Polling Rate among Gaming Mice Analysis: 1000hz Is Not Enough

Everything about latency. Tips, testing methods, mouse lag, display lag, game engine lag, network lag, whole input lag chain, VSYNC OFF vs VSYNC ON, and more! Input Lag Articles on Blur Busters.
deama
Posts: 224
Joined: 07 Aug 2019, 12:00

Re: Polling Rate among Gaming Mice Analysis: 1000hz Is Not Enough

Post by deama » 08 Jul 2020, 17:50

NDUS wrote:
13 Jan 2020, 20:23
BTRY B 529th FA BN wrote:
13 Jan 2020, 18:15
How can I get my mouse to poll at 8000Hz?

Using a Bloody SP30
Sensor: PMW 3360 Optical Engine
Can only be done with the following mice:

- any mouse with MLT04 sensor (but it is pointless, because MLT04 is an ancient sensor that maxes out at 1.5m/s, defeating the purpose of 8000hz polling)

Or, one of the two mice that use qsxcv's custom PMW3360 firmware that is designed for 8000hz polling:
- Zaunkoenig M1K (probably shipping in 1-2 months)
- Ninox Astrum (apparent vaporware with no update in a year)

If you have one of these options, this overclock.net thread has the right driver
Could other mouses be overclocked to 4000hz at least?
Also, I hear that a lot of mouses these days have an artificial cap on the polling rate, is there some way to hack it and disable that?

konchy
Posts: 10
Joined: 11 Dec 2018, 04:54

Re: Polling Rate among Gaming Mice Analysis: 1000hz Is Not Enough

Post by konchy » 22 Jul 2020, 14:08

Would like some clarification.
Why would all these high refresh rate mice matter when our monitors can't even display them?

For instance, if my monitor can only display 240 Hz, and my mouse is performing at 1000 Hz, it already means that 3 out of 4 mouse updates are being skipped by the monitor. What more with mice that are polling at 4k or 8k Hz.

On another note, my mice (Razer Viper Mini and Logitech G02} don't even poll at a constant 1000 Hz unless I'm moving/flicking it fast. Most of the time it hovers around 200-700 Hz, averaging about 500-600Hz. How fast would you have to flick a mouse to reach 4k or 8k Hz polling rate?

deama
Posts: 224
Joined: 07 Aug 2019, 12:00

Re: Polling Rate among Gaming Mice Analysis: 1000hz Is Not Enough

Post by deama » 23 Jul 2020, 05:23

konchy wrote:
22 Jul 2020, 14:08
Would like some clarification.
Why would all these high refresh rate mice matter when our monitors can't even display them?

For instance, if my monitor can only display 240 Hz, and my mouse is performing at 1000 Hz, it already means that 3 out of 4 mouse updates are being skipped by the monitor. What more with mice that are polling at 4k or 8k Hz.
It's mostly to do with flick shots or quick camera movements, e.g. in an fps game, if you do a 180 degree spin, you have to slice the 1000hz polling rate between the each of the 180 degrees; whereas with 8000hz this slicing ends up making it much smoother because you have more hz.

konchy
Posts: 10
Joined: 11 Dec 2018, 04:54

Re: Polling Rate among Gaming Mice Analysis: 1000hz Is Not Enough

Post by konchy » 29 Jul 2020, 12:00

But we won't be able to see or feel that additional smoothness because our monitors aren't updating as fast as our mice is?

To truly take advantage of a 4k Hz mouse we probably would have to wait until we get pcs that can consistently output 4000 fps on a 4000 Hz monitor

User avatar
Chief Blur Buster
Site Admin
Posts: 8831
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: Polling Rate among Gaming Mice Analysis: 1000hz Is Not Enough

Post by Chief Blur Buster » 29 Jul 2020, 13:43

konchy wrote:
29 Jul 2020, 12:00
To truly take advantage of a 4k Hz mouse we probably would have to wait until we get pcs that can consistently output 4000 fps on a 4000 Hz monitor
FALSE
There is a beat frequency effect problem!
You don't need 4000fps. There's microstutter problems from Hz-vs-poll aliasing effects too!
I've got a photo here, of this very example:

Image

Today, it's even worse at 240Hz, 360Hz and 480Hz, as refresh rates starts to get closer to Hz. Read below!

The aliasing effects change with sync technology setting, blur reduction setting, VRR setting, and all, but it's all there at lower poll Hz. 1000Hz isn't enough, I already see limiting factors during strobed operation for 1000Hz polling, with my very own eyes... ... ...
konchy wrote:
29 Jul 2020, 12:00
But we won't be able to see or feel that additional smoothness because our monitors aren't updating as fast as our mice is?
Actually, yes we can. There's a beat frequency effect between poll rate and frame rate.

As refresh rates start to enroach poll rates, the beat-frequency effects are becoming more amplified.

There's a more noticeable beat-frequency audio effect between a 499Hz audio tone and a 1000Hz audio tone, than a 101Hz audio tone and a 1000Hz audio tone. Not as noticeable as a 999Hz audio tone and a 1000Hz audio tone, but definitely 499Hz-vs-1000Hz is more noticeable han 101Hz-vs-1000Hz. Even though I am deafie, I am quite aware of how stronger beat frequencies become when they are closer-together in octaves.

Likewise, you see those raising refresh rates and frame rates, combined with falling motion blur (that normally hides tiny microstutters)? Bingo, Einstien.

The same effect happens to microstutter harmonics / beat frequencies. Some of the harmonic microstutter mathematics are explained in Blur Busters Law: The Amazing Journey To Future 1000 Hz Displays

During VRR operation and low-motion-blur operation (1ms MPRT, strobe backlights), these microstutters are also additionally made more visible, and even sub-millisecond microstutters will become visible (The Amazing Human Visible Feats Of The Millisecond) as screen resolution becomes higher simulaneously with much lower motion blur.

For example a retina 4K 1000Hz display with a 1000fps GPU, and given sufficiently wide FOV and fast motion speed (The Vicious Cycle Effect of high resolution & refresh bringing tinier motion defects to human-visibility noise floor) in this theoretical case, a 0.5ms microstutter can become human noticeable.

Think in simple Blur Buster Law mathematics. Two screenwidths per second motion, about 8000 pixels per second, times 0.5ms stutter equals 4 pixel stutterjump. If the display's motion blur is sufficiently tiny (blur no longer hides the microstutter), and stutterwidth sufficiently big, it is a much more human visible stutter if the stutter jump width isn’t retina angular resolution. It’s an even bigger problem for VR, with its 180 degree FOV creating more opportunity for motion issues to be noticed before it falls off the screen edge.

Simultaneously, for motion-quality purists, this is why I recommend 1600dpi or 3200dpi settings for smoother mouse slowturns (I now use the 3200dpi setting of a 12000dpi sensor). Most esports still use 400dpi, 800dpi and sometimes 1600dpi, partly because past mice often degrade in flick accuracy at the higher DPIs, and the cursor was too fast at Windows Desktop, but newer sensors now can do higher DPI reliably if you keep your mouse feet clean and use a proper fine-textured mousepad that the sensor tracks well on. Use DPI-switching features to slow down cursors when exiting games to Windows. High DPI is the only way to get "TestUFO smoooth slow mouseturns" in games.

Be mindful, the computer can be the limiting factor (i.e. its ability to accept >1000Hz smoothly without lag), but the polls will be better-aligned with frametimes -- that's the important part. Ideally, one needs sub-millisecond mouse input alignments for smooth motion -- polltime:frametime, for good low-blur and good VRR operation.

If you understand audio harmonics then stutter harmonics is easy to learn! (Including advanced/deep/glorified versions of "118fps at 120Hz creates 2 stutter per second").

The higher the refresh rates, the microstutter beat-frequency problem with current mouse poll rates starts to become a weak link in the refresh rate race. It's already happening today.

If you are a software developer, please correctly study up on the "gametime-vs-frametime-vs-photontime" relativity jittering problems.

It is all too easy to make accidental false scientific assumptions about specific milliseconds not mattering. Although that thread is not about mouse, the mouse is a quantization error margin to the frametime:photontime factor, since mousepolltime:gametime time-relativity jitter creates human-visible microstutter that's often fixable by raising poll Hz.
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

       To support Blur Busters:
       • Official List of Best Gaming Monitors
       • List of G-SYNC Monitors
       • List of FreeSync Monitors
       • List of Ultrawide Monitors

Brainlet
Posts: 62
Joined: 30 May 2020, 12:39
Contact:

Re: Polling Rate among Gaming Mice Analysis: 1000hz Is Not Enough

Post by Brainlet » 29 Jul 2020, 15:32

Chief Blur Buster wrote:
29 Jul 2020, 13:43
Simultaneously, for motion-quality purists, this is why I recommend 1600dpi or 3200dpi settings for smoother mouse slowturns (I now use the 3200dpi setting of a 12000dpi sensor). Most esports still use 400dpi, 800dpi and sometimes 1600dpi, partly because past mice often degrade in flick accuracy at the higher DPIs, and the cursor was too fast at Windows Desktop, but newer sensors now can do higher DPI reliably if you keep your mouse feet clean and use a proper fine-textured mousepad that the sensor tracks well on. Use DPI-switching features to slow down cursors when exiting games to Windows. High DPI is the only way to get "TestUFO smoooth slow mouseturns/strafes" in games.

Be mindful, the computer can be the limiting factor (i.e. its ability to accept >1000Hz smoothly without lag), but the polls will be better-aligned with frametimes -- that's the important part. Ideally, one needs sub-millisecond mouse input alignments for smooth motion -- polltime:frametime, for good low-blur and good VRR operation.
This is why Windows 7 is great. It's capable of sustaining really solid polling (even under load):

Image

This is the surface of a Glorious Helios:

Image

and it indeed supports higher DPIs in the 4 digit range (2000-3000+). As a generally safe range for most modern mice and pad surfaces I recommend 1600-2000. Some sensors and firmware implementations might do more, some might do less. As a nice extra, higher dpi also decreases input lag noticeably.
Starting point for beginners: PC Optimization Hub

senny22
Posts: 93
Joined: 03 May 2019, 17:40

Re: Polling Rate among Gaming Mice Analysis: 1000hz Is Not Enough

Post by senny22 » 30 Jul 2020, 09:15

Brainlet wrote:
29 Jul 2020, 15:32
[quote="Chief Blur Buster" post_id=55421

and it indeed supports higher DPIs in the 4 digit range (2000-3000+). As a generally safe range for most modern mice and pad surfaces I recommend 1600-2000. Some sensors and firmware implementations might do more, some might do less. As a nice extra, higher dpi also decreases input lag noticeably.
How does increasing dpi reduce input lag? I've never read that before and can't do how that would be the case.

User avatar
MaxTendency
Posts: 59
Joined: 22 Jun 2020, 01:47

Re: Polling Rate among Gaming Mice Analysis: 1000hz Is Not Enough

Post by MaxTendency » 30 Jul 2020, 10:04

senny22 wrote:
30 Jul 2020, 09:15
How does increasing dpi reduce input lag? I've never read that before and can't do how that would be the case.
Image

Image

Sources are linked into the pics. Idk about the 0 smoothing claim, some say there's smoothing above 2k dpi even on newer sensors.

Post Reply