3 ms·

Nice to see new art work in browsers! Some comments pointed out that there are latency issues, which are mainly caused by tonejs. I used tonejs a few years ago

Nice to see new art work in browsers! Some comments pointed out that there are latency issues, which are mainly caused by tonejs. I used tonejs a few years ago for my first live coding project: https://quaverseries.web.app/ https://quaverseries.web.app/ Interestingly, I also used the same interactive rhythm prompts as op in the page title, but the timing was not accurate at all. See this classic article: https://web.dev/articles/audio-scheduling https://web.dev/articles/audio-scheduling Later, in order to refine my idea, I used rust + wasm and made https://glicol.org/ https://glicol.org/. It basically solved the audio latency problem, and I also made js bindings: https://glicol.js.org/ https://glicol.js.org/ I'd love to get as much feedback as possible. I feel that this sequencer is particularly suitable for glicol's seq syntax.

strickinato 2y ago
Oooh!! I've played a bunch of glicol.org - but had no idea about the js bindings! Might I recommend you put a link to `glicol.js.org` from your main site. I might have given this a go instead of tone.js on a recent project.
- chaosprint 2y ago
  Thanks for your suggestion, I will add it to the next version. I am currently rebuilding the homepage with solidjs.
nox101 2y ago
You don't need to use rust + wasm to "solve the latency problem" but it's cool that you solved it for yourself and made a cool library. https://musiclab.chromeexperiments.com/Song-Maker/song/6339181125697536 https://musiclab.chromeexperiments.com/Song-Maker/song/63391...
- chaosprint 2y ago
  Partially agree. It's not necessarily Rust. But C/C++/Rust -> wasm -> AudioWorklet <- SAB -> Main Thread is basically the ceiling of web audio. Of course, there is still room for improvement in interaction. Your example is a project from 2018, also based on tonejs. The main problem is not latency, but the impact of GC and the main thread on the clock.
riffraff 2y ago
random note: on glicol.org you have "optinal semicolon" which I presume is "optional". Cool stuff anyway :)
- mejutoco 2y ago
  Other small typos on the js page in case it helps: recommanded, reletively.
thelastparadise 2y ago
I'm getting buffer underrun city on glycol it sounds like crap.
iamsaitam 2y ago
Can you explain further or point to, how exactly wasm allows more precise timing? I also agree that the timing with JS and Webaudio is really poor, not as bad as MIDI with 24 PPQN, but I think a few years back from my experiments, the max I could get was a fluctuating 96PPQN.

strickinato 2y ago

Oooh!! I've played a bunch of glicol.org - but had no idea about the js bindings! Might I recommend you put a link to `glicol.js.org` from your main site. I might have given this a go instead of tone.js on a recent project.

chaosprint 2y ago
Thanks for your suggestion, I will add it to the next version. I am currently rebuilding the homepage with solidjs.

nox101 2y ago

You don't need to use rust + wasm to "solve the latency problem" but it's cool that you solved it for yourself and made a cool library. https://musiclab.chromeexperiments.com/Song-Maker/song/6339181125697536 https://musiclab.chromeexperiments.com/Song-Maker/song/63391...

chaosprint 2y ago
Partially agree. It's not necessarily Rust. But C/C++/Rust -> wasm -> AudioWorklet <- SAB -> Main Thread is basically the ceiling of web audio. Of course, there is still room for improvement in interaction. Your example is a project from 2018, also based on tonejs. The main problem is not latency, but the impact of GC and the main thread on the clock.

riffraff 2y ago

random note: on glicol.org you have "optinal semicolon" which I presume is "optional". Cool stuff anyway :)

mejutoco 2y ago
Other small typos on the js page in case it helps: recommanded, reletively.

thelastparadise 2y ago

I'm getting buffer underrun city on glycol it sounds like crap.

iamsaitam 2y ago

Can you explain further or point to, how exactly wasm allows more precise timing? I also agree that the timing with JS and Webaudio is really poor, not as bad as MIDI with 24 PPQN, but I think a few years back from my experiments, the max I could get was a fluctuating 96PPQN.