mastouille.fr est l'un des nombreux serveurs Mastodon indépendants que vous pouvez utiliser pour participer au fédiverse.
Mastouille est une instance Mastodon durable, ouverte, et hébergée en France.

Administré par :

Statistiques du serveur :

585
comptes actifs

#webaudio

0 message0 participant0 message aujourd’hui

About to release this JavaScript library 📚

It's a clock/scheduler which schedules notes for Web MIDI or Web Audio.

It can drive a MIDI clock 100% accurately, just like professional software such as Logic Pro or Ableton can do.

There’s no slowdown or drift as it uses an Audio Worklet to schedule using the audio thread 🧵

A répondu dans un fil de discussion

@sarajw

UPDATE: These libraries abstract away much of the complexity of the Web Audio API.

tonejs.github.io/

howlerjs.com/

alemangui.github.io/pizzicato/

The Pizzicato basic example (in the screenshot below) is helpfful because it clearly highlights the contrast in complexity against Web Audio API code that accomplishes exactly the same.

WaveSurfer is a bit different, and I'm including it because it has a very rich suite of examples
wavesurfer.xyz/examples/

I could build a reproducible test case of how Firefox is ignoring my linearRampToValueAtTime for gain nodes
soledadpenades.com/posts/2024/

I don't know why I get so excited!! But I am.

While I try to file the bug, my post explains how I captured the output of the audio context using MediaRecorder, and made it so that you can download the files straight from the browser to visualise them in Audacity. Even in Safari 😆

Suite du fil

Now, if anyone would have some pointers how to easily get a WebAudio AudioWorklet going with @vite, then that example could be updated to produce realtime audio and I could listen to that for hours in the background... 🥰🎧

AFAICT support for worklets is still outstanding and so far I only managed it via pre-compiling/bundling the worklet code in a separate build step, but it's not great to work with that way during dev...

8 years ago I partially started going back to my roots to begin working on a number of #Forth based DSLs & REPLs for livecoding... not just for audio/music, but also #GenerativeArt, #shaders, #geometry and GUI/layout generation. Even today, I still think Forth-like #concatenative languages have unique potential & freedoms and are severely under-explored in all of those domains mentioned. Even though all of the examples in this thread are early, pretty rough prototypes, I'd like to share some videos/screenrecs of my previous research, maybe also to inspire others exploring this approach of programming more...

First off, a #WebAudio prototype from 2015, made in the browser REPL at forth.thi.ng (audio starts at ~2:15)

There's only a single audio buffer used here (boilerplate setup for that in the first 2 mins of the video) and all subsequent operations just write to that buffer or manipulate its contents...

1/4