- This topic has 31 replies, 4 voices, and was last updated 1 month ago by
.
-
Posts
-
Hello everyone,
New on the forum, thank you !
I recently bought a Dell S2722QC for my MacBook Pro 13 inch M1 2020 and my PC (the first through USB-C and the second with HDMI). I have a datacolor SpyderX Pro calibration tool and I tried to calibrate both displays (MacBook and Dell) with it.
I first started with the included SpyderX software and then did a calibration with DisplayCal. The aim is to color match the two displays, both on macOS and on my PC.I tried several settings but struggle to get consistent results on macOS and Windows.
The best I could manage with DisplayCal on the MacBook at 120 cd/m2 is : deltaE avg: 0.36, max 1.35 100% SRGB 85% adobeRGB
I think it’s great for a laptop screen and especially with an entry-level MacBook.
On the other hand, when I try to calibrate the Dell monitor, I can either repeatedly get an average deltaE of 0.8 and max of around 3, with 98% sRGB coverage and 80% adobeRGB, or one time (that I’ve not been able to reproduce) deltaE of 0.52, max 2.0 and 99% sRGB/83% AdobeRGB. The Measurement report is attached. It’s not bad actually, but the blues are way off when I do a verification against sRGB.
I tried to do the calibration again with the same settings or others, but get worse results everytime.
Could you please help me understand why, and recommend me settings for the best calibration ?
Thank you !
SpyderX Pro on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.Spyders are known to be bad. If you can, get an i1d3 instead (I recommend the calibrite colorchecker display plus)
A few novice mistakes. 2.2 absolute gamma will crush blacks on a LED panel. You choose your offset and Relative 2.2 gamma, so it is black compensated. It is pretty normal for a consumer display.
Unknow if your display type correction is white led. Yours is not HDR. Check your blue hue visually with a blue filter screen. Maybe the blue hue is off. It could be the sensor and correction though. Your gamut picture shows you reach targets for all colors and that is great. Blue is just off slightly in high end with its hue. It might be corrected and make the rest wrong so it’s good to check.
Anyone else can help. I wrote a different message before, and it got erased. Hope this helps.
What does the simulation profile none mean in the report? It is compareing to a straight 2.2 curve but the display is calibrated to black compensated it looks.
Your report is not using a simulated colorspace like Rec 709. It is comparing your display color space to its native colorspace. You need a checkbox checked there. Use simulation profile as display profile should not be checked to measure your profile to Rec 709 or a colorspace. Checked measures your hardware whitebalance and color management to the colorspace. A profile will adjust your hue in color managed apps. So your report is just wrong.
I recalibrated and rerun the verification.
The results are attached for the details.
So, average dE : 0.6, max 2.46 (still the blue, but much better than before), WP 6580K
I’ll leave it to that, I’m very satisfied with the calibration. I guess I made something wrong before to have so much deviation for the blues…
Thank for the great advice !
Unknow if your display type correction is white led. Yours is not HDR. Check your blue hue visually with a blue filter screen. Maybe the blue hue is off. It could be the sensor and correction though. Your gamut picture shows you reach targets for all colors and that is great. Blue is just off slightly in high end with its hue. It might be corrected and make the rest wrong so it’s good to check.
He choose the right built in mode.
What does the simulation profile none mean in the report? It is compareing to a straight 2.2 curve but the display is calibrated to black compensated it looks.
It is comparing to whatever ICC profile says. Profile is created after calibration so it does not care about your nominal gamma target, just about whatever gamma it got (ARgyll priorizes grey neutral) and since it was built in macOS it’s likely that it sores actuall TRC till it got near black where ICC would asume infinite contrast.
I recalibrated and rerun the verification.
The results are attached for the details.
So, average dE : 0.6, max 2.46 (still the blue, but much better than before), WP 6580K
I’ll leave it to that, I’m very satisfied with the calibration. I guess I made something wrong before to have so much deviation for the blues…
Thank for the great advice !
That is a wrong setting, stick to your first one.
If (by chance) your SpyderX was accurate and its buit in mode for “all wled PFS” and “all white led” is accurate, whitepoint is likely to be off from those coordinate values.Regarding blue on both measurement report it is caused because measured blue and predicted blue by ICC profile do not match.
Predicted blue ramp by ICC profile is made with ti3 mesurements and then simplified to fix in the extra simplified ICC profile type that macOS can handle.
So first task is to find if those predicted values were calculated from wrong data or if they were oversimplified from after callibration measurement data, a “best fit” model from a highly irregular 3D mesh which doe snot seem to be the case.Also calibrated display shows a mid grey color error issue, try lowering calibration speed to low for full 12-24-46-96 grey ramp measurements, but it may take long time, IDNK Spyder X speed.
Vincent is right about the color profile. 87% P3 is not White Led. It is HDR rec 2020 almost.
This Mac OS stuff is crazy to me. Mac has a different ICC profile. Funny I see ICMs everywhere in windows and all RGB. RGB is REC 709 but I get a different report on the checked and unchecked simulation profile. The simulated box is to see what color managed apps see?
OK. I understand that I got good results by chance and not because I selected the right settings ?
So, i should rerun calibration with LCD PFS Phosphor WLED instead of White-LED, 6500 K white point, black level as measured, Gamma 2.2, set Single Curve+Matrix, as macOS doesn’t support LUT corrections and lower the calibration speed to low.
Is that correct ?
Then which testchart should I use in the profiling tab ? 34 patches ? 54 ? More ?
It is a high gamut display and White Led is wrong. The link on the bottom helps on the displaycal main page.
Yes. I can not use more than 54 patches in Single Curve+Matrix. Not sure it makes a difference.
It is very undocumented to me. Your screen was showing 1.6 to 2.2 gamma in the lower dark shadows range of white. Your 1% black would be elavated compared 2.2 at infinite contrast- a screen with 0 brightness at 0 level white. Black Compensated takes it in to account. Your result matches 1.6 gamma matches Black Compensated gamma better. It will look better. 1 of 255 should just be visible or maybe invisible in a lighted room. http://www.lagom.nl/lcd-test/black.php It should be even upgraded but hardware does not get that right only the 1d lut tone curve from a profile. Display Cal Help will help some and research. AVS forum is good, Some sub redits on Reddit are good. I browse the colorists sub reddit. Reddit almost took over forums and websites.
I prefer 3 curves + matrix . I should try 1 curve + matrix some day and see what is better. 1 curve means your RGB white balance is right and all 3 rgb are exactly the same difference off. My blue is way of in 1%. .016 for RG and B is .3 . This is only Calibrite colorchecker plus meter reading. I think Spyder X is fine and tests proven it very close to Xrites meters. The dies in xrites last a long time and datacolors meters had worse protection for the die that reads them. And xrites meters have better user calibrations and for future display techs.
Use D65, not 6500K. Gamma 2.2 is for bright environments, BT.1886 or 2.4 for dim. Sorry I can’t help more. Busy ATM
eye opener. I just looked at the 34 patch test chart. It is the editor function in the profileing section. It only adjusts the grey scale basicly. Profileing checks 100 red, green and blue. Amazing it works. It does work though and can see the measurment difference from built in icc profiles like srgb and the user made ones. It is very complicated.
