I wrote a post about JerkMon a few days ago. I was already a tad obsessed with accurate observations of the behaviour of my pointer. My fixation has only gotten worse since, as I started playing at 480 Hz.
The latest result is a program which captures VSync frame timing, mouse events through buffered raw inputs in Microsoft Windows, and exposes the collected data over WebSocket.
Made JerkMonPro to observe it in great detail. Unlike JerkMon, its portable counterpart, it can afford to feature histograms.
What I've found is that there are at least 3 types of USB3 ports on my workstation:
- Some where pauses exceed 10ms every few seconds,
- Some where pauses exceed 2ms multiple times per second, and 4ms every few seconds;
- Some where pauses never exceed 1ms.
That's now been confirmed with 2 different mice, the same USB cables (tried a few), and the same receiver placement.
For frame-perfect inputs at high refresh rates, this seemingly innocuous choice appears critical.
Edit: Produced some analysis of USB controllers in a new post.