Hi all, and thank you so much for this great software that you're making available for all of us :)
I was thinking about a way to eliminate the delay completely (sort of) while still retaining the desired smoothness, and that's how it would work.
The algorithm I had in mind would get the raw pen events, print them (retaining the jags), and at the same time save their info in memory.
Once N events have been cached and printed, another thread (or process) in parallel would start, compute what it needs to compute with those N events and redraw only that part of the screen. All while the main thread/process still collects and draws the raw strokes.
Doing this way, the latency should roughly stay the same and the jags should only be present for a small amount of time after they've been drawn.
I don't know how feasible that would be, but in theory it should work, and it's a solution I would personally adopt in order to have a super clean result without sacrificing latency (if it takes 1s to update what I've written before I don't mind, I'm focused on what I'm writing right now and I'd like a super clean final result)
Hi @giorgio-arena, this would indeed be the best solution! Unfortunately, we would need to have access to xochitl's source code to do that, which reMarkable is not releasing (yet?).
Hi all, and thank you so much for this great software that you're making available for all of us :)
I was thinking about a way to eliminate the delay completely (sort of) while still retaining the desired smoothness, and that's how it would work.
The algorithm I had in mind would get the raw pen events, print them (retaining the jags), and at the same time save their info in memory. Once N events have been cached and printed, another thread (or process) in parallel would start, compute what it needs to compute with those N events and redraw only that part of the screen. All while the main thread/process still collects and draws the raw strokes.
Doing this way, the latency should roughly stay the same and the jags should only be present for a small amount of time after they've been drawn.
I don't know how feasible that would be, but in theory it should work, and it's a solution I would personally adopt in order to have a super clean result without sacrificing latency (if it takes 1s to update what I've written before I don't mind, I'm focused on what I'm writing right now and I'd like a super clean final result)
Let me know what you think and thanks again :)