- This topic has 15 replies, 3 voices, and was last updated 6 years, 11 months ago by
.
-
Posts
-
Hi,
I’ve been calibrating my displays with the X-rite software for the Colormunki display since I got it. I’ve also used DisplayCal in the past when I previously had a Spyder4. I felt like the X-rite software did a decent job, but I decided to give DisplayCal a whirl with the colormunki colorimeter. I’m calibrating an Acer XB271HU (IPS, WLED).
I configured the RGB balance in displaycal and calibrated to a 6500K target with gamma 2.2, using the “High” speed setting. All other settings were left at default, and colorimeter corrections were imported from the software. Compared to calibration from the X-rite software using the same 6500K and gamma 2.2 targets, the displaycal calibration has a warmer white visually. The X-rite software looks more green in the whites, while display appears more red. I’m not surprised theres a difference, but I am surprised at how much of a visual difference there is. Is this normal?
Also the two calibrations are vastly different in the gamma characteristics of the darker gray tones. Using the Lagom LCD test page (http://www.lagom.nl/lcd-test/black.php), calibration with the X-rite software is much darker from boxes 1-10, and in particular boxes 1-3 are barely visible. The displaycal calibration is lighter in these tones, and boxes 2-5 are distinctly more visible. Why would the two softwares be handling gamma so differently in the lower end? Is there a setting in displaycal I should be changing? I see there are settings for black offset and such.
Just wondering if these differences can be attributed to anything, since both programs are calibrating using the same colorimeter. Which software should I trust? 🙂
Thanks!
Calibrite Display SL on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.Hi,
your questions should be answered by the FAQ I’ve just added.
Which software should I trust?
The more accurate one 🙂
Thanks for the FAQ. Can I assume that “the more accurate one” means DisplayCal? haha
Follow up question: does the bit depth of the panel affect the calibration in terms of low level, or near black gamma performance? My previous monitor, a Dell U2414H (8 bit, 6 bit + FRC) performs quite differently in lower level grays compared to the Acer XB271HU (10 bit, using 8 bit + frc) with both monitors calibrated with displaycal and the colormunki display colorimeter. The gamma appears lower on the Dell, with darker squares appearing more visible than on the Acer. The Acer almost appears to have some black crush. I am taking care to make sure that the associated profile for each monitor is active when viewing them separately.
Should lower level gray squares be pretty difficult to see on a display calibrated to gamma 2.2?
Can I assume that “the more accurate one” means DisplayCal? haha
Possibly, although that probably comes down to the amount of patches measured, and the type of profile created – DisplayCAL measures more than most other profilers and creates cLUT profiles by default. I would assume that when you set i1 Profiler up to create cLUT profiles as well and measure a similar amount of patches (which can be done in the advanced interface), that the differences would become small.
Follow up question: does the bit depth of the panel affect the calibration in terms of low level, or near black gamma performance?
If the effective bit depth is lower than 8-bit, this might affect the time taken during calibration and the attainable accuracy (although probably only negligibly), because it’ll become harder to match the delta E tolerances due to the coarser sampling.
The gamma appears lower on the Dell, with darker squares appearing more visible than on the Acer.
It probably comes down to a difference in contrast – I suspect the Acer to be capable of lower black level than the Dell (although probably not a big difference). If you want to match them, calibrate the lower black level display to the same (higher) of the other one (and same white level as well obviously).
Should lower level gray squares be pretty difficult to see on a display calibrated to gamma 2.2?
It depends on the increment step size, but generally yes.
Possibly, although that probably comes down to the amount of patches measured, and the type of profile created – DisplayCAL measures more than most other profilers and creates cLUT profiles by default. I would assume that when you set i1 Profiler up to create cLUT profiles as well and measure a similar amount of patches (which can be done in the advanced interface), that the differences would become small.
I have the colormunki display. The X-rite software doesn’t seem to offer the options you are referring to. Maybe thats on a higher version?
It probably comes down to a difference in contrast – I suspect the Acer to be capable of lower black level than the Dell (although probably not a big difference). If you want to match them, calibrate the lower black level display to the same (higher) of the other one (and same white level as well obviously).
The Acer does indeed have a lower black level. I’m measuring close to 0.10, where on the Dell it was about 0.12. Does this mean the colormunki device is able to detect the lower levels better on the display that has a lower black level? (and thus correctly apply the gamma function)
It depends on the increment step size, but generally yes.
Taking into account what you said about the lagom test patterns in the FAQ, I’m referring to the black level test on that site. On the Acer, the squares from 1-3 are almost invisible, I believe I can just BARELY see them in the dark with my face to the screen. On the Dell these were distinctly visible.
I have the colormunki display. The X-rite software doesn’t seem to offer the options you are referring to. Maybe thats on a higher version?
Yep, I think the ColorMunki software is stripped down, only i1 Profiler will have them (although I’m not sure it’ll unlock them with the ColorMunki, that may only work with the i1 Display Pro).
Does this mean the colormunki device is able to detect the lower levels better on the display that has a lower black level?
No, it just means that the displays have different black levels.
Taking into account what you said about the lagom test patterns in the FAQ, I’m referring to the black level test on that site. On the Acer, the squares from 1-3 are almost invisible, I believe I can just BARELY see them in the dark with my face to the screen. On the Dell these were distinctly visible.
As mentioned in the FAQ, these tests are not particularly useful in a color managed environment.
Thanks for the replies. I guess my OCD will have to just live with the differences between the two displays and calibration software packages.
I’ll run a more detailed calibration using DisplayCal on the Acer and see how the results turn out.
I ran a “medium” speed calibration on the Acer XB271HU using the colormunki, and the results are similar to calibration generated at “high” speed. It seems gamma in the low end is consistent across subsequent calibrations. Though different from my previous Dell U2414H, I’ll assume that the calibration is doing its job. Its quite hard to see shadow detail in practical use however.
I have one additional question, if you don’t mind 🙂
Why does it seem that each time I run the interactive display adjustment step prior to calibration to adjust the RGB channels, I get different nominal results for R, G and B. Do displays really drift that significantly, that quickly? Can it be attributed to variability in the colorimeter?
It also seems that no matter how close I dial in the RGB channels to acheive as close to 6504K as possible, as well as matching the color coordinates, the color temp of white changes significantly post calibration. For example I dialed the RGB channels in to about 6510K pre calibration, but post calibration I’m seeing it jump to close to around 6600K. I’ve even tried using “native” and a 6500K target for white point, and both seem to alter the white point after calibration is completed. I guess they look fairly similar visually, without any way to effectively A-B them, but I’m just confused as to why this happens. Shouldn’t the 6500K target option yield a final white point thats closer to 6500K?
Thanks for your replies. I seem to be adding to the FAQ with each post 🙂
I ran a “medium” speed calibration on the Acer XB271HU using the colormunki, and the results are similar to calibration generated at “high” speed.
That’s not what I meant though. You should set an identical black level target, i.e. on your case 0.12 cd/m2 (which is the higher one of the two monitors, and the lowest that both can achieve).
Do displays really drift that significantly, that quickly? Can it be attributed to variability in the colorimeter?
It’s not really significant, but yes.
So I’m assuming on an LCD display calibrated to a gamma 2.2 curve, the few shades just above black are extremely difficult to see, or maybe can’t be seen. I’ll take this as “normal” behavior. Is there a setting in displaycal that would allow me to alter the gamma at the low end slightly to bring out more shadow detail? If I understand correctly from your previous replies, if I manually set a higher black level (for example 0.12), it may make the darker tones more visible?
The issue I’m having is that after calibration, shadow detail in photos, test patterns and games (i play a lot of games) is extremely difficult to see. It appears like there is black crush. While I’m assuming this is correct behavior based on the calibration on my particular display, I was just wondering if there is a way to remedy it in the software settings?
I did not have this issue on the Dell U2414H after calibration. Darker shades were very dark, but still distinctly visible. Is it possible that the contrast setting on the Acer is too high? I left it at default (in this case, a value of 50).
Thank you!
So I’m assuming on an LCD display calibrated to a gamma 2.2 curve, the few shades just above black are extremely difficult to see, or maybe can’t be seen. I’ll take this as “normal” behavior.
Correct. They should only be visible under ideal conditions.
Is it possible that the contrast setting on the Acer is too high? I left it at default (in this case, a value of 50).
Contrast should only be determined by the black and white level. The contrast control should be left alone in almost all cases on computer LCD monitors as it distorts the signal and either clips or lowers contrast.
Gotcha, thanks. I think I’m well equipped to create a decent calibration at this point. I ran another on Fast speed, gamma 2.2 and native color temp (dialed the RGB channels in close to 6500K with DeltaE of about 0.3) and the results were excellent.
I think I’ll play with the black output offset to try to rise out of black quicker. I was reading up on the documentation on the home page. Am I understanding correctly that 100% output offset would give the fastest rise out of black, or do I have that reversed? If so there is really no point in me tinkering with it.
Thanks again!!
Am I understanding correctly that 100% output offset would give the fastest rise out of black, or do I have that reversed?
You got that reversed, it’s the other way around. But note that this doesn’t affect color managed applications anyway, only stuff that is not color managed like the desktop.
Ok, thanks for answering all of my questions! You do such excellent work, really appreciated!
The contrast control should be left alone in almost all cases on computer LCD monitors as it distorts the signal and either clips or lowers contrast.
What about the RGB controls? Is it advisable to ignore those as well (even if the RGB bars are grossly uneven) and only adjust the brightness?
