`OfflineAudioContext` support about faust-loader HOT 5 CLOSED

I had a quick look and I think I know what's creating at least the immediate exception in my app. There's this cache for modules, so that modules that have already been added aren't added again.

This cache doesn't consider the context they were added to, just the URL. So it returns a module that has been initialized to another context in my application (since I use a normal AudioContext at initialization, but am trying to initialize these modules with in OfflineAudioContext afterwards, when the user tries to bounce the tracks).

This explains at least one of the errors I get:

loadProcessor.js:144 DOMException: Failed to construct 'AudioWorkletNode': AudioWorkletNode cannot be created: The node name 'Panner' is not defined in AudioWorkletGlobalScope.

The node is certainly not defined in AudioWorkletGlobalScope, because we skipped that part and returned one that's defined in another audio context!

I'll try to fiddle with this (maybe not today though).

from faust-loader.

Last update, I ended up commenting out the lines that check the cache real quick:

// if (existing) { 
//     return existing; 
// } 

No exceptions! The bounce works! Unfortunately it generates a buffer with no audio; it's a completely mute wav. I guess I just confirmed at least API-wise one could swap the AudioContext for the OfflineAudioContext and get no errors, but getting no audio back seems to be back in the realm of "could OfflineAudioContext acutally be used though?". Logging off for real now though. ✌️

from faust-loader.

I had been eyeing the NPM downloads, but I'm elated to hear for certain that someone besides me is getting use out of the package!

I appreciate the investigation; we should definitely be caching per-context as well. The loader should support N contexts, and absolutely should support Faust nodes in an OfflineAudioContext.

I had built this loader as part of a much larger project so I moved on once it reached MVP level, but you've prompted me to take a look at cleaning it up a bit. Right now I'm updating the FaustAudioProcessorNode code to pass type checks and expand the typing for the returned node. After that I'll look at creating a working test case for multiple (possibly offline) audio contexts.

In the meantime, if you're able to get offline rendering audio output working, PRs are absolutely welcome!

from faust-loader.

I just published v1.2.2, which fixes this. I also added a simple test case rendering audio from a faust node in an OfflineAudioContext and confirmed that it works, at least with a trivial processor.

from faust-loader.

Awesome @Smona! I haven't worked on the project in question since I posted this issue, but I'm looking forward to trying it out. Thanks for looking into this 🙏

from faust-loader.