4 kHz polling rate mice for high refresh rate monitors

[Inco^]

Hello,

I was wondering about your experience with 4000 Hz polling rate mice. Specifically, with 4 kHz wireless mice. Using XL2566K, so far I've tested:

- DAv3 Pro and Viper V2 Pro from Razer,
- Lamzu Atlantis OG V2 4K,
- Pulsar X2H.

From my experience, only Razer delivered an effective 4 kHz wireless polling rate experience. For the Lamzu and the Pulsar, 4 kHz feels stuttery that it defeats the purpose of a higher polling rate.

It's too bad I don't like either the shape or the scroll wheel of the Razer mice.
What's your take on the different 4K wireless mice on the market?

[Dalek]

phpBB [video]

[axaro1]

phpBB [video]

I didn't like his video, his testing methodology is, as always, questionable.

He didn't choose a fast speed to test high polling rates (it looks at most like 3cm/s), omitted DPI until called out on twitter, refused to take the right steps to sync frames or ensure good frame pacing. If you compare his 4k/8k vs 1k slideshow with the one that Battlenonsense made years ago with the V8k, you can clearly see that the latter was much better.

For some reason now people think that this video from Optimum Tech completely invalidates proper studies like the one from Sunjun Kim http://kuaa.net/publications/2021-DoWeN ... rMouse.pdf , which clearly highlighted a correlation between refresh rate/polling rate jitter size and jitter perception.

If you ask about personal experience, I can maybe feel a difference in latency between scuffed wireless 1k and 4k/8k, but not necessarily with a good 1khz wired mouse like the XM1r vs an 8khz Dav3 (non pro).

But the real difference is related to visibility, at 360hz I can 100% tell the difference between 1khz and 4k/8khz, asynchronicity induced jitters are perceivable (especially, but not necessarily, if you use motion blur reduction).

[Chief Blur Buster]

8Khz can really hurt performance, so that is true, but it should be packaged in a more technology-understanding way:

But high poll rates DO help, if done properly.

There are some special considerations, about properly optimizing 2KHz+ for human visible motion quality:

1 -- You need fast motion speeds;
2 -- You need high DPI (1600dpi+) and low sensitivity;
3 -- It is more visible at 360Hz+ or strobed.
4 -- Adjust until your CPU can handle it. So use 2KHz or 4KHz instead of 8KHz if your CPU is lagging due to 8Khz.
5 -- Not all games benefit. Some do! Also, sometimes it's about motion quality, not latency.

I see quite noticeable differences between 1KHz and 4KHz when I enable strobing WHILE using high DPI. It makes a lot of the "strobe-amplified microjitters" disappear, if optimized properly.

[Haste]

Does chief have a way to talk to optimum? Maybe he could later revisit this and make a new video with better methodology.

[chandler]

8Khz can really hurt performance, so that is true, but it should be packaged in a more technology-understanding way:

But high poll rates DO help, if done properly.

There are some special considerations, about properly optimizing 2KHz+ for human visible motion quality:

1 -- You need fast motion speeds;
2 -- You need high DPI (1600dpi+) and low sensitivity;
3 -- It is more visible at 360Hz+ or strobed.
4 -- Adjust until your CPU can handle it. So use 2KHz or 4KHz instead of 8KHz if your CPU is lagging due to 8Khz.
5 -- Not all games benefit. Some do! Also, sometimes it's about motion quality, not latency.

I see quite noticeable differences between 1KHz and 4KHz when I enable strobing WHILE using high DPI. It makes a lot of the "strobe-amplified microjitters" disappear, if optimized properly.

what if I use 800 instead of 1600 ? (input lag wise its not much of a difference according to videos I saw on youtube)

or use the 360hz non-strobed in games that dont benefit from MBR that much ?

[Chief Blur Buster]

what if I use 800 instead of 1600 ? (input lag wise its not much of a difference according to videos I saw on youtube)

DPI won't help CPU utilization of high-Hz. It only reduces the number of reports/sec at low frame rates. Besides, 400dpi at 1 inch/sec = only max 400 reports per second regardless of 1000,2000,4000,8000. And when you move slower, it's ugh-grainy, bad for slowtracks (e.g. sniper tracking style movements) if you've got a high sensitivity setting.

[chandler]

what if I use 800 instead of 1600 ? (input lag wise its not much of a difference according to videos I saw on youtube)

DPI won't help CPU utilization of high-Hz. It only reduces the number of reports/sec at low frame rates. Besides, 400dpi at 1 inch/sec = only max 400 reports per second regardless of 1000,2000,4000,8000. And when you move slower, it's ugh-grainy, bad for slowtracks (e.g. sniper tracking style movements) if you've got a high sensitivity setting.

I think u misunderstood me , I was talking about the latency of the sensor ( 400 vs 800 vs 1600 vs 3200 dpi ; etc)

like battle(non)sense explains here in his video (Ive timestamped it to 1:55) :
https://www.youtube.com/watch?v=6AoRfv9 ... on%29sense

you recommend using 1600 dpi but the differences between this and 800 according to his testing are very small

[Inco^]

Thanks for replies.

@Dalek: I already had seen this video.

@Chief:
1. I play quite a fast-paced game.
2. I used 25600 DPI (and also tried lower just to make sure).
3. I used 360 Hz XL2566K with strobing (Persistence = 23 in Blur Busters Strobe Utility).
4. My CPU is I9 12900KS, but the bad 4K experience with Lamzu and Pulsar probably is not due to CPU limitations, because it works well with Razer mice. Besides, for Razer mice, it becomes a bit unstable/unreliable at 8K (with both wireless Razer mice mentionned as well as with Razer Viper 8K wired).
5. It really is the motion quality that is obviously not good (faulty, worse than 1 kHz) with Lamzu and Pulsar at 4 kHz, whereas it is quite good with Razer. I wouldn't be able to tell about latency.

Also I used a PCI-E card for the mouse USB port.

[Chief Blur Buster]

2. I used 25600 DPI (and also tried lower just to make sure).

That's pretty high but I like the setting if it's native non-interpolated; I hope that's a good sensor! I love the use of higher DPIs these days. I now use 3200dpi and 6400dpi for some things.

3200dpi is now my new minimum default setting, simply because it's the highest dpi that doesn't feel turbo-cursor to me (you can reduce Windows Pointer Speed setting as long as you aren't using older non-rawinput FPS games).

1600dpi was just me egging those old 400/800dpi esports veterans. There WERE good reasons, but...

This was when the first "really good" 1600dpi mouse came out. 3200dpi+ (and even 6400ish) is good nowadays with the best sensors.

400 & 800 was often much less problematic in the day of old CS:GO, but newer engines are now getting better with the high-DPI. If you're using the new CS2 engine. The problem is the old esports mindset and flawed math in some old engines that often make them feel a bit less accurate. You can make ~3200dpi behave more exactly like ~800dpi, but with upgrades to slowtrack only. Or whatever multiple that is optimal for your sensor (I keep hearing odd numbers like 900 and 1200, you can simply double it up while halving your sensitivity, to maintain your eDPI). Basically getting the same "400-800dpi" like performance even at 3200dpi+. What magically happens is your slowtracks becomes far less jittery, without interfering with your fastflicks.

I often prefer higher now for modern games, the benefits outweighs the drawbacks if you have a really good sensor and a really good mouse surface, and if it doesn't interfere with your max-flick speed one iota.

It's a big rabbit hole, and even old CS:GO can behave okay at high DPI if you've got things configured at correct math multiples (etc), but there's something about the old less-precise math in the older engines that isn't intuitive (sometimes weird, sometimes good) if you don't properly calculate with a site such as www.mouse-sensitivity.com

One big problem is syncing your edpi across all your games so your flickfeel is the same. That is a big problem sometimes. But you can do multiple things:
1. Use your mouse automatic profile switching, low and high DPI in different games
2. Use www.mouse-sensitivity.com calculator to configure per-game
3. Preserve your muscle memory across all games

It has to be a great sensor whose fastflicks behaves the same at 400dpi through 3200dpi+. That was unobtainum, but not anymore -- you can sync your muscle memory today now, as long as you're not using an outlier/b0rk3d game.

Then you can have 800dpi in one game, 3200dpi in another game, and 12800dpi in another game, and they still flick-turn exactly the same, even if you prefer to arm-flick or wrist-flick. Except your high-DPI games allows you to slowtrack perfectly. 400dpi at 0.25inch/sec = only 100 reports/sec = only 100 mouseturn frames per second = sabotage for refresh rate race. Sometimes in some games you DO need to track that slowly (e.g. games that force you to track slower-moving distant objects, like laser-highlighting a target for guided missles, or continuous "spraying" at a distant enemy like a vehicle going sideways a half a clik away, etc),

And the mouseturnfeel during slowtrack is VERY grainy at low DPI. So, more DPI the better, IF you can optimize the high-DPI problems out of your workflow.

Sites like mouse sensitivity calculators are boon for helping you out if you want to aberrate one game away from the old esports standards (that was more user friendly "set and forget" for older games).