Acer GN246HL - Troubles with Linux

[RealNC]

Hm. This does make testufo run at 120FPS for me, but scrolling is still 60FPS.

[Chief Blur Buster]

These tests all run perfectly fine without any ghosting/blur/motion problems. Sure the browser is now forced to 120fps. But it certainly gives you the ability to compare results with lightboost related visuals on Windows/Linux.

Yes it certainly makes TestUFO more useful for 120Hz Linux users.
(Though it means a manual step to hardcode a browser to match the refresh rate)

[jigoku]

Hm. This does make testufo run at 120FPS for me, but scrolling is still 60FPS.

I've noticed that too. Although figured that most compositors seem to lock most applications at 60fps.
For example, if i use KDE5/kwin as the desktop and compositor, firefox seems to scroll at 60fps despite the layout.frame_rate being over rided. If you disable the kwin compositor (eg; desktop effects) firefox then scrolls much more nicely most of the time. There is still the odd stutter, due to it not being vsynced however.

I've found better luck using Awesome Window Manager, along with a simple compositor like xcompmgr or compton. These don't seem to affect firefox when scrolling at 120fps.

I'll keep trying to figure out these problems, if there's any work arounds, but so far for this monitor i've posted a bunch of information i've found out so far at this small repository, which may be helpful for others trying to fix blur related problems on this monitor with Linux;
https://github.com/Jigoku/Acer-GN246HL-linux

I'm still getting a bit of corona/blur in the bottom half of the screen, and cannot seem to find any commands over the ddc protocol, where i may be able to adjust the strobe phasing, which i have heard helps with other monitors having this effect on part of the screen. I'll keep on the look out for more solutions though.

[Chief Blur Buster]

I'm still getting a bit of corona/blur in the bottom half of the screen, and cannot seem to find any commands over the ddc protocol, where i may be able to adjust the strobe phasing, which i have heard helps with other monitors having this effect on part of the screen. I'll keep on the look out for more solutions though.

Non-Strobed Mode
If you're using non-strobed mode, it's possible that panel heating may be affecting pixel transition speeds asymmetrically on different parts of the panel. The power supply is usually hidden behind the bottom edge of the panel, which makes the bottom edge of panel hotter, which can cause ghosting to look different. Cold LCDs will ghost differently from hot LCDs. But this can affect the same panel surface if one part of the panel is cold and the other part of the panel is hot.

For Strobed/LightBoost Mode
This is simply strobe crosstalk, the asymmetry between the top-to-bottom LCD scanout versus the all-at-once strobe flash. For more reading, please see Advanced Strobe Crosstalk FAQ. It is normal for the top/bottom edge of LightBoost to have much more ghosting then the rest of the panel. For monitors that have adjustable strobe phase (BenQ/Zowie monitors) it is possible to put the crosstalk band somewhere else in the screen (e.g. even the middle!)

[jigoku]

I'm still getting a bit of corona/blur in the bottom half of the screen, and cannot seem to find any commands over the ddc protocol, where i may be able to adjust the strobe phasing, which i have heard helps with other monitors having this effect on part of the screen. I'll keep on the look out for more solutions though.

Non-Strobed Mode
If you're using non-strobed mode, it's possible that panel heating may be affecting pixel transition speeds asymmetrically on different parts of the panel. The power supply is usually hidden behind the bottom edge of the panel, which makes the bottom edge of panel hotter, which can cause ghosting to look different. Cold LCDs will ghost differently from hot LCDs. But this can affect the same panel surface if one part of the panel is cold and the other part of the panel is hot.

For Strobed/LightBoost Mode
This is simply strobe crosstalk, the asymmetry between the top-to-bottom LCD scanout versus the all-at-once strobe flash. For more reading, please see Advanced Strobe Crosstalk FAQ. It is normal for the top/bottom edge of LightBoost to have much more ghosting then the rest of the panel. For monitors that have adjustable strobe phase (BenQ/Zowie monitors) it is possible to put the crosstalk band somewhere else in the screen (e.g. even the middle!)

Thanks for the explanations. I've had a look at that page previously on strobe crosstalk. You mention heat issues, i assume that could be part of it, as the monitor does get mildly warm at the bottom, the casing behind protrudes half way down the back where the connectors are etc, and generates a little bit of heat. However, this doesn't seem to affect the 144hz non-light boost mode (there is normal ghosting/coronas as expected with the tests which are equally present across the entire screen).

For crosstalk, it seems this monitor doesn't support configurable crosstalk settings. I've tried exploring the DDC features using `ddcutil probe` on Liunux, and also checked out the manufacturer(?) menu (by holding down the menu button whilst powering on), but doesn't offer anything related to that.

There does seem to be some settings that can't be adjusted however;

Code: Select all

   Feature xe0 - Manufacturer Specific
   Feature xe1 - Manufacturer Specific
   Feature xe7 - Manufacturer Specific
   Feature xe9 - Manufacturer Specific
   Feature xeb - Manufacturer Specific
   Feature xfa - Manufacturer Specific

I figured out that 0xfa is the feature to adjust lightboosts brightness, but the other undescribed features don't seem to alter anything. I did read that most lightboost monitors don't actually support adjusting the crosstalk/strobe/phasing settings, as they are allegedly pre-tuned at factory.

Either way, this monitor is more than usable for now. Whilst there is corona/blur on the bottom half of the screen using lightboost, it's not actually noticeable in any gaming or fast camera-panning in any videos.

PS; there were two other features discovered by `ddcutil probe`

Code: Select all

   Feature x20 - Horizontal Position (Phase)
   Feature x30 - Vertical Position (Phase)

I'm not sure if they are relevant to crosstalk/strobe, as they cannot be written to over DDC.

Maybe Acer locks out this setting from being adjustable?

[Chief Blur Buster]

I think that's for analog VGA.

As for strobe crosstalk, don't worry about it -- it's normal for top/bottom edges.

As for LightBoost brightness, it's a clarity versus brightness tradeoff, see LightBoost 10% vs 50% vs 100%. It adjusts the strobe flash length. Easiest way to calibrate is via TestUFO Ghosting Deluxe (tiny text) or the TestUFO Panning Map Test (street name labels).

I believe 0xFA is indeed strobe length adjustment for LightBoost, but that's it. You can adjust it in the monitor's OSD however...