- This topic has 6 replies, 2 voices, and was last updated 1 week, 3 days ago by
.
-
Posts
-
Hello! I’ve been trying to calibrate my Samsung G95SC for over a month now. Generating 3DLuts and ICC profiles. I’ve used hard clipping, roll off, all different gamma settings, the picture comes out severely crushed or blue/red tinted (depending on the observer). Nothing seems to be working. I’m mainly trying to do HDR but SDR doesn’t even come out right. Usually the profiles are worse then the original. I’m using a matrix file for the G93SC because I’ve heard they have the same panel. If someone does have a correct matrix profile or any help id really appreciate it!
Hardware:
Ted’s PiGen 4 on a raspberry pi 4
Calibrite Display Plus HL
No spectrophotometer (idk if this would make a huge difference. I’m sure for the matrix it would maybe?)
Attached are a few images of measurements / how the profiles look
Calibrite Display Plus HL on Amazon Calibrite Display SL on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.First of all CCMX correctiosn are not portable between different i1d3 colorimeters, so find a CCSS correction suitable for your blacklight type, whatever it is QLED, QDOLED… whatever it is.
Once you find a suitable one use display RGB gain controsl to fix whitepoint. If it does not look white, then use a visual whitepoint editor approach to get a “visually pleasing white” or a match to a reference white display. If you do this approac, all LUT3D must be relative colorimetric, not absolute (there is an alt version using alt whitepoint std profiles but try easiest first).Also it seems that you do not use verification reprot properly, you are making comparisons between madVR profile and madVR output… which cannot match if you have a LUT3D applied (sine that is the intention of a LUT3D). When you have a LUT3D applied to a display/pipeline/whatever you make a direct comparison to whatever you are trying to simulate., not to display profile.
Ohhhh okay. Thank you so much for the response! That clears a lot up! I’m new to this and find color cal really interesting. I must have misread something down the line than. I swear ccmx was the “universal” correction and ccss wasn’t. I do have the ccss file that was done in the G93SC which uses the same qd-oled panel. However, there are 3. I assume I use the v2, but what does the WG one mean? The green coordinates seem to be the biggest difference. I already see a big diffrence even with doing HDR! Though, it seems to be tinted red a little bit. I’ve heard using a different observer can correct this.
As for the verification, if I understand correctly from other posts. If I’m verifying an non profiles or calibrated monitor, I check use simulation profile as target profile, and select the color space that I’m using? And when I do apply a profile I want to uncheck it and only have simulation profile set as the one I generated? Does this mean I have to apply the ICC/Icm profile before starting the verification?
I’m still a little confused on verifying the LUT3D. Would I disable the LUT3d and then run the verification using the simulated profile as the one used to generate the LUT3d?
So i did calibrate it and things are looking alot better on the verification side. At least i believe, ill attach an image if you wouldn’t mind verifying i did it correctly. colors seem a bit washed out but, this could just be my lut3d being setup wrong or because I’m doing HDR.
I must have misread something down the line than. I swear ccmx was the “universal” correction and ccss wasn’t.
All colorimeters support CCMX (matrix) corrections … but you have to make those corrections FOR THEM. Only some colorimeters support CCSS (EDR, spectral samples) but you can share them since they store no information about colorimeter (we may argue about inter display spectral variations).
I do have the ccss file that was done in the G93SC which uses the same qd-oled panel. However, there are 3. I assume I use the v2, but what does the WG one mean? The green coordinates seem to be the biggest difference. I already see a big diffrence even with doing HDR! Though, it seems to be tinted red a little bit. I’ve heard using a different observer can correct this.
Share them, or link them. I’ll inspect them and answer you.
As for the verification, if I understand correctly from other posts. If I’m verifying an non profiles or calibrated monitor, I check use simulation profile as target profile, and select the color space that I’m using? And when I do apply a profile I want to uncheck it and only have simulation profile set as the one I generated? Does this mean I have to apply the ICC/Icm profile before starting the verification?
I’m still a little confused on verifying the LUT3D. Would I disable the LUT3d and then run the verification using the simulated profile as the one used to generate the LUT3d?
-LUT3D enabled (in MadVR, in a LUT box, inside a high end monitor…)=> verify against what yoy are trying to simulate, for example Rec709 gamma 2.4 (set simulation profile + use simulation as display profile)
-LUT3D simulated (device link) => same as above but you have to provide LUT3D equivalent in ICC (device link)
-No LUT3D, just verify if the display profile in OS (a custom one made by you) matches display => no simulation profile at all.
-No LUT3D, just verify if the display profile in OS (a custom one made by you) matches display and is able to render images tagged as colorspace X properly in a color managed app like Photoshop => use “X” (for example sRGB) as simulation profile but DO NOT use use simulation profile as display profile (because they won’t match)
-No LUT3D, just check if my custom made profile makes my display behave like sRGB (or other) => not supported by displaycal right now, you can trick it by using Virtual machine.So i did calibrate it and things are looking alot better on the verification side. At least i believe, ill attach an image if you wouldn’t mind verifying i did it correctly. colors seem a bit washed out but, this could just be my lut3d being setup wrong or because I’m doing HDR.
Start by using properly the native gamut SDR modes of that OLED: how to use it properly for color managed apps, how to use it for madVR or Resolve, how to force sRGB-like behavior for non HDR games…
When you can handle this, then go to HDR, but in HDR mode (properly configured) MS win will try to simulate sRGB desktop and you may be applying double color management. Also since monitor expects Rec2020 PQ data as input you’ll have to configure madVR for that.
Share them, or link them. I’ll inspect them and answer you.
Here is the link to the corrections
-LUT3D enabled (in MadVR, in a LUT box, inside a high end monitor…)=> verify against what yoy are trying to simulate, for example Rec709 gamma 2.4 (set simulation profile + use simulation as display profile) ………
Thank you! This makes a lot more sense, i appreciate the detailed explanation!!
Start by using properly the native gamut SDR modes of that OLED: how to use it properly for color managed apps, how to use it for madVR or Resolve, how to force sRGB-like behavior for non HDR games…
When you can handle this, then go to HDR, but in HDR mode (properly configured) MS win will try to simulate sRGB desktop and you may be applying double color management. Also since monitor expects Rec2020 PQ data as input you’ll have to configure madVR for that.
Will do! it Definity makes more sense to start with SDR then hoping straight into HDR. i just prefer the look of HDR even on the windows desktop on this display, SDR makes everything including the desktop heavily clip for some reason, on all devices I’ve tried on it but, i guess this would be a really good way to learn SDR calibration.
I’ll try this an update you. again, i appreciate all your information! i hope these questions don’t seem to dumb, I’ve done a lot of reading on it but i keep running into either conflicting information or recommendations that just don’t work, especially when it comes to QD-OLED. You’ve actually provided me with information that works and visually shows a night and day difference. Thanks again!
