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

[NDUS]

I made a 1-minute video which superimposes 500hz/1000hz/8000hz polling rates against some gameplay footage.

https://www.youtube.com/watch?v=mcsbDDyfLLU

It turns out that when you either:

1) flick very quickly at a target that is nearby,
or
2) flick slowly at a target that is far away,

then 1000hz is inadequate.

Here's some higher quality stills:





In case you were curious, the way you get 8000hz polling is by using this driver: https://www.overclock.net/forum/375-mic ... 00-hz.html

To get 8000hz polling requires either a mouse with MLT04 sensor (which is quite old, and bad except for its 8khz polling support - it maxes out at 1.5m/s) or a mouse with firmware specifically designed to allow 8000hz polling (currently just the Zaunkoenig M1K.)

Here is a fun fact: it's actually possible for USB devices to poll at 8000hz without any additional work, but the input device has to be something called a "USB High Speed" device which apparently requires an expensive 4+ layer PCB design (hence why no peripherals company bothers.) Why you can achieve 8khz polling on a normal PCB with this driver is beyond me, but there you go.

[JordanV]

Faster is better but I don't think characterizing 1000Hz as "not enough" or "far too little" is true when it comes to the potential to perform a flick shot in a single frame at 60fps. Let's assume a 60fps environment with negligible lag. A player calculates they need to move the mouse x pixels right and y pixels up to get a flickshot in the next frame. As long as the mouse is accurate, they can do this even with a 60Hz polling rate. In an environment with lag where refresh, fps, and polling rate are not in sync, this won't be the case and improving those rates will provide better consistency as well as better information to make such a shot more likely.

I don't know what the upper limit would be for increasing mouse polling rates to not matter for humans if fps and refresh are infinite but given the state of electronics these days, you can have multi MHz or GHz polling rates with a PCIe mouse which, I am going to go out on a limb here, ought to be enough for anyone.

[NDUS]

https://www.youtube.com/watch?v=OLmADXyssZo

Here is what 500hz mouse polling looks like to the naked eye. These are not even very fast flicks.
You can even see a polling error/microstutter in the video!



https://www.youtube.com/watch?v=acGBZv1891Y < 125hz polling in Overwatch

[NDUS]

A probably more realistic aiming scenario in Overwatch:

https://www.youtube.com/watch?v=AcU8kb8NOL4

[Q83Ia7ta]

but who shoot like this? in the middle of the move

[NDUS]

but who shoot like this? in the middle of the move

Well, I do, and I'm pretty good.
It's much easier to merely click at the right time, rather than to stop the entire mouse at the correct moment and then click. Mice are pretty heavy objects with feet made of teflon. And the faster your flick is, the more difficult it becomes to stop the mouse.

I think even if you do stop the mouse before shooting, high polling rate is still crucial. Possibly even more crucial, because there is intense deceleration involved, and the mouse needs to update its position as rapidly as possible to "catch" the deceleration in the very short window right before you fire. But this is not as easy to show in an image.

[hleV]

but who shoot like this? in the middle of the move

Every decent Unreal Tournament player.

[Chief Blur Buster]

but who shoot like this? in the middle of the move

Many Quake gamers in the CRT days could shoot in the middle of a turn without pausing the turn. At least when they upgraded to those great early Intellimouse Explorers and used 3Dfx Voodoo2 SLIs to max-out their motion fluidity at the time. The lack of motion blur of the CRT era made it practical to eye-track enemies mid-turn. Many pro players didn't adapt to LCD-favourable tactics such as stationary-gaze on crosshairs.

Now that 240 Hz is here and 360 Hz monitors are coming (essentially behaving as strobeless ULMB), professional gaming tactics will evolve to include more eye tracking, now there will be fewer stroboscopic effects of finite frame rate.

We are 100% in favour of 8000 Hz mice in this refresh rate race to future retina refresh rates. Not only it is not just a "cross-the-finish-line-effect" (the millisecond that wins an Olympics race even if you can't feel the millisecond), there ARE even human visible benefits of a 1000Hz vs 8000Hz mouse -- the microstutter difference has been tested by Blur Busters to be human-visible, especially in sub-1ms-MPRT strobed situations. While it is not necessary to see the millisecond to win by the millisecond, this is one of those situations where the millisecond surprises. (Much like unexpected benefits of retina refresh rates surprises us -- that there are actually still human visible benefits -- ASUS and NVIDIA now has a roadmap to 1000Hz displays, as reported by PC Magazine and my direct conversations with reps at CES 2020).

[BTRY B 529th FA BN]

.

[NDUS]

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