- This topic has 11 replies, 2 voices, and was last updated 4 years, 6 months ago by
.
-
Posts
-
I’m trying to calibrate my ViewSonic VP2768-4K monitor and am second guessing everything. It’s supposedly pre-calibrated, but I’m not confident in the accuracy. I have the XRite i1 Display Pro. I do video production, so I need the monitor to be as color accurate as possible for grading footage. Really struggling with finding the proper settings. No matter what, when I try to verify the accuracy of a profile, it comes up way off/unsuitable. I’m just all confused now as the settings for both processes are probably off. Can anyone shed any light on the proper settings? Any help would be greatly appreciated!
Here’s what I currently have:
Calibrite Display Pro HL on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.No matter what, when I try to verify the accuracy of a profile, it comes up way off/unsuitable. I’m just all confused now as the settings for both processes are probably off. Can anyone shed any light on the proper settings? Any help would be greatly appreciated!
You are not verifying the accuracy of the profile, if you wish to do that you should uncheck simulation profile.
I do not remember the exact settings but with that configuration in verification setting you are validating display colorspace & native gamma/grey (+ optional white point tunning in OSD) against your target colorspace and gamma.While your display is sRGB-like, its native colorspace is not Rec709/sRGB, its a little bugger in some locations, hence if you use your settings ther will be some errors. Also native display gamma won’t be rec1886 in a 1000:1 display (…because it’s useless) but close to 2.2 unless your display OSD has some 2.4 gamma preset and its accurate.
You need to know that GPU calibration will fix white point, grey color and gamma… but it won’t map the display’s slightly bigger native gamut to rec709. If you need to do that you’ll need a HW calibration with a lut-matrix-lut/LUT3D or an application that supports software LUT3D like Resolve or an application that uses color management with ICC profiles (which maybe your situation if you use some video editing tools under macOS).
If you explain which editing tool you use people can help you in better way.AFAIK that display should have Viewsonic Colorbration support but I’m not sure if display has a lut-matrix-lut for reducing colorspace to Rec709 or itf it’s just a 1D LUT like GPU but internal to monitor.
So my recomendation:
-Do not use Rec1886, it is useless with a 1000:1 display. Read what Rec1886 menas, there should be ITU-R papers available for free. Use 2.4.
-Try Colorbration first if supported (Viewsonic & Dell have proper specatral corrections for their displays, or almost all their displays, Benq doesn’t).
Then validate results with DisplayCAL. Grey color errors or gamma errors or small white point deviations could be corrected in GPU with an additional DisplayCAL calibration. If Colorbration calibration is OK, no need to apply DisplayCAL calibration but it’s the best tool to verify Colorbration results.
If your display has some kind of gamut emulation capabilities in its HW calibration (Colorbration tool from Viewsonic) then you can use your current verification settings. Otherwise you should not enable simulationprofile.-I’m against calibration speed = “high”, unless you have a super visually neutral grey, but YMMV.
Hey Vincent,
Thanks for the info and suggestions! I use FCPX, but also use Resolve for a few things (basic color grades being one of them). So, I tried a number of calibrations with Colorbration (sRGB and rec709, 120 brightness, D65 white point, 2.2 gamma and 2.4 gamma) and no matter what, when I verify it the test comes back as “failed” in Colorbration and DisplayCal. I’ve attached the fail for my sRGB calibration from Colorbration. I’ve unchecked Simulation Profile and just click measure. It just seems like a few things are always a bit off. Any thoughts?
This reports shows huge errors in a* for grey. Are you sure that you are using OSD mode where colorbration stored HW calibration?
It seems abnormaly high. We know that those “lowcost” HW calibration solutions can have some issues in grey, perhaps 1.x in a* but…3??? It looks like “uncalibrated”.
Please make sure that each time you validate, OSD mode/preset is the use you used for calibration, no matter it is for Colorbration or DisplayCALFor example, you made a Rec709 2.4 colorbration and stored it in “CAL1” or an OSD preset named like that. Colorbration made a profile. Such profile is valid ONLY if you set active preset to that CAL1.
If you made later a DisplayCAL calibration for “standard” OSD preset, that displayCAL profile is only valid for THAT preset.
If you change OSD mode you have to manually change default display profile and maybe restart application.
If you wish to validate Colrobration profile, monitor shoudl be in CAL1 and default display profile should be colorbration’s one (use ICC version 2 if requested)Also make sure that all auto dimming features are disabled in your computer and display.
Note: when you use Colorbration it is possible that due to that software limitations & over simplifications “asumed vs measured whitepoint” is “Not OK”. DO not worry too much about it, it can be corrected later in DisplayCAL if you wish to, but a*b* values for grey should be “grey” , something like a* = 0.x, 1.3 at most… not 3.x.
Thank you! I made sure that the Display Profile I made aligns with the HW Cal1 that Colorbration created. I wasn’t sure if I needed both or if the Cal1 profile negated the need for a Mac Settings Display Profile, but I’m using the Cal1 profile and the ICC profile together in the above test. Even if the monitor is uncalibrated, this should correct it back, right?
With Colorbration, since it’s the Viewsonic version of i1Profiler, there are only options for RGB Primaries (set to Rec. 709), Luminance (120), then profile settings (Bradford, Version 4, and Matrix based respectively).
So in DisplayCal, in the Calibration settings, for Tone Curve I want Rec 709, right? Do I leave Black Output Offset at 100%? I’m using the Screen Correction – Spectral: LED White LED Family (AC, LG, Samsung), perhaps that’s not correct?
Progress! (?) – I randomly tried these settings and it got me seemingly very close to being ok:
Display and Instrument – Correction: Auto
Calibration – Whitepoint: 6500, White Level: 120, Black Level: As measured (should this be something specific?), Tone Curve: Rec. 709 with Black Output Offset at 100%.
The spectrum coverage is alright, but the best I’ve gotten it so far. Results are better overall, but have a few massive bad ones in the shadows.
Spectral: LED White LED Family (AC, LG, Samsung), => yes, this one. Use it. It means common sRGB-like LED display, 99% of displays and no P3/HDR LED TVs
Grey is grey now (a*=0.X), but do not use Rec709 TRC for display. It does not mean what you think. Same for Colorbration Rec709, or sRGB preset in Colorbration. Use gamma 2.4 or 2.2 respectively o those two setups.
Default settings (which I think that include black output 100% at calibration), medium/slow speed to get that grey, WhiteLED correction for common displays, D65 white.
You can discuss about gamma 2.2 or 2.4 but for Rec709 video 2.4 is the most common choice although some people argue about 2.2 for web deliveries. Put 2.4 and if OSD in Viewsonic allows 2.4 preset use it.
Since you use macOS for common desktop work you should use default profile settings (single curve+matrix + black point compensation), because macOS desktop color engine has too many faults and does not work properly with profiles that include 3 independent TRC.
For making a software LUT3D for Resolve you can make a more accurate XYZLUT profile, but use it just for that, for making a LUT3D. After that, default display profile should remain in the simple profile made with default settings for profile in macOS.Thank you very much! I think I got it figured! Gamma 2.4. The lower end varies a bit (1.55 is the biggest variance, but it is on black). That’s not a problem, right?
Attached a photo and the full PDF of the results. That black point variance is ok, right?
You cannot correct black in GPU LUT calibration unless you sacrifice a lot of contrast (and no one is going to do that).
Gamma in black are going to be lower that your target value because contrast is finite and you want to keep some separation between darkest greys.You didn’t choose gamma 2.4, if you want gamma 2.4 re do it again. Do not choose Rec1886, as explaned before it’s useless in a 1000:1 display.
These were my settings in DisplayCal of the verification results I sent in the last post. I have the ViewSonic set to Gamma 2.4 as well. Thoughts?
