9 ms·
Is there any kind of objective study that shows: 1) What percentages of the population are able to distinguish between hZ rates in terms of normal object smoot
by crazygringo 3y ago
Is there any kind of objective study that shows:
1) What percentages of the population are able to distinguish between hZ rates in terms of normal object smoothness of movement, and
2) Same question but about mouse latency, given lots of hZ headroom so it's measuring specifically the mouse and not display?
At some point improvements don't matter any more, and I'm curious where that point is.
And of course there are always going to be ways to detect any hZ if underlying object motion is fast enough, but I'm not talking about frame rate artifacts like wheels seemingly moving backwards. I mean regular linear motion at regular UX movement speeds.
- Etherlord87 3y ago"able to distinguish" is a very ambiguous description. Imagine the following scenario: you play a video game, where you are a cowboy in a standoff, as soon as your opponent grabs his gun, you need to click to shoot yours. You are punished for a false start, and rewarded based on your reflexes (maybe statistically: the sooner you shoot the more likely you're to win the duel). There is no lower bound to latency, and therefore upper bound to FPS, that helps you statistically. Eventually the resolution of the study (e.g. sample size and therefore the statistical error) will say that "the population are able to distinguish" up to some Hz refresh rate. But when you then increase the resolution of the test (sample size), you will again start to see a difference. Of course I'm not saying a research like that wouldn't be interesting, all I'm arguing is that there's no definitive answer about maximum FPS that makes sense.
- noqc 3y agothis heavily depends on whether or not there is any interpolation, the complication is that you can detect a strobe at essentially arbitrary frequency, provided it is illuminating a moving object, as you will see the rate of the strobe as a spacing between images. Similarly, if you are rendering pong with no interpolation, and the ball is moving fast enough you'll just see copies of the ball. This makes the test somewhat ill posed. Interpolation is also somewhat ill posed, computationally, but if you were to do it very intelligently, or to only a first approximation, I suspect we have more trouble detecting it.
- crazygringo 3y agoIndeed -- is it time to start applying motion blur to scrolling, dragging, and cursor movement? This wouldn't affect hZ for gaming, but would it be an improvement for general UX? Can anyone tell the difference between e.g. scrolling at 960 hZ without motion blur, and scrolling at 120 hZ with the exact right amount of motion blur?
- bob1029 3y agoThink about the human brain's ability to distinguish 2 nearly-identical sound frequencies. Our neural nets are masters of time domain signal analysis. Even the slightest shift in phase or frequency is usually perceptible. I think saccade plays a big role in how we perceive smoothness of motion and could also explain why some people don't really mind 60hz whereas others find it to be an affront to civilization. Just a little bit of difference between human visual systems could explain a large difference in final perception.