top | item 24296350

(no title)

h3ctic | 5 years ago

Yes, apparently it only has a latency of 21ms

discuss

order

ksd482|5 years ago

21ms sounds really low, which is good of course.

But as humans, would we notice 21ms lag while writing even if we paid close attention?

jm_l|5 years ago

You would definitely notice a 21ms lag while writing. Ideally you want to get below 10ms, but for physical-object-like responsiveness 1ms is the standard. See this old video from Microsoft research which demonstrates

https://www.youtube.com/watch?v=vOvQCPLkPt4

braythwayt|5 years ago

The original Apple Pencil and first generation of iPad Pro boasted of a 20ms lag. That was considered impressive, but perceptible if you look at your writing carefully.

I found it fine for note-taking, but lots of people would still notice the latency, especially those using creative tools where there is a strong feedback loop between what you see and what you draw.

They are now claiming 9ms lag. I suspect this is imperceptible for the use case of note-taking and marking up PDF documents (e.g. highlighting, making notes in the margin).

But then again, 20ms is going to be more than fine for that use case as well.

infogulch|5 years ago

I think you have to get below 10ms for humans to be unable to notice. Drawing is a worst case scenario for exposing screen latency; it's hard to match reality's 0ms and a pen with a fine tip doesn't help hide any of it like a comparatively chunky finger does. That said, 40ms -> 21ms is a really big improvement which could make the experience go from awkward to quite usable.

nomel|5 years ago

I can absolutely tell the difference between an early iPad Pro (20ms) and later (9ms), when using a pencil. With the lower, it feels more like paper, like color is coming out of the pencil.

TheRealPomax|5 years ago

"Close"? Anything over 3~4ms starts to be noticable, and anything over 10ms is "clearly" noticable for folks who expect the response of a pencil on paper.

TheRealPomax|5 years ago

That's not an "only", that's still at least twice what it needs to be to cross the uncanny valley of input latency.