Buffering Audio from files

[yard]

In most generic scope we want to be able to work with long files, ideally even when working with AudioBufferSourceNode and other AudioBuffer related nodes.

Potential implementations:

• typical streaming from local/remote files (similar/same as StreamerNode) that could be precisely scheduled. Most likely yet another node to implement that is not standard to web audio api
• StreamableAudioBuffer / BufferableAudioBuffer / MoreSaneNameAudioBuffer
  1. Buffer that would work the same way described above that could be usued within AudioBufferSourceNodes
  2. In streaming from within memory:
     1. We load the still encoded file to the memory and decode from it on the fly
• Other option goes here - feel free to come up with your own solutions

TheMostImportantNode: This issue is potentially huge it should be split into smaller ones once we decide on the approach

The purpouse of the task is to provide design document that will be used as a base for further development

[poneciak57]

Yeah we have plans to support more utilities for such scenarios but we are currently discussing the possible solutions and do not have clear view of this feature yet. Hovewer, if you only need to play the file in the graph (no playback rate modification and so on) you can check out streamer node it works with local files too docs

[Catsvilles]

Would be cool to make it possible to stream-render long audios without OfflineAudioContext, especially on the web. Use case: I would like to take a longer audio file, apply reverb to it and render it directly to user's file system without writing the whole audio into the RAM. Would that be feasible in the future?

[jerryseigle]

Yes this would be great. Also if there was a way to seek, loop, have playback rate, maybe detune. I would also like to see it not decode everything into memory maybe do it in chunks to keep memory usage down which would also allow us to play lots of audile files the same time

[Slavezax]

+1