who can use Linux but not windows? Share your Linux setup

jordan

A Hackintosh built on my Gigabyte Z390 Aorus Master with i9-9900K has some issues (common with many Z390 and 9th gen CPUs setups). There is a challenge in configuring the graphics because the best profile for UHD 630 is the Mac Mini, but there is no Mac Mini based on 9th generation processors. By default, the integrated GPU doesn't work, and a discrete AMD graphics card is needed. I used the RX6600, but this setup has significant dithering issues and is not functional. I managed to get UHD 630 (i9-9900K) working (which is quite an adventure and can take at least half a day), it's not comfortable for the eyes. The bootloader option 'enableDither=0', which is supposed to disable dithering on the iGPU, doesn't work with that setup.

Therefore, for a comfortable experience, I would recommend using Apple Silicon M1. It's better to first find someone you know and try working on M1. The Mac Mini 2018 based on the i7-8700 is very noisy; the only option is the one based on the i5-8400, but that's not as powerful a machine. However, since you have i7-8700 like in the Mac Mini 2018, you might want to give it a try with Hackintosh and see if it works out.

WhisperingWind these guys(which i got to the through jordan's comments) are academic researchers in the field of vision. So they built a library to do vision experiments.

A part of this library is to verify What goes to the screen: http://psychtoolbox.org/docs/BitsPlusIdentityClutTest

They even fixed a bug in ubuntu, related to dithering: https://psychtoolbox.discourse.group/t/dithering-with-ubuntu-22-04-md-radeon-pro-wx-3200-viewpixx/4517

autobot but it seems only for their device. So not sure if this helps…

autobot

It seems they use displays that can show 10 or 12 bits. This is the maximum for some cards and doesn't allow them to dither. They broke something in the Linux kernel for the 10/12-bit modes for some AMD cards, which seems to have caused dithering. Then they fixed it. This fix has been in the kernel for about two years.

I have come across comments in the amdgpu kernel module code stating that for some cards, dithering should not be enabled at 12 bits. For older cards, I think this value could be 10 bits.

ryans

I made a spontaneous large number of edits in the i915 module and it seems I forced a smaller number of bits per channel, which resulted in improvements on my monitor with FRC. Later, after studying the i915 code more carefully, I realized that dithering for Intel iGPU is disabled by default in the i915 (except for 6 bits in certain modes). I additionally removed these 6-bit checks from the i915 code so that dithering would not be enabled. I will check if there is dithering at 6 bits on the vanilla kernel, then I can post the fix along with the build guide. But it is only for intel igpu and 6 bit mode.

autobot

Please answer a few questions when you have time.

1. If you enable hardware acceleration in OS / Firefox, will your eyes feel strained?
2. How intense will it be?
3. Will the issue recur when watching videos and reading text?
4. Could you try focusing your gaze on the dot of the letter "i." Will it feel like your gaze wants to slip off with hardware acceleration on and off?

Sorry for the many questions; I'm just trying to understand how hardware acceleration affects comfort in Linux.

...............

............

WhisperingWind

For me on "bad screens", what I see around letters is a blue glow.

The first time I was able to notice this was after I used a good screen for a few hours with a pure red screen filter, i.e. blue and green was entirely disabled (NegativeScreen app).

After doing that, I looked at a bad screen again, and both these bluish glows around text and other contrasting items + weird extra "subtle shadows" around UI elements was suuuuuuper obvious to me

Another really interesting difference between bad screens and my "one good screen" is on that good screen, using a pure red screen filter looks entirely red, I do not see any other colors at all even while scrolling.

On essentially all other screens, even though I can confirm that only red subpixels are activated with a macro camera, I still see "green and blue trails" while motion happens on the screen (and sometimes, even when there's no motion, I can see green+blue+yellow colors vibrating at edges of contrasting objects that increases when I move my head around). Macs are especially bad with this. In addition, setting most screens to only use blue pixels will show red motion trails.

I don't think it's the other pixels "secretly" activating, they look like they are disabled, but more about how the way the LCD RGB color filters were physically manufactured somehow creates more than just the color they're supposed to. It also might be connected to LCD response time "overdrive" methods as those are frequently associated with colored motion trails.

The strangest part about this though is how my "one good screen" mysteriously is a total outlier here, since it doesn't have the above issue at all.