Is there any way of exporting audio contained in an audio context?

[ddvxx]

I am developing a DAW app and I wonder how can I export the composition i am making in the timeline.

[maciejmakowski2003]

Currently you can design your composition via OfflineAudioContext like in one of our examples in the app, but the output of that operation is AudioBuffer - raw numerical data, so at that point our support ends. We will add your request to our road map. There is a thread about the same issue in Web Audio API

[marksyzm]

I achieved this using the buffer features from the wavefile package:
https://www.npmjs.com/package/wavefile

Once I collected all the audio buffers into an array, I could then create a writeable audio file:

const waveFile = new WaveFile()
  waveFile.fromScratch(
    channelsPerFrame,
    sampleRate,
    bitsPerChannel.toString(),
    audioBuffersRef.current,
  )

  await RNFetchBlob.fs.writeFile(
    waveFilePath,
    waveFile.toBase64(),
    'base64',
  )

I made an ffmpegkit wrapper like this:

export const useFfmpeg = () => {
  const ffmpeg = useCallback(async (command: string) => {
    await new Promise(async (resolve, reject) => {
      await FFmpegKit.executeAsync(
        command,
        async session => {
          const returnCode = await session.getReturnCode()
          console.log(session, returnCode)
          if (ReturnCode.isSuccess(returnCode)) {
            console.log('Audio Converter: Success')
            resolve(null)
          } else if (ReturnCode.isCancel(returnCode)) {
            console.log('Audio Converter: Cancelled')
            reject(new Error('Audio Converter: Cancelled'))
          } else {
            console.log('Audio Converter: Failed')
            const output = await session.getOutput()
            console.log(`Audio Converter: Failed\n${output}`)
            reject(output)
          }
        },
        () => {},
        () => {},
      )
    })
  }, [])

  return convertAudio
}

The files can then be combined using ffmpegkit, like this:

const inputs = [{
    input: voiceMp3FilePath,
    delay: voiceTrackHistory?.cursorStart,
  }, {
    input: backingMp3FilePath,
    delay: backingTrackHistory?.cursorStart,
  }]


await ffmpeg(
  `${inputs
    .map(({input}) => `-i ${input}`)
    .join(' ')} -filter_complex "${inputs
    .map(
      ({delay}, index) =>
        `[${index}]adelay=delays=${delay}ms:all=1[s${index + 1}];`,
    )
    .join(' ')} ${
    inputs.length > 1
      ? '[s1][s2]amix=inputs=2[aout]" -map [aout]'
      : '" -map [s1]'
  } ${path}`,
);

[jerryseigle]

Thanks, this is much needed.

I am also building a DAW. I was thinking it would be nice to have a multitrack node. We currently can do this but it would clean up the code a lot if we could just provide an array of files to a node and they are all played same time

[maciejmakowski2003]

We are planning to add support for EncoderNode, that will save processed audio directly to the file 📁 . Should be supported in 0.12.0