- This topic has 20 replies, 3 voices, and was last updated 5 years ago by
.
-
Posts
-
I have a Dell U2713H monitor with 2 presets for hardware calibration and profiling, win 10, nVidia Quadro graphics card. I have i1Displaypro X-Rite colorimeter.
I want to use the largest gamut possible for photo printing and editing with Lightroom am sRGB fo internet etc. I used DCCS to calibrate and profile CAL 1 in sRGB color space using xy coordinates, 100cd/m2 2.2 g D6000 (v2). I used CAL 2 for native color space, 2.2g 100cd/m2 D5800 (v2). Then verified with display.ca. Follows the 2 firsts the links for those reports.
Then I choosed in osd menu personnal color preset and calibrate and profile in native color space, D5800, 2.2 g 100cd/m2 with Displaycal to get an icm file. Finally I did a last profile for osd preset sRGB D6000 2.2g 100cd/m2. I verified those 2 profiles. Please take a look and 3rd and 4th links here.
The 2 reports made for the CAL hardware profile indicates what seems to me bad results for “Measured vs. assumed target whitepoint ΔE*00”.
Why’s that?
Does-it matter (I supposed it those)?
What did I do wrong with DCCS? What can I do to improve the result.
Are the 2 software profiles with Displaycal better that harwares ones ( Ithought that displays with hardware cal were better choices for photogaphy even if I am only amateur level) ?
Calibrite Display Pro HL on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.What did I do wrong with DCCS? What can I do to improve the result.
Did you use “GB LED” or “RG Phosphor” display type in DUCCS?
Are the 2 software profiles with Displaycal better that harwares ones
You should skip software calibration in DIsplayCAL due to your monitor’s hardware calibration. Set tone curve on the calibration tab to “As measured” and disable interactive display adjustment. Also disable black point compensation on the profiling tab.
Did you use “GB LED” or “RG Phosphor” display type in DUCCS?
I have not such a choice . DUCCS found my display automatically and the colorimeter.
You should skip software calibration in DIsplayCAL due to your monitor’s hardware calibration. Set tone curve on the calibration tab to “As measured” and disable interactive display adjustment. Also disable black point compensation on the profiling tab.
The 2 reports made for the CAL hardware profile indicates what seems to me bad results for “Measured vs. assumed target whitepoint ΔE*00”.
Why’s that?
Because of DUCCS many faults.
There should be a log un your user folder in AppData (hidden), Local or Roaming subfolder, then another Dell folder.Calibration is (was on older versions IDNK now) computed without considering target brightness, after internal LUT are wrote DUCCS lowers brightness and this may change whitepoint.
Does-it matter (I supposed it those)?
White is not very close to “daylight” (“natural”) whites.
If you cannot live with that, correct it with DisplayCAL over CAL1/CAL2. I would be a GPU calibration limiting one or two channel maximium output by an small amount.
Quadro should not cause banding with this task.What did I do wrong with DCCS? What can I do to improve the result.
AFAIK, DUCCS should reset LUT contents before reading CAL/CAL2 uncalibrated behaviour… but just in case, factory reset monitor and try again.
Your monitor also is a model from 1st gen HW cal from Dell, so it is excluded from Dell SDK support which can be used (if you know some C++ programming) to write a DisplayCAL calibration (*.cal contents) in native gamut to CAL1/CAL2.
Complaining to Dell about DUCCS white point under performance and the lack of support from Dell SDK is another way to try to solve this.
(Your monitors LUT can be written, so it’s Dell’s fault if such functionality is not available to your U2713H in Dell SDK. It’s a Dell’s policy choice, not a monitor’s HW or firmware limitation).Are the 2 software profiles with Displaycal better that harwares ones ( Ithought that displays with hardware cal were better choices for photogaphy even if I am only amateur level) ?
Do you see banding while inspecting a perfectly smooth gradient in a non color managed enviroment (MS paint, even IE) while using ”
PERSO 2.2 2019-04-04 100cdm² D5800 XYZLUT+MTX” ? Does it bother you? If you do this is a good reason to use HW cal (or a GPU capable of doing graphics card calibration without banding)Anyway, “sRGB factory mode” is too far away from your desired calibration target (not from its target, IDNK that), so when you calibrate it using DisplayCAL you ended with 87% unique grey levels (4-th report). I would say that it is not very good. For that target it would be better to use DUCCS to calibrate CAL1 then “fine tune” white point in graphics card with DisplayCAL if getting that white point is importnat for you.
Did you use “GB LED” or “RG Phosphor” display type in DUCCS?
I have not such a choice . DUCCS found my display automatically and the colorimeter.
PasProfile is right, no EDR choice, but auto selected EDR choice by DUCCS should be visible in “profile patch size” choice screen … o maybe in the screen with a “yellow”-like information screen in the left before you click for starting HW cal measurements. I do not remember the exact screen.
Check out if it says “GB-LED” or something like that.I forgot something important, after DUCCS has done its job, it is going to try to use its LUT loader (or Windows loader) to load a null/identity calibration into GPU (output=input).
In latest windows 10 versions this will cause GPU LUT contents truncation to ~8bit… so if you load a DisplayCAL calibration for whatever OSD mode it is very likely to cause banding.
You need to reboot after DUCCS calibration and maybe ensure that Xrite gamma loader does not start.So many information to process. I’ll come back soon. Thanks you all in the meantime. One thing certain, I have no skill for C++ programming, when I began photography, it was with a Minolta SRT101, black and white argentique!! and my own dark room with chem baths!
Because of DUCCS many faults.
There should be a log un your user folder in AppData (hidden), Local or Roaming subfolder, then another Dell folder.Is it the file you speak of?
AFAIK, DUCCS should reset LUT contents before reading CAL/CAL2 uncalibrated behaviour… but just in case, factory reset monitor and try again
OK Vincent, I’ll try and report
Do you see banding while inspecting a perfectly smooth gradient in a non color managed enviroment (MS paint, even IE) while using ”
PERSO 2.2 2019-04-04 100cdm² D5800 XYZLUT+MTX” ? Does it bother you? If you do this is a good reason to use HW cal (or a GPU capable of doing graphics card calibration without banding)Yes I see banding. My Nividia quadro can isn’t it?
Anyway, “sRGB factory mode” is too far away from your desired calibration target (not from its target, IDNK that), so when you calibrate it using DisplayCAL you ended with 87% unique grey levels (4-th report). I would say that it is not very good. For that target it would be better to use DUCCS to calibrate CAL1 then “fine tune” white point in graphics card with DisplayCAL if getting that white point is importnat for you.
To achieve this, I open Displaycal, skip calibration, select .current. in setting and run a profile isn’t it??
I forgot something important, after DUCCS has done its job, it is going to try to use its LUT loader (or Windows loader) to load a null/identity calibration into GPU (output=input).
In latest windows 10 versions this will cause GPU LUT contents truncation to ~8bit… so if you load a DisplayCAL calibration for whatever OSD mode it is very likely to cause banding.
You need to reboot after DUCCS calibration and maybe ensure that Xrite gamma loader does not start.I think that you find the reason of my banding . How do I make sure that Xrite gamma loader does not start after rebooting?
Log:
There should be several files. One with connection data sent or read from USB & DDC /CI, another one (maybe mixed) with xyY measurements.
As you will see, after calibration is written, DUCCS try to fix white luminance from default 50% OSD value (~200cdm2?) to your desired target. It can cause some issues… but it is the way DUCCS works (at least not too old versions)Banding:
Several sources for this issue. If you let Windows 10 or other loaders to play with GPU LUT it is very likely that contents will be truncated to 8bit until you reboot (so there is banding when you load a not null/identity calibration). That includes apps like i1Profiler, OEM variants of i1Profiler like Dell, Viewsonic… also Basiccolor Display, Microsoft Night mode for Windows 10, or even monitor going to stand by and then awaking: banding until you reboot.There is another source of banding caused by GPU driver or limited LUT precision like intel iGPUs or some older nvidias.
If your banding is caused by the first source, then use task manager to disable those LUT loaders in “Startup” tab and if it does not show, use a known & secure tool for that like “Autoruns” (Sysinternals in Microsoft site) and uncheck “XRgamma.lnk” in startup/logon. Disabling it won’t mess or conflict with DUCCS, it does not need it.
Anyway… after a DUCCS calibration you will need to reboot if you want to load “properly” a not null/identity GPU calibration (no matter what is your GPU vendor),.It seems that something has gone wrong in latest Win10 versions regarding GPU LUT content/loading.
DisplayCAL in CAL1/CAL2 after DUCCS:
Do as you did in DisplayCAL for “sRGB factory mode”,
-choose desired white (white luminance: native)
-choose desired gamma
-choose desired profile type (typical curves+matrix/curve +matrix for this kind of well behaved displays)
-Click in calibrate & profile.
Since CAL1/CAL2 has locked RGB gain controls, when popup shows asking to tweak white, just click in start calibration.
The key difference is that you cannot modify factory sRGB mode white point (or gamma)… but you can modify white in DUCCS making this white closer to your desired target.HI, I want to check banding. I have a psd file smooth grey gradient opened with MS viewer (not paint) and I see bandings.
1-In task manage, I found CalibrationLoader 5.0 application from Logo Kommunications. It is stored with Dell UCCS files. It is a XRgamma.ink file. So you think that I can desable it without messing with my pc?
2- In Nvidia Quadro K600 I have a choice for (colors settings) between “by default” or “Nvidia colors settings “RGB” 10bits, do I choose the second one?
HI, I want to check banding. I have a psd file smooth grey gradient opened with MS viewer (not paint) and I see bandings.
- Banding in color managed applications (even partially like old MS viewer) is different from calibration banding. That is why you should use MS paint (or equivalent app without any kind of color management).
- Sometimes those smooth gradients PSD/PNG are not smooth, or are 16bit gradients. Use a smooth 8bit gradient for testing calibration banding.
If you open this PNG in MS Paint it should not have banding:
http://www.lagom.nl/lcd-test/img/gradient-h.png
You can user other gradients if they are smooth and even post them here for checking. - Banding = “cuts”, “steps” with a well defined border.
If you see “smooth” pink or green coloration in grey in such gradients it is not banding, it is neutral grey error in calibration. There are verly likely caused by too few measurements of uncalibrated behavior by calibration app (like DUCCS).
In task manage, I found CalibrationLoader 5.0 application from Logo Kommunications. It is stored with Dell UCCS files. It is a XRgamma.ink file. So you think that I can desable it without messing with my pc?
Yes, no problem disabling it. Use DisplayCAL trap app to switch between ICCs when you switch from one OSD to another.
IDNK if there is conflict with that app and DisplayCAL tray loader… but dispcal tray app is better so there is no need to run that XR gamma at startup/logon.
