Windows 10 Version 2004
On this note @Slacor has just added the User Bio feature back in - this is really handy for writing your symptoms and triggers for quick reference by others. If you hover over my username you'll see what I mean - really helpful when reading replies from people because they're only really relevant if they have similar symptoms and triggers to you.
As for drivers, I think you guys are right, I'll experiment with older drivers and configs on the 9500 in W10 v2004 before returning it. Just wish I didn't have to trigger such severe symptoms to experiment - if only there was a test that could be run so we know exactly what's going on. @Seagull seems to be making good progress with this though, maybe one day we'll have a reliable system for testing for dithering and PWM in advance of getting sick!
si_edgey This throws up some interesting questions, like why the XPS 9560 would give migraine symptoms in 30 minutes of use with 1909, but the same graphics hardware and drivers allow me to use 2004 all day? And is it the hardware or the drivers in new machines that are the problem, if Windows isn't?
I suspect it's a bit of both. 2004 to my eyes definitely looked more comfortable. Like others mentioned, can you try the same 9560 drivers on the 9500?
The VBIOS defines the video output regardless of the OS/Driver settings, so it could be forcing temporal dithering at the GPU output level. Checking my NUC technical notes there are options to change this, but I can't see them in my BIOS. I guess it's locked down.
si_edgey In my case i don't think the PWM is an issue because one of my favourite device is the iPhone X, and I can use it for hours and hours!
I've tried to disable the FRC on my monitor thanks to the AMD driver where I can choose 6 bit or 8 bit. But I didn't felt any difference about my symptoms, instead the image quality was pretty bad without the FRC.
I really don't know what the problem is!
- Edited
Now moving back to WIndows 7 to do a comparison. My impression is that in 2004, they fixed whatever was making it difficult to focus on (text in particular is hugely better), but somehow it still hurts over time. Certainly an improvement over 1909. I will continue to run 2004 on my laptops, but on my desktop where Windows 7 has no compatibility issues at all, I think I will run Windows 7. My comparison will be on World of Tanks, which I spend a fair amount of time playing on my desktop, and which I had pain over time using 2004.
- Edited
It seems you can install the not-yet-released Nvidia driver 451.22 (which enables the latest WDDM 2.7) right now if you download a big developer file: https://forums.guru3d.com/threads/geforce-driver-451-22.432623/
I give up guys. It doesn't work for me. I can use it for some times but after 1/2 hours I die. I've tried WDDC 2.6 and 2.7 but I didn't see/felt any difference.
Now I have to format again and try the 1803, let's hope to fix it now
My current setup is: AMD ryzen 1700x, RX 480 and alienware AW2518H 240hz.
I will try more thing at the end of the year with the new Nvidia 3000 series and maybe a second SSD to use 2 W10 at the same time
- Edited
I also don't think 2004 is good to me. Maybe i can try older drivers later, but i feel pain in eyes right now. The text is sharper, but strain is there after a while. There is something wrong with colors. The screen is just to bright.
- Edited
I ran some experiments on my two setups:
- Dell XPS 9560 (Intel 630 / Nvidia 1050) - W10 v2004 (updated direct from 1511) - no tweaks, working perfectly - I can use this all day for the first time in years, 1607 - 1909 caused migraines and nausea
- Dell XPS 9500 (Intel UHD / Nvidia 1650) - W10 v2004 - no tweaks, migraines and nausea
I couldn't get the 9560 drivers working on the 9500, so I swapped in the 9560 SSD and booted the unusable machine with the usable OS and drivers. The experiment didn't really work as the laptop auto-installed drivers for the 9500 on startup, and it was unusable.
However, when I swapped the SSD back to the 9560 and reverted to the same display drivers I had been using (in the original post), it was STILL unusable. On my previously perfect setup.
I gave my eyes a day without computers to reset, then tried deleting and installing drivers and tried everything else I could think of - still unusable, and I couldn't find anything that was different, version numbers, BIOS version etc. But something must've updated within Windows somewhere when I booted the SSD on the unusable machine and stayed like that. Whatever changed made my perfect setup unusable for me, on the same laptop, graphics drivers and OS.
I had cloned my drive before I did this (not my first rodeo!) and when I restored from the clone the 9560 is back to being PERFECT again - like looking at a piece of paper on 2004 with no tweaks.
As always, I find myself confused again - some interaction between Windows and the graphics drivers must be at play here. Now I have a perfectly working clone I'm tempted to start fiddling with drivers to see if I can work out what is causing the issue. Any thoughts?
diop System restore got confused when I swapped the SSD to a new machine and back sadly, as my first thought was the same as yours. On the plus side, I'm potentially close to having a system where I can make it 'bad' and 'good' very quickly, on a seemingly identical setup, which will be great for testing.
On Windows 7, I can play World of Tanks for much longer before pain starts. Oh well. At least Microsoft is still fiddling with the rendering. I'd be more concerned if nothing was changing at all between builds.
Over the weekend I freshly installed Win 10 2004 on several Laptops by using an iso image from the techbench site.
(https://tb.rg-adguard.net/public.php)
The following laptops where tested with respect to visual comfort:
- Thinkpad X1 Extreme FHD with Intel UHD 630 Graphics WDDM 2.6 / 2.7 drivers from 2020 (Display: 8 bit)
- Acer Travelmate P645 S with Intel HD 5500 Graphics WDDM 2.0 driver from 2017 (Display: 6 bit + FRC)
- Acer Travelmate P645 MG with Intel HD 4400 Graphics WDDM 2.0 driver from 2015 (Display: 6 bit + FRC)
(The Latest Firefox 77.0.1 browser was used for reading tests on all laptops).
On all laptops I noticed the following improvements over Windows 1909:
- Easier to read and focus on text. Particularly reading in Firefox has improved.
- The screen's content seems to be steadier and calmer.
- No neurological problems. Neither migraine-like symptoms nor any tightening/tingling of the head was noticed.
Nevertheless on all three laptops I experienced some eye strain to a varying degree. The least strain was noticeable on the older Acer latops. There I only got slightly watered, irritated eyes after some minutes of reading in the browser. I got the impression that I could use these laptops for several hours without developing any more serious strain.
The most strain I got on the newer Thinkpad. My left eye accumulated some serious pain that was creeping up after around half an hour of using it. Both eyes got also much more irritated than on the older laptops. By the way upgrading Intel drivers from WDDM 2.6 to 2.7 made no difference.
I assume that perhaps the lower level of WDDM support made a positive difference here. The laptops based on WDDM 2.0 seem to be much more eye friendly than WDDM 2.6/2.7 on the Thinkpad. Perhaps the intel UHD 630 drivers still need some fine-tuning to catch up in eye friendliness.
In short. On the Acer laptops the latest Win 10 iteration is better than the previously installed 1909. Regrettably the same cannot be said for the Thinkpad.
tfouto I don't know to be honest. Right now I am still sick from Saturday, so I can't test anything!
I am sure that the version 1809 is 100% fine on my laptop (dell latitude E5550), intel driver from 2017 with WDDM 2.0 + ditering.exe enable.
With my desktop I was using 1809 + AMD driver from 2017 with WDDM 2.2 and it was 80/90% fine. No brain problem, just some burning eyes.
Last Sunday i've formated the 1909 and installed the 1803 (i didn't have any ISO of 1709 or 1809 right now) but i can't give you any feedback right now. I need more time to recover my brain.
I think there is something that triggers my esophoria on or off, and maybe it could be the dithering that cause a less stable image?! I should try to start some visual training and/or using prismatic lens!
P.S. I think that only bill gates can help us
I'm sorry to see that there aren't that many success stories in here - currently I would view W10 v2004 as a revelation for my temporal dithering / subpixel rendering sensitivity. However, my experience is completely dependent on the Intel drivers.
I've got 7th gen Intel graphics in my laptop (HD 630) and I provided I use very old drivers I can use it all day. I have experimented with new Intel drivers (2018 - 2020) and they cause the same 'unable to focus', squiggly eye feeling that I had with Windows 1909, which will build to nausea and migraines if I continue. However, revert back to the oldest drivers I could find (Intel HD 21.20.16.4475) and the display is calm, easy on the eyes and I can use it all day on my XPS 9560.
I'm going to use this information to really push Intel to investigate.
I have also updated the Nvidia drivers of my onboard 1050Ti to the current version (26.21.14.4614) and I am experimenting with gaming on my laptop for 1 hour at a time, and have yet to get any symptoms - bear in mind this is the first time in 8 years I can game on a laptop.
Feeling boosted by all this, I went to work trying to get my brand new XPS 15 9500 to work (Intel UHD 630 on board). After days of experiments, I feel there is nothing that can be done to make this new integrated card work for me - it triggers crazy nausea, migraines and dizziness from working on it.
Overall I still can't quite believe I have a working setup, on good hardware, and a modern OS that is basically 'vanilla'. Really hope there are further success stories to come from v2004.
- Edited
JTL So frustrating isn't it? Don't suppose you're in Scotland, or even the UK...? I'm willing to chip in some money on finding a solution to this for everyone, or even advancing our knowledge of it. So if you have hardware I could use or purchase to help with investigation then I'm 100% game.
In reality though, this isn't our problem to solve, it's Intel's / AMD's / Nvidia's. I've been speaking with my contact at Microsoft and he has been speaking to a graphics guy who agreed with our hypothesis of hardware bit-depth up-scaling (using temporal dithering) that can happen at the output stage of a graphics adapter, so that confirms why some systems are 'bad' and will always be 'bad'.
But on 'good' hardware that is made bad by drivers - that is unacceptable, and will be an easy fix if we can get someone to care. These companies don't want bad press so let's start thinking collectively about how we can make a loud noise about this and get it sorted once and for all.