hello Florian,
I use the 1d calibration for desktop use/playing Forza Horizon 3 (doesn’t support 3dlut) as my monitors gamma is either too dark/too bright without it. If I could add noise to the 1d I would in a heartbeat.
I found an image which looks very similar to what I’m seeing after applying the profiles 1d, notice the red banding? I see stripes of green as well mixed in.
I need to establish if my monitor is faulty or is this normal? (Considering I’m using an Nvidia gaming gpu) before my 30 day return window expires.
Thank you.
I use the 1d calibration for desktop use/playing Forza Horizon 3 (doesn’t support 3dlut) as my monitors gamma is either too dark/too bright without it.
Does the banding/colorization go away if you reset the video card gamma table to linear?
I found an image which looks very similar to what I’m seeing after applying the profiles 1d, notice the red banding? I see stripes of green as well mixed in.
I assume this is a photo of a monitor in unknown state. So, graphics card -> monitor -> camera -> camera processing -> JPEG -> browser color management -> graphics card -> monitor. Too many steps and components involved to really draw any conclusions without knowing the exact specifics of each.
Does the banding/colorization go away if you reset the video card gamma table to linear?
Yes, its not visible when i set the video card gamma to linear.
I think I’ve found why I’m seeing all these red/pink stripes near black, at 0 there is way too blue..
Also creating a uncalibrated measurement report in DisplayCal (against srgb) displays gray balance having too much blue near black.
Unfortunately I only have Gain controls on my monitor, and therefore can’t correct near black this using the monitor alone. 🙁
Is there anything I can do to help correct this in DisplayCal Florian?
