Google Chrome Temporal Dithering

machala

Hi all,

I'm opening this topic for Google Chrome specifically, but my issue is general like for many on this forum. I'm struggling to find "good" hardware - laptops, PCs, tablets, phones that don't cause eye strain or even motion sickness. I used to have 0 issues prior to, say, 2011/2012, but since that time it's getting worse and worse and I know it's not me as old hardware I still have is still perfectly OK for me.

Unlike a lot of people on this forum I am rather confident the source of these issues are various forms of flickering, interestingly I don't believe the usual suspect of PWM is the source of any of this, although of course it is certainly better to avoid PWM, especially with lower frequencies. On the contrary, speaking of PWM, I read some forum posts claiming that some monitors have non-PWM brightness control implemented so bad that it may actually cause current fluctuations that result in a more serious low frequency flicker, worse than anything that PWM may create.

My symptoms are probably better than many other people as I've been able to get used to majority of the new gear finally (after some research and testing), but every second of using it I realize it is not OK and not comfortable, it's just acceptable.

In this particular instance I have a question about Google Chrome. I used it for a short time longer time ago and never touched ever since (I'm on FF) until now in order to try out Google Stadia. The browser is terrible, causing eye strain and motion sickness. I am quite certain it is caused by temporal dithering, but I don't remember it from before so it had to be introduced in the meantime, probably several years back.

My questions to the forum users are following:

1/ Are you experiencing the same - no issue in Firefox while you get eye strain with Chrome?

2/ Are you aware of any method to turn this dithering "feature" off in Chrome?

3/ If you do experience the same, can you support this in the following Chrome Help community thread and, more importantly, the bug report to Chromium devs?

https://support.google.com/chrome/thread/87823703?hl=en#

https://bugs.chromium.org/p/chromium/issues/detail?id=1157382

Any help, support or comment is greatly appreciated.

Thank you!

Seagull

machala

Chrome or any other browser does not dither. I've tested it with my capture card / software.

Your problem is more likely being caused your gpu / monitor dithering the colours chrome uses. However, unless you can figure out how to change the colour scheme in chrome there is isn't much you can do about it.

optix

machala

1) I'm experiencing exactly the same on a otherwise safe system. Firefox is ok.
2) Unforunately not
3) WIll do.

JTL

Seagull I'm beginning to feel that a more fruitful line of investigation for this particular issue is to compare the frame of a "good" application versus a "bad" application, to see what exactly is different.

I mean, while there (probably) isn't any motion going on. I still feel optical illusion type tricks are possible.

Unfortunately I haven't had time to work on VideoDiff as of recent. Contributions welcome.

machala

Hi, definitely not, I'm 100% sure of what I said. Firefox on the same OS, same PC, same screen is causing zero issues (just like all other applications) while I get eye strain after several minutes with Chrome. Additionally, see this:

They turned dithering on in Chrome in order to solve color banding back in 2015:

https://bugs.chromium.org/p/chromium/issues/detail?id=41756

You can even see someone complaining about eye strain in the middle of this bug chain - that they started to experience eye strain after they enabled this. The user was totally ignored, see for yourself under the link above.

However, dithering was still not fuctioning on GPU-accelerated setups, which they "fixed" under this bug in 2017:

https://bugs.chromium.org/p/chromium/issues/detail?id=731693

I hope this makes it clear. And of course, the best proof for me is what I see with my own eyes...

I was looking into it further, perhaps the only way would be to compile Chromium yourself with dithering disabled, but it's a lengthy process...

I understand what you're trying to say: that it's the graphics card/monitor mysteriously dithering differently in basically every application on the PC and differently in Chrome. But I do not believe this is likely... plus see the bug reports above, Chrome (Chromium) is using dithering very officially...

Seagull

machala

https://ledstrain.org/d/895-gpu-dithering

I used this methodology to test Chrome and found no dithering from Chrome itself.

Best of luck in solving your problem.

Quad43

wow..... this explains a lot for high browsing usage eyestrain. would be great to find a solution

machala

Quad43
Have you ever used Firefox for browsing? If not, can you do that and see if it makes things better for you?

Please keep in mind that the issue is not with Chrome, but with its Chromium rendering core - core that is used by many other browsers, including MS Edge. I'm not using Apple products, I'm on Windows and Linux (BTW I have the issue with Chrome on both systems, no difference), but as far as I know Safari is using dithering as well.

Firefox is the only browser I know of that should not be using dithering.

machala

Seagull
Hi again,

thanks a lot!

2 questions if I may:

what Chrome version did you test this with?
Can you describe your setup please? OS, GPU, driver versions?

It is possible that both of us are right - and there are setups and conditions under which Chrome does not use dithering - which could lead to an alternate solution for me.

diop

machala HI there,

Have you tried the latest MS Edge browser? Interestingly I get strain from Firefox/Chrome/Chromium but Chromium Edge is comfortable for me.

FWIW, I have trouble with modern media players and emulators too. Kodi causes eyestrain (despite dithering being switched off in settings), retro emulators (e.g. Retroarch) cause discomfort compared to a 2010 emulator (Kega Fusion) which is like looking at paper.

Seagull

JTL

As I explore here, the slightest of changes in colour shade can have a significant effect on how a GPU, and in all probability, how a monitor dithers. I used to have problems like this, where a software update would cause me eyestrain, till I switched to a GPU which does not dither.

I'm sorry, I don't see the value here in investigating further. Every time I have investigated a piece of software for dithering, or other unseen the motion, the outcome has always been nothing. I have only ever seen GPUs dither. Honestly, I am getting tired of repeating it. If people want to ignore it, they are free to do so, but its one of the few opinions here backed up by testing which didn't relying upon human perception.

AgentX20

I have experienced eye strain with Chrome and Firefox in the past on a known safe system. Launch Chrome and the strain starts, quit it and all is good. I've seen the same thing with version 4 of the flight simulator called Prepar3d.
Strangely, and it's a good thing, the new MS Edge browser has been safe so far.

JTL

Seagull As I explore here, the slightest of changes in colour shade can have a significant effect on how a GPU, and in all probability, how a monitor dithers

I wasn't ignoring you. In fact I'm planning to implement similar TTC histograms in VideoDiff when I have time.

Seagull the slightest of changes in colour shade can have a significant effect on how a GPU, and in all probability, how a monitor dithers. I used to have problems like this, where a software update would cause me eyestrain, till I switched to a GPU which does not dither.

I plan to do more research with the cards I have at some point

KM

I'd be willing to do a lot of testing, but I don't know which capture card to buy that is the cheapest but would still suffice for our needs (which are not having perfect color accuracy but detecting mere color changes). But I don't have a fast PC and don't know how much money would need to be spent to buy one that's fast enough. Probably too much to justify testing. It seems some cards can save directly to SD card, so maybe it could be possible to do it without a PC. But I don't know whom to ask for more info. I don't want to spend hundreds of Euros nor end up buying something that turns out not to be usable eye-strain-wise (PC, GPU, OS) or is insufficient to detect temporal dithering at all. I try to gather information about this every now and then, but the market for capture cards is overwhelming and not made with temporal dithering testing in mind. If we had info on what to buy that does what we need but costs the least money it would be easier to continue.

Seagull

I use this. I know it is sufficient for recording dithering by using ditherig.exe on intel integrated graphics. I can set ditherig to temporal dithering, or none, or any other setting and get different results from the captured footage.

https://www.blackmagicdesign.com/uk/products/decklink/techspecs/W-DLK-33

You'll need a card which will record uncompressed. Any compression at all could mess up your results. My software is pretty basic to use and can be provided.

JTL

KM What is the fastest desktop PC you have access to?

I bought a DVI2PCIe card ($2000 MSRP), but you can sometimes find them cheaper on eBay and such. (Think I personally paid around $400 CAD for mine), and with a quad core i5 from 2012 it's certainly fast enough for my VIdeoDiff research

Sunspark

I agree. While spatial (this is fine) dithering might have been implemented, I really really doubt that Chrome with all the other things they need to work on decided they needed to focus on implementing FRC temporal dithering inside a browser when it already exists in monitors and GPUs.

I am primarily a Firefox user, but I do have Chrome, Edge, etc. installed. With Edge I have it in hardware mode settings, and in Chrome I have it in software mode. I would suggest that in Chrome you click on the menu, select settings, select advanced on the left sidebar, select system and disable "use hardware acceleration when available". This would give you a data point on whether or not it's the GPU.

machala

Big thanks to everybody for their replies. Yes, I tried a lot of other Chromium-based browsers, including the new MS Edge and to me they have EXACTLY the same effect on my eyesight and giving me same headaches. Chrome, Vivaldi, MS Edge... all of them both in Windows and Linux, on different systems (2 laptops, 2 PCs, 3 of them with Linux and Windows installed, one with Windows only).

I know that I am very sensitive to flickering and motion, back in the CRT days, I could detect monitors of up to 85Hz and even guess what refresh rate they were using (60, 75, 85), only 100Hz and above was stable for me. I personally do not believe this issue with Chromium-based browsers is caused by different colors and their interaction with graphics card or monitor dithering as I see exactly the same effects on various systems and no other application is causing this on any of them. I can sometimes almost see the picture instability in Chrome. I hope that the devs will react to the bug report and I again urge anybody with the same issue with Chrome to star the report or even comment there confirming the issue.

Thank you very much!

Quad43

Super interesting. Reading this site on FF all good. Opened up Chrome and browsed regular page/sites, bam. Within a minute or two. I'm definitely less inclined to use FF due to speed but based off this I may have to. Great discovery machala! I hope we can get it resolved.

Lauda89

Maybe it's a problem related to the font?
I am reading your post with Chrome and firefox with zoom 110% (my daily use) and the font is really different.

Do you have problems with Chrome even when you watch a youtube video on full screen?

Gurm

If you install older Chrome (v68 works for me) you'll see that it is better. Eyestrain started for me with Chrome v69. Problem is v68 is out of support now. 🙁