- This topic has 81 replies, 13 voices, and was last updated 4 months, 4 weeks ago by
.
-
Posts
-
Are your sure you are using the “simplest” and more compatible ICC type? single curve + matrix + BPC.
I didn´t understand it previously, but now it should be simple, compatible ICC, right?
Chrome, Edge, Davinci Resolve and Adobe Premiere pro are showing colours oversaturated. Photoshop and Windows gallery are showing it correctly, tho.-or oversaturation which is linted to app not usig OS info about display profile or that OS has not assigned as default profile the profile that describes display in its current preset & config.
It looks like it is this problem, then.
I have it set to Displaycal to load my colours instead of letting the operating system handle calibration loading.@MW Thanks, I´m sending my ICC profiles here.
2. Ensure that DisplayCal profile loader is started with windows. If not you can enable it by rerunning the Displaycal installation exe.
It started with Windows, and I did a few reinstalls of Displaycal, so it should be all right.
Chrome, Edge, Davinci Resolve and Adobe Premiere pro are showing colours oversaturated.
“matrix” profile, yes. But Resolve is not color managed in Windows, it won’t use display profile… or if it can IDNK how. In MacOS it will use display profile.
For windows you’ll need a LUT3D fro Resolve (source Rec709, destination your display profile). As a general rule for a general purpose display that is not going to be connected through a decklink card or something like that, rememberto DO NOT add VCGT to LUT3D, since you’ll keep VCGT calibration for the whole system.
novideo_srgb seems to be broken and crashes at launch, at least with the latest nvidia driver. Not sure if it’ll ever get fixed. Just browsing through this thread I’m trying to figure out the best way to create a profile with displaycal for windows acm.
My understanding for Calibration for an LCD-type panel (Mine is fairly close to 2.2 flat, not the srgb tone curve)
CIE1932 2 degree
white point 6504K
white level as measured
black level as measured
tone curve gamma 2.2 relative
black output offset 100%’
Calibration Speed HighProfiling:
single curve + Matrix
black point compensation enabled?
profile quality highAny other settings that are recommended? Will Windows ACM actually use the 1D LUT created by setting the tone curve in the calibration settings?
I’ll add my display is wide-gamut with garbage fake HDR, so I basically use it as an SDR display. It doesn’t have a clamped mode, so I’ve used novideo_srgb to clamp and do some calibration. I want to emulate the same thing. I want non colour-managed applications to clamp correctly. I don’t use any colour-critical software.
I made a test chart in displaycal for novideo_srgb. Should I use the same one here?
Test chart (novideo_srgb.ti1)
“Single channel” = 21, “Neutral” = 256, “Multidimensional” = 5
white patches = 4, black patches= 4If it can clamp it can also do more exotic changes like inverting red and blue which is easy to test with a custom profile. The one attached also has a 1D LUT that darkens so you can also try that.
Calibration Speed High
IDNK how bad is your display greyscale uncalibrated, but if you have a fast device like i1displaypro… medium speed should be your safe approach. Slow, if you notice some color cast not related to the lack of dithering or high bitdepth LUTs in your GPU.
I never use it on high speed.
Profiling:
single curve + Matrix
black point compensation enabled?Yes, enabled, “fake infinite” contrast is the safe choice.
On displays very limited contrast, because display is bad or by your own choice (250:1 or lower for some print preview), you may want to store actual black level (BPC disabled), but otherwise enabled by default.
Any other settings that are recommended? Will Windows ACM actually use the 1D LUT created by setting the tone curve in the calibration settings?
AFAIK it will use EDID information regarding TRC and primaries. So a widegamut set on hardware level to an sRGB preset, when you enable ACM it will further desaturate that sRGB because ACM believes that it is a widegamut.
Your custom 1DLUT in GPU through displaycal loader shoudl be there AFAIK but I’ve not tested this.Calibrite Display Pro HL on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.Thank you for the responses. My display is around 2500:1 (VA panel). Doesn’t have an sRGB preset or any way to clamp at the hardware level, so I don’t think I have to worry about ACM desaturating the image too much.
From what I’m reading ACM expects the sRGB tone curve if I load an application that is NOT colour managed. That gets a little tricky if a game or other application is actually expecting 2.2. It was nice to use novideo_srgb and be able to change the tone curve on the fly if a game looked washed out or crushing black. Oh well. I’ll give it a shot.
you can find some newer forks of novideo_srgb on github which may have fixed issues with newer Nvidia drivers. Or you can use dwm_lut.
After creating my new icc I realized I could verify the results with HCFR with auto color management enabled. It’s definitely clamping my wide-gamut down to srgb colour space correctly. It’s not observing the vcgt in the profile, so the tone curve is not calibrated but my monitor tracks relatively close to a flat 2.2 anyway. I may try that tool that converts displaycal profiles to MHC2 compatible ones to see how it looks.
In terms of visual quality, I think it actually looks much better than what I was getting with novideo_srgb, and I wonder if that’s because of the black point compensation being enabled. The github for novideo_srgb always suggested having it disabled, but I think I likely had some black crush (especially with a VA panel). novideo_srgb does say that it requires accurate reporting of the monitors black point to get good results.
Tried using MHC2Gen … not really sure it’s doing the right thing. I just took the profile I’m using now and tried converting it to an MHC2 profile for ACM. When I switched to that profile that gamma looked like 1.8 or something. I’m not sure what it was doing. The tooling around this is bad.
I can live with whatever clamping it’s doing now until novideo_srgb gets fixed or one of the branches.
Can you attach the profiles? Did you enable HDR for the MHC2 to work, or did it work in SDR as well?
Can you attach the profiles? Did you enable HDR for the MHC2 to work, or did it work in SDR as well?
I used this command: MHC2Gen sdr-acm [–calibrate-transfer]
I think I tried both with and without –calibrate-transfer, so in the MHC2 profile attached I’m not sure which option was specified. I overwrote the file when experimenting. All I know is after using the MHC2Gen tag the black level was elevated and checking in HCFR showed a really odd gamma that didn’t match anything, when the monitor is set roughly 2.2 flat.
I’ve also played a little bit more and created some new profiles, one that was 2.2 and one with srgb tone curve. All other settings in displaycal were kept the same. Flipping back and forth while playing a game, I can see there are differences in the tone curve/gamma, so even without the MHC2 tag Windows ACM does something with the colour profile that adjusts the image on non-managed applications besides clamping to SDR gamut. I took a look at Warhammer Space Marine 2 because it has a simple in-game brightness adjustment that allows you to adjust both a low and high to avoid black and white crush. With my 2.2 profile, I don’t have to adjust anything. With the srgb profile I need to adjust both the low and high brightness images to expand the contrast.
So maybe I’m misunderstanding how this works. I know non colour-managed applications should default to srgb, but I thought it only used the monitor EDID or maybe the lumi tag in the profile. I thought the vcgt was completely ignored.
