cstoquer / audio-encoder Goto Github PK

Working with audio-encoder, I discovered an issue as described here. My colleague reported it and luckily the maintainer of lamejs package released a new version 1.2.1 to fix that.

It would be great if you could update the dependency to the most recent version (1.2.0 -> 1.2.1) so that I don't have to emulate the bugfix manually each time I build.

Would be great if WAV encoder would support at least 22050 and 48000.

Hi there,

the encoder to mp3 doesn't work anymore.

Here is my code :

import encode from 'audio-encoder';

export const bufferToMp3 = (buffer) => {
        return new Promise((res, rej) => {
            encode(buffer, 320, null, blob => res(blob));
        })
    }

An error is thrown :

Lame.js:135 Uncaught (in promise) ReferenceError: MPEGMode is not defined
    at lame_init_old (Lame.js:135:1)
    at Lame.lame_init (Lame.js:229:1)
    at new Mp3Encoder (index.js:100:1)
    at encodeMp3 (encodeMp3.js:49:1)

Currently encodeWav expects a callback that will be passed the blob after a 30ms timeout.

Why is the timeout necessary in a function that appears to be otherwise fully synchronous? I tried swapping the setTimeout for a simple return blob; and it works fine.

Assuming there's some asynchronicity that has to be waited for, why not return a promise that resolves with the blob?

In my use case, I'm encoding many different chunks of audio and then putting the blobs into a zip file for download, so I have to wait for all blobs to be returned. The callback was not a practical solution for me, but maybe other people using this library typically don't have this problem and are okay with the callback solution.

I could submit a PR with a very simple change to promisify encodeWav, but I wanted to understand the library's rationale first.

Thanks!

Support streaming Float32Array input.

Currently it appears the only valid input to audioEncoder() is a Web Audio API AudioBuffer.

AudioWorklet and MediaStreamTrackProcessor (https://w3c.github.io/mediacapture-transform/) provides a means to get "real-time" audio data as Float32Arrays.

We should be able to stream Float32Arrays to audioEncoder() and later call a method to encode the data once required input is achieved.

Hi this is great. Can you include in the readMe what is possible to get out from onProgress and an example.
I would like to make a encode progress bar. How do I get progress data out?