node-record-lpcm-16

Records a 16-bit signed-integer linear pulse modulation code WAV audio file.

This module uses Node.js streams to minimize memory usage and optimize speed, perfect for embedded devices and "the internet of things".

These audio files are fully compatible with both the Google Speech to Text API (v2) and the Wit.ai Speech API.

Installation

npm install node-record-lpcm16

Dependencies

Generally, running npm install should suffice.

This module requires you to install SoX and it must be available in your $PATH.

For Mac OS

brew install sox

For most linux disto's

sudo apt-get install sox libsox-fmt-all

For Windows

download the binaries

Options

sampleRate    : 16000  // audio sample rate
channels      : 1      // number of channels
threshold     : 0.5    // silence threshold (rec only)
thresholdStart: null   // silence threshold to start recording, overrides threshold (rec only)
thresholdEnd  : null   // silence threshold to end recording, overrides threshold (rec only)
silence       : '1.0'  // seconds of silence before ending
verbose       : false  // log info to the console
recordProgram : 'rec'  // Defaults to 'rec' - also supports 'arecord' and 'sox'
device        : null   // recording device (e.g.: 'plughw:1')

Please note that arecord might not work on all operating systems. If you can't capture any sound with arecord, try to change device (arecord -l).

Usage

var record = require('node-record-lpcm16')
var fs = require('fs')

var file = fs.createWriteStream('test.wav', { encoding: 'binary' })

record.start({
  sampleRate : 44100,
  verbose : true
})
.pipe(file)

The library will automatically attempt to stop when it encounters silence, if not you can stop the recording manually.

var record = require('node-record-lpcm16')
var fs = require('fs')

var file = fs.createWriteStream('test.wav', { encoding: 'binary' })

record.start().pipe(file)

// Stop recording after three seconds
setTimeout(function () {
  record.stop()
}, 3000)

This module uses Node.js streams, if you're unfamiliar with them I'd suggest reading Substack's excellent stream handbook.

Example

Here's how you can write your own Siri in just 10 lines of code.

var rec = require('node-record-lpcm16')
var request = require('request')

var witToken = process.env.WIT_TOKEN; // get one from wit.ai!

exports.parseResult = function (err, resp, body) {
  console.log(body)
}

rec.start().pipe(request.post({
  'url'     : 'https://api.wit.ai/speech?client=chromium&lang=en-us&output=json',
  'headers' : {
    'Accept'        : 'application/vnd.wit.20160202+json',
    'Authorization' : 'Bearer ' + witToken,
    'Content-Type'  : 'audio/wav'
  }
}, exports.parseResult))

elbywan / node-record-lpcm16