waveform-data.js

waveform-data.js is a JavaScript library for creating zoomable representations of audio waveforms to enable visualisation of audio content.

waveform-data.js is part of a BBC R&D Browser-based audio waveform visualisation software family:

audiowaveform: C++ program that generates waveform data files from MP3 or WAV format audio.
audio_waveform-ruby: A Ruby gem that can read and write waveform data files.
waveform-data.js: JavaScript library that provides access to precomputed waveform data files, or can generate waveform data using the Web Audio API.
peaks.js: JavaScript UI component for interacting with waveforms.

We use these projects within the BBC in applications such as the BBC World Service Radio Archive and browser-based editing and sharing tools for BBC content editors.

Install

Use npm to install waveform-data.js, for both Node.js and browser-based applications:

npm install --save waveform-data

Usage and examples

dist/waveform-data.js is delivered as a UMD module so it can be used from a <script> tag, or as a RequireJS or CommonJS module.

Importing waveform-data.js

Using a script tag

Simply add waveform-data.js in a script tag in your HTML page:

<!DOCTYPE html>
<html>
  <body>
    <script src="/path/to/waveform-data.js"></script>
    <script>
      var waveform = new WaveformData(...);
    </script>
  </body>
</html>

Using ES6

import WaveformData from 'waveform-data';

Using RequireJS

define(['WaveformData'], function(WaveformData) {
  // ...
});

Using CommonJS (Node.js)

const WaveformData = require('waveform-data');

Receive binary waveform data

You can create and initialise a WaveformData object from waveform data in either binary or JSON format, using the Fetch API, as follows.

Binary format

Use audiowaveform to generate binary format waveform data, using a command such as:

audiowaveform -i track.mp3 -o track.dat -b 8 -z 256

Copy the waveform data file track.dat to your web server, then use the following code in your web application to request the waveform data:

fetch('https://example.com/waveforms/track.dat')
  .then(response => response.arrayBuffer())
  .then(buffer => WaveformData.create(buffer))
  .then(waveform => {
    console.log(`Waveform has ${waveform.channels} channels`);
    console.log(`Waveform has length ${waveform.length} points`);
  });

JSON format

Alternatively, audiowaveform can generate waveform data in JSON format:

audiowaveform -i track.mp3 -o track.json -b 8 -z 256

Use the following code to request the waveform data:

fetch('https://example.com/waveforms/track.json')
  .then(response => response.json())
  .then(json => WaveformData.create(json))
  .then(waveform => {
    console.log(`Waveform has ${waveform.channels} channels`);
    console.log(`Waveform has length ${waveform.length} points`);
  });

Using the Web Audio API

You can also create waveform data from audio in the browser, using the Web Audio API.

As input, you can either use an ArrayBuffer containing the original encoded audio (e.g., in MP3, Ogg Vorbis, or WAV format), or an AudioBuffer containing the decoded audio samples.

Note that this approach is generally less efficient than pre-processing the audio server-side, using audiowaveform. To avoid blocking the browser's UI thread, he audio will be processed using a Web Worker, if supported by the browser.

const audioContext = new AudioContext();

fetch('https://example.com/audio/track.ogg')
  .then(response => response.arrayBuffer())
  .then(buffer => {
    const options = {
      audio_context: audioContext,
      array_buffer: buffer,
      scale: 128
    };

    return new Promise((resolve, reject) => {
      WaveformData.createFromAudio(options, (err, waveform) => {
        if (err) {
          reject(err);
        }
        else {
          resolve(waveform);
        }
      });
    });
  })
  .then(waveform => {
    console.log(`Waveform has ${waveform.channels} channels`);
    console.log(`Waveform has length ${waveform.length} points`);
  });

If you have an AudioBuffer containing decoded audio samples, e.g., from AudioContext.decodeAudioData then you can pass this directly to WaveformData.createFromAudio:

const audioContext = new AudioContext();

audioContext.decodeAudioData(arrayBuffer)
  .then((audioBuffer) => {
    const options = {
      audio_context: audioContext,
      audio_buffer: audioBuffer,
      scale: 128
    };

    return new Promise((resolve, reject) => {
      WaveformData.createFromAudio(options, (err, waveform) => {
        if (err) {
          reject(err);
        }
        else {
          resolve(waveform);
        }
      });
    });
  })
  .then(waveform => {
    console.log(`Waveform has ${waveform.channels} channels`);
    console.log(`Waveform has length ${waveform.length} points`);
  });

Drawing a waveform image

Once you've created a WaveformData object, you can use it to draw a waveform image, using the Canvas API or a visualization library such as D3.js.

Canvas example

const waveform = WaveformData.create(raw_data);

const scaleY = (amplitude, height) => {
  const range = 256;
  const offset = 128;

  return height - ((amplitude + offset) * height) / range;
}

const ctx = canvas.getContext('2d');
ctx.beginPath();

const channel = waveform.channel(0);

// Loop forwards, drawing the upper half of the waveform
for (let x = 0; x < waveform.length; x++) {
  const val = channel.max_sample(x);

  ctx.lineTo(x + 0.5, scaleY(val, canvas.height) + 0.5);
}

// Loop backwards, drawing the lower half of the waveform
for (let x = waveform.length - 1; x >= 0; x--) {
  const val = channel.min_sample(x);

  ctx.lineTo(x + 0.5, scaleY(val, canvas.height) + 0.5);
}

ctx.closePath();
ctx.stroke();
ctx.fill();

D3.js example

const waveform = WaveformData.create(raw_data);
const channel = waveform.channel(0);
const layout = d3.select(this).select('svg');
const x = d3.scale.linear();
const y = d3.scale.linear();
const offsetX = 100;

const min = channel.min_array();
const max = channel.max_array();

x.domain([0, waveform.length]).rangeRound([0, 1024]);
y.domain([d3.min(min), d3.max(max)]).rangeRound([offsetX, -offsetX]);

const area = d3.svg.area()
  .x((d, i) => x(i))
  .y0((d, i) => y(min[i]))
  .y1((d, i) => y(d));

graph.select('path')
  .datum(max)
  .attr('transform', () => `translate(0, ${offsetX})`)
  .attr('d', area);

In Node.js

You can use waveform-data.js to consume or generate waveform data from a Node.js application, e.g., a web server.

const WaveformData = require('waveform-data');
const express = require('express');
const fs = require('fs');
const app = express();

app.get('/waveforms/:id.json', (req, res) => {
  res.set('Content-Type', 'application/json');

  fs.createReadStream(`path/to/${req.params.id}.json`)
    .pipe(res);
});

The following example shows a Node.js command-line application that requests waveform data from a web API and resamples it to a width of 2000 pixels.

#!/usr/bin/env node

// Save as: app/bin/cli-resampler.js

const WaveformData = require('waveform-data');
const request = require('superagent');
const args = require('yargs').argv;

request.get(`https://api.example.com/waveforms/${args.waveformid}.json`)
  .then(response => {
    const waveform = WaveformData.create(response.body);
    const resampledWaveform = waveform.resample({ width: 2000 });
    const channel = resampledWaveform.channel(0);

    process.stdout.write(JSON.stringify({
      min: channel.min_array(),
      max: channel.max_array()
    }));
});

Usage: ./app/bin/cli-resampler.js --waveformid=1337

Data format

The file format used and consumed by WaveformData is documented here as part of the audiowaveform project.

JavaScript API

Please refer here for full API documentation.

Browser support

Any browser supporting ECMAScript 5 will be enough to use the library - think Array.forEach:

IE9+, Firefox Stable, Chrome Stable, Safari 6+ are fully supported;
IE10+ is required for the TypedArray Adapter;
Firefox 23+ and Webkit/Blink browsers are required for Web Audio API support.

Development

To develop the code, install Node.js and npm. After obtaining the waveform-data.js source code, run npm install to install Node.js package dependencies.

Credits

This library was written by:

Thank you to all our contributors.

This program contains code adapted from Audacity, used with permission.

License

See LICENSE for details.

Contributing

Every contribution is welcomed, either it's code, idea or a merci!

Guidelines are provided and every commit is tested against unit tests using Karma runner and the Chai assertion library.

fducom / waveform-data.js Goto Github PK

waveform-data.js's Introduction