Best possible settings

[Franilustrador]

I’ve been using DisplayCal for more than a year. I have a Dell U2417H monitor, and an i1 Display Pro calibrator, on OS Manjaro Linux. I have always proceeded in the same way for the calibration and profiling of my monitor, but I do not know if I do it in the correct way:

1.- I start width the factory settings of my monitor. The factory settings are: brightness 75, brightness 75. Standard mode. Here I have great doubts: Is it understood that the Sstandar Mode is the default setting? Or should I go to the manual setting of the red, green and blue channels, and set them all to 100%?

2.- In Displaycal, spectral correction, obtained from the program itself (Auto) .Whithe point 6500. White level 120 cd / m2. Black level: As measured. Gamma 2.2. Calibration speed: Low. Profiling type: Curves + Matrix. Profile quality: High. No black compensation. Ammount of Patches: between 500 and 800.

I am an illustrator, and my work always aims to print on paper. Precision in color is very important.

Thanks in advance.

Calibrite Display Pro HL on Amazon  
Disclosure: As an Amazon Associate I earn from qualifying purchases.

[Vincent]

1-Use “Custom Color”/”User” mode. Naming varies between manufacturers. This way you cand modify RGB gains. In other oSD modes without accesing RGB gains white point will be corrected using GPU and you may not want that.

2-“WLEDFamily_07Feb11” for LED sRGB screens and it looks that you own one of these.
Calibration speed = number of grey patches used for calibration. Slower means more grey ramp and smaller grey ramp jumps iterations until error is less than some value. “Low” aims for best grey so it is OK.
Since you are an illustrator banding may be a headache. Check out after calibration banding in non color managed enviroment (choose a SMOOTH gradient and open it in MS Paint or similar simple non color managed app for Linux)… if it shows banding check some tips in some threads of this subforum buy may be related to graphics card HW (unsolvable unless you get a new one with high bitdepth LUTs and dithered outputs).
If you wish to minimize banding in color managed programs like Krita/GIMP (which is not the same as previous one and could be caused by software, unsolvable) it would be helpful to choose an extremely idealized (but maybe not so accurate) profile like single curve + matrix. You do not need more than 100 profiling patched for it. That kind of profile should minimize rounding errors caused by color management.
So there is a balance:
-less banding in color managed apps => single curve + matrix
-more accurate => XYZLUT+matrix with a few hundreds of patches but you may experience banding in gradients even with a perfectly smooth calibration (no banding in no color managed enviroments)

Your IPS screen should validate after calibration against its own ICM profile with calibration with low dE error, it should be a well behaved screen with reasonably 9x% srGB coverage, so my 1st try will be to use a single curve + matrix profile, then validate and test gradients in your editing software. If all is OK, you are done.

[Franilustrador]

First of all, thanks for your answer, and after reading it, I opted again for a profile from XYZLUT matrix, with almost 600 patches. As you told me, the bands on the gradients may be a problem, but my work in illustration is similar to a traditional oil paint, with visible brushstrokes, so supersoft and perfect transitions are not a problem for me. Validation of the profile is good, with delta errors E always lower than 2, sRGB 99.9%, 6492ºK, etc … so I think everything is fine. I am more interested in good color accuracy.

However, I still have some doubts.

1.- Are the adjustments prior to the calibration of the monitor important? As I mentioned, the factory settings, in my case are the following: Brightness 75, Contrast 75, and Standard Mode. For the last calibrations, before DisplayCal does its work, I have started by adjusting the Red, Blue and Green channels to 100% each, and then readjust them for the desired white point. I do not know if it is more convenient to start the calibration from the Standard mode, or if this changes something the final result.

2.- I used the correction that advised me, “LCD White LED IPS (WLED AC LG Samsung) <WLEDFamily_07Feb11.ccss>”. I think it gives better results, observing the numbers, than the one I used until now, from the public database, which took into account my monitor and my colorimeter. Although it is not certain if a correction or another is better …

3.- In some forums, I have observed that with the type of profile I am using, (XYZLUT matrix), they use Black point Compensation. It is convenient? Should I consider it?

[Vincent]

1.- Are the adjustments prior to the calibration of the monitor important? As I mentioned, the factory settings, in my case are the following: Brightness 75, Contrast 75, and Standard Mode. For the last calibrations, before DisplayCal does its work, I have started by adjusting the Red, Blue and Green channels to 100% each, and then readjust them for the desired white point. I do not know if it is more convenient to start the calibration from the Standard mode, or if this changes something the final result.

AFAIK “Standard” mode should have RGB gains locked in Dells unless you change to “Custom color” OSD mode… but if youcan change RGB gains while in Standard mode, it’s fine. It is good to have that fucntionality in more OSD modes than 1.

Costrast= keep factory setting,
brightness & RGB gains use them untill you get want you want.

3.- In some forums, I have observed that with the type of profile I am using, (XYZLUT matrix), they use Black point Compensation (BPC). It is convenient? Should I consider it?

BPC=ON is like to store in profile “perfect black” even it is not true. As with single curve + matrix, it’s about idealizations when you use limited precision computing, like when doing color management.
If you do not see strange & disgusting color cast in black or near black (I mean, softproof OFF), you are fine.

It’s the same as banding: if you have issues, you choose what you need to prioritize, minimize strange colro artifacts or use accuarte profiles.
The better behaved  a display is, the more an idelaized matrix+single curve matches that display… so using XYZLUT matrix is no longer useful (weighted against its posibble issues)  for display profiling.

• This reply was modified 4 years, 11 months ago by Vincent.

[Author]

Posts