- This topic has 17 replies, 5 voices, and was last updated 3 weeks ago by
.
-
Posts
-
Greetings,
In the same boat here but for a Samsung G60SF. Colors are either way off or I have been seeing colors wrong my whole life.
After reading quite a few of the other threads I somehow got the idea in my head that the NEC MDSVSENSOR3 would be able to fix my issues, or at least help. If I am wrong please point me on the path and I will send this back when it arrives.
If that will work do I also need NEC SpectraViewII software?
TIA
Greetings,
In the same boat here but for a Samsung G60SF. Colors are either way off or I have been seeing colors wrong my whole life.
Define wrong. 99% likely that what you are seeing is not caused by “bad calibration” but due to lack of color management in the software you used in your computer.
If your previous displays were sRGB-like and contnet you play on your computer was sRGB/Rec709 transformations done by color managament would be output = input (with minor gamma correction), so in a “no color management envrioment” you got colors close to expected.
Now woth a widegamut QD-OLED this does not apply.So the first step is learing how to use color management and apps with color management support
OR
system wide sRGB simulation (inside monitor by using sRGB presets, inside GPU: novideo_sRGB, AMD driver, DWMLUT)Remember that using any kind of sRGB simulation systemwide (2nd alternative), display ICC profile in OS control panel should be whatever you are simulating (sRGB in this example). If you keep display ICC installed with driver with native gamut all color managed apps will desaturate in a wrong way, because you are “lying to OS” about display behavior.
After reading quite a few of the other threads I somehow got the idea in my head that the NEC MDSVSENSOR3 would be able to fix my issues, or at least help. If I am wrong please point me on the path and I will send this back when it arrives.
You can use this i1d3-variant to create ICC profiles with DisplayCAL… but you’ll need how to use that ICC profile (color management, my 1st reply)
If you want to use it for gaming or something like that 1st step would be “simulating sRGB” systemwide. Learn color management later.
If that will work do I also need NEC SpectraViewII software?
No, that software is for NEC monitors, to write calibration inside this manufacturer’s monitors. Not yours
TIA
Calibrite Display Pro HL on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.“If you want to use it for gaming or something like that 1st step would be “simulating sRGB” systemwide. Learn color management later.”
This is what I am looking for. I still have the calibration device arriving this week.
Any chance you have a link diving into simulating systemwide RGB?
It truly might just be me not being used to the brightness/contrast. I just feel like the blue and green is off and there is like a yellow/green haze over whites and blues.
This was out of the box.
“If you want to use it for gaming or something like that 1st step would be “simulating sRGB” systemwide. Learn color management later.”
This is what I am looking for. I still have the calibration device arriving this week.
Any chance you have a link diving into simulating systemwide RGB?
As was witten before:
(in GPU)
-novideo_sRGB (nvidia), google for its github
-AMD driver (amd cards), lots of manuals, google AMD driver sRGB simulation.
-DWMLUT (all cards), full LUT3D running on GPU shaders. LUT3D is created with displaycal and this implies learning about ICC profiles, etc. DWMLUT is not allowed by some games because it could be used for cheating (rising shadows so you can see enemies too early)(in monitor)
read your particular display model user manual. Usually it has locked whitepoint. Sometime even brightness and that’s a no go.It truly might just be me not being used to the brightness/contrast. I just feel like the blue and green is off and there is like a yellow/green haze over whites and blues.
Usually contrast OSD, leave at factory settings.
Brightness lower till you find it confortableThis was out of the box.
Alright so I have been playing with it a little bit over the past week. The only corrections I have seen available that would possibly be applicable are the XG27AQDPG (which has anti-reflection) and the AW2725DF (which I am pretty sure also has anti-reflection).
- The G60SF has a different matte like anti glare.
I have run it with those profile and I have run it with auto.
- Settings:
- Display and Instrument:
- White level drift compensation: Checked
- Instrument: I1 Displaypro
- Mode: LCD(Generic), however the AW2725DF does profile to Refresh (generic)
- Calibration
- Interactive Display Adjustment: Checked
- White point: 6500
- White levels: between 120-150
- Black level: as measured
- Observer: CIE 2012 2
- Tone Curve: sRGB
- Black output offset: 100%
- Ambient light level adjustment : Box not checked
- Black point correction: Box not checked
- Calibration speed: High
- Profiling:
- Profile Type: Curves + Matrix
- Black point compensation: Unchecked
- Profile Quality: High
- Testchart: Auto-Optimized and default Testchart. Increasing neutral patches to 256 when able.
- Patch Sequence: Minimize display response delay
I set the monitor to Original, which is in Warm 1, that puts it relatively close in adjustment (Standard results in blue hitting the -40 range).
Color Space: Auto
Nvidia control Panel:
-
<li style=”list-style-type: none;”>
- Color Depth: 32 bit
- Output color format : RGB
- Output Color Depth: 8 bpc
- Output Dynamic Range: Full
Novideo sRGB:
-
<li style=”list-style-type: none;”>
- Use ICC profile (whichever ICC profile is being tested)
- Dithering: Enable
- Mode: SpatialDynamic2x2
- Bits: 8 bit
I feel like I am forgetting stuff and I probably am. It seems to be close but I am still hitting a noticeable max delta (I believe) close to or over 3… it is not to noticeable until it hits on the right object.
What can I do better here?
If you have a different monitor/laptop screen for which it’s easier to get a good spectral correction:
- Display a white/grey patch on the second monitor that’s very close to D65 (after measuring).
- Display a white/grey patch of similar brightness on your QD-OLED.
- Adjust the QD-OLED (OSD, GPU settings, …) until you see the same grey on the QD-OLED as on the second display.
- Measure the QD-OLED (using the closest correction you can find) and note the x,y coordinates of your new grey reference.
- Calibrate the greyscale/gamma using the new reference and generate a matrix profile.
You should get good greyscale/gamma. Colors maybe not perfect (not ideal correction, metamerism, probably non-linearities) but enough to correct for major gamut differences.
Thank for the reply. I will work on that right now. The only other monitors I have are IPS and TN. I’ll try it on the IPS.
Would you mind taking a look at the results from these calibrations please? Am I being over critical for my intended purpose?
DO NO use CCMX that you did not make. Only CCSS are portable.If you make your own CCMX with an spectro = fine, but DO NOT use other’s.
Do not use i1d3 uncorrected, excluding trying to figure where primaries are.Also your sRGB simulation is under saturated, so it looks that you are applying twice, monitor + gpu, twice on gpu … you cannot use AMD driver / novideo sRGB and validate against display profile whith whatever color managed simulated profile.
IF you simulate something on GPU or monitor, THAT you simulate is display profile. But DisplayCAL doe snot make it easy to simulate exact match to some standard profile and also applying VCGT calibration.I’m sorry, I am so confused here. No clue on the double GPU. GPU is rtx4090, could it be picking up the motherboard video? I’m not plugged into it, but I guess it could cross somehow. It runs the profile. I save it and don’t load it in novideo, clamp and then run the verification.
The only setting in Nvidia I could see interfering would be override to reference mode, but that is not checked.
So do not use the i1d3 uncorrected and then the XG27AQDPG profile was CCMX which was my fault I should have grabbed the CCSS. I am guessing Auto* is not considered a correction? Soooo, then I would use the XG27AQDPG CCSS file instead of the CCMX?
I’m trying here.
If you are using novideo_sRGB, you are simulating sRGB or other generic colorspace. “Your display” (display + GPU sRGB simulation by novideo_sRGB) is sRGB now.
Hence you cannot use DisplayCAL color managed validation where you say (by configuration) that your display is “whatver profile you made in its widegamut presets” and you want to simulate “color managed” (like Photoshop will do) how it will render an sRGB image. => This is what you did in your HTML report: you enable sRGB simulation on GPU and keeping widgeamut profile as display profile you asked displaycal “how will photoshop render an sRGB image”, hence the under saturation you see in lower a*b* plot.
You are applying it twice.When you simulate something on monitor (by preset) or in GPU your display profile and the colorspace you want to simulate become the same.
This is usually tested by “use simulation profile as display profile”, but this will?/would? undo VCGT grey calibration if using novideoSRGB or AMD driver sRGB simulation. On DWMLUT should be no issue.An alternative for this will be using HCFR (beware disabling VCGT on config), but report is not as detailed as Displaycal regarding a*b* grey range.
Another bypass would be using another DisplayCAL in a virtual machine and validate with “use simulation profile as display profile” in virtaul machine, so VCGT won’t be cleared.
AFAIK DisplayCAL does not allow right now to verify “an exact match to an arbitrary generic profile” (use simulation profile + use simulation profile as display profile) and at the same time keep VCGT calibration.So having the monitor in Auto rather than native and the Nvidia control panel as 8 bpc, Full, RGB. Is contradicting? Or acting one in the same?
Is there any way to only apply it once?
Real quick too, DWMLUT is this going to get me torched in game for shading issues? Not trying to cheat, but the anti-cheat software does not understand that.
The contradction is your DisplayCAL validation report, the way you configured it as explained in my previous message, as the suggested alternatives.
I understand it a little bit now. The use Nvidia setting I had place were simulating RGB and novideosRGB was simulating as well. Then my monitor was simulating is too? Since it was in Auto rather than Native?
I might be way off here. If you point me to it I will learn it. Is there a masterclass or book on this?
Vincent,
What is my best option here? What if I used two separate systems to do this? Or use the second one to just measure the calibration?
I read what you write, I promise…I just am not knowledgeable in this. If I ran the calibration on my system and then calibrated using my other system (4080 vs 4090). Would that be okay?
I’m sorry man, I am just in it to deep now to let go.
When you simulate something on monitor (by preset) or in GPU your display profile and the colorspace you want to simulate become the same.
This is usually tested by “use simulation profile as display profile”, but this will?/would? undo VCGT grey calibration if using novideoSRGB or AMD driver sRGB simulation. On DWMLUT should be no issue.An alternative for this will be using HCFR (beware disabling VCGT on config), but report is not as detailed as Displaycal regarding a*b* grey range.
Another bypass would be using another DisplayCAL in a virtual machine and validate with “use simulation profile as display profile” in virtaul machine, so VCGT won’t be cleared.
AFAIK DisplayCAL does not allow right now to verify “an exact match to an arbitrary generic profile” (use simulation profile + use simulation profile as display profile) and at the same time keep VCGT calibration.This, as explained before. With the cautions about VCGT not preserved.
