| Summary: | HDR desaturation issues with after updating to 6.4 from 6.3.5 | ||
|---|---|---|---|
| Product: | [Plasma] kwin | Reporter: | pvpdmail |
| Component: | colour-management | Assignee: | KWin default assignee <kwin-bugs-null> |
| Status: | RESOLVED FIXED | ||
| Severity: | major | CC: | avery-j, kde, kdedev, nate, xaver.hugl |
| Priority: | NOR | Keywords: | regression |
| Version First Reported In: | 6.4.0 | ||
| Target Milestone: | --- | ||
| Platform: | Arch Linux | ||
| OS: | Linux | ||
| Latest Commit: | Version Fixed/Implemented In: | ||
| Sentry Crash Report: | |||
| Attachments: |
6.3.5 kscreen-doctor -o at 0% sRGB intensity
6.3.5 kscreen-doctor -o at 100% sRGB intensity 6.4.2 kscreen-doctor -o at 0% sRGB intensity 6.4.2 kscreen-doctor -o at 100% sRGB intensity 6.3.5 image at 100% sRGB intensity 6.3.5 image at 0% sRGB intensity 6.4.2 image at 100% sRGB intensity 6.4.2 image at 0% sRGB intensity 6.3.5 drm_info at 0% sRGB intensity 6.3.5 drm_info at 100% sRGB intensity 6.4.2 drm_info at 0% sRGB intensity 6.4.2 drm_info at 100% sRGB intensity |
||
|
Description
pvpdmail
2025-06-20 16:08:50 UTC
I am also having this issue with a GSM LG ULTRA monitor on Arch as well. Same problem here: washed out colors since plasma 6.4. Persist regardless of connection (DP v1.4 & DP v1.4 over Thunderbolt; same for v1.2) and GPU (AMD Radeon 780M & AMD Radeon 9700 XT)
$ ddcutil detect
Display 1
I2C bus: /dev/i2c-13
DRM_connector: card0-DP-5
EDID synopsis:
Mfg id: GSM - Goldstar Company Ltd (LG)
Model: LG ULTRAWIDE
Product code: 40577 (0x9e81)
Serial number: 201NTSU1H406
Binary serial number: 51406 (0x0000c8ce)
Manufacture year: 2022, Week: 1
VCP version: 2.1
I don't see this on git-master with an NVIDIA GPU. Hopefully someone with an AMD GPU can test this on git-master Can one of you who is seeing this bug please share a screenshot of washed out SDR content with some HDR content to highlight the difference for folks who may not know what to look for? Thanks (In reply to TraceyC from comment #4) > Can one of you who is seeing this bug please share a screenshot of washed > out SDR content with some HDR content to highlight the difference for folks > who may not know what to look for? Thanks I tried taking screenshots with HDR enabled and disabled but they hardly look any different. It doesn't match what I see in person. I forgot to mention I'm using a NVIDIA RTX 4070 SUPER. It would seem both AMD and NVIDIA users are having the same issue. (In reply to Avery J. from comment #5) > I tried taking screenshots with HDR enabled and disabled but they hardly > look any different. It doesn't match what I see in person. I forgot to > mention I'm using a NVIDIA RTX 4070 SUPER. It would seem both AMD and NVIDIA > users are having the same issue. Thanks for the additional information. Can anyone who is affected take a picture of their monitor with a phone or another camera that shows the problem? Thanks. Just to be sure, if you disable DDC/CI, then disable HDR and enable it again, does that make a difference? Also, please attach the output of kscreen-doctor -o and drm_info (as files). Ideally for both 6.3.5 and 6.4 Created attachment 182915 [details]
6.3.5 kscreen-doctor -o at 0% sRGB intensity
Created attachment 182916 [details]
6.3.5 kscreen-doctor -o at 100% sRGB intensity
Created attachment 182917 [details]
6.4.2 kscreen-doctor -o at 0% sRGB intensity
Created attachment 182918 [details]
6.4.2 kscreen-doctor -o at 100% sRGB intensity
Created attachment 182919 [details]
6.3.5 image at 100% sRGB intensity
Created attachment 182920 [details]
6.3.5 image at 0% sRGB intensity
Created attachment 182921 [details]
6.4.2 image at 100% sRGB intensity
Created attachment 182922 [details]
6.4.2 image at 0% sRGB intensity
Here is both kscreen-doctor -o from 6.3.5 and 6.4.2 at both 0% and 100% sRGB intensity and some sample images of the difference. Personally, I lost track of which 6.4.2 image was at which intensity as they are exactly the same. For 6.3.5 you should probably look at the blue bar of "Display & Monitor" where the effect shows up best on camera. It goes without saying that the effect is a lot more noticeable in person than in camera (I had to put the white remote there just to try and have a constant white balance). I couldn't provide the drm_info as I didn't have it installed. On another note, I had a slight suspicion that having the sRGB intensity slider at 0% when updating and then increasing it might fix the issue, but no. Also, disabling HDR, DDC/CI and turning the monitor off and on again didn't help with the issue. It feels bad to have to revert back to 6.3.5 due to the HDR bug when 6.4.2 is so good :( The color intensity slider not working most likely has the same cause as bug 505971, and should be fixed in 6.4.3. The output of drm_info would allow me to make sure of that. If your display clips colors to the HDR metadata we provide, then the same fix might also help with this problem. If you want to test it before the release of 6.4.3, you could compile the Plasma/6.4 branch of KWin yourself, or by adding only https://invent.kde.org/plasma/kwin/-/commit/cd4f1ebd3d316de965bdb303121639b3a90a4fdb to the PKGBUILD of the current kwin package and rebuilding that. Created attachment 182951 [details]
6.3.5 drm_info at 0% sRGB intensity
Created attachment 182952 [details]
6.3.5 drm_info at 100% sRGB intensity
Created attachment 182953 [details]
6.4.2 drm_info at 0% sRGB intensity
Created attachment 182954 [details]
6.4.2 drm_info at 100% sRGB intensity
Hope these are fine. Good, that's exactly the same as in bug 505971: > │ │ │ ├───Display primaries: > │ │ │ │ ├───Red: (0.6400, 0.3300) > │ │ │ │ ├───Green: (0.3000, 0.6000) > │ │ │ │ └───Blue: (0.1500, 0.0600) > │ │ │ ├───White point: (0.3127, 0.3290) Let's wait for 6.4.3 then before doing any more digging. It's almost certainly gonna be fixed there :) Update 6.4.3 seems to have fixed this issue for me. Colors now look just as good (maybe even better due to the calibration) as in previous versions. |