Table of Contents

1. Guide
   1. Get The Sample Rate
   2. Process the Audio with audiowaveform
   3. Import The Processed Audio Into Motion Canvas

Audio visualisation with Motion Canvas and audiowaveform.

Guide

For the purposes of this guide, I will use src/data/test.mp3.

Get The Sample Rate

You can acquire the sample rate using ffprobe:

ffprobe src/data/test.mp3

Which should yield something like:

ffprobe version 5.1.3 Copyright (c) 2007-2022 the FFmpeg developers
  built with gcc 12.2.0 (GCC)
  configuration: ...
Input #0, mp3, from 'src/data/test.mp3':
  Duration: 00:00:03.79, start: 0.025057, bitrate: 85 kb/s
  Stream #0:0: Audio: mp3, 44100 Hz, mono, fltp, 85 kb/s

Here, we can see that test.mp3 is at 44100 Hz, but we do not need anywhere close to this resolution. I find that about 5 samples a second works for me, so I would use 44100 / 5 = 8820 for the next step.

Process the Audio with audiowaveform

audiowaveform -i src/data/test.mp3 -o src/data/test.mp3.json -z 8820

The -z 8820 argument refers to the level of “zoom,” or how many samples should be used for a single entry into the new JSON file. This should be calculated based upon the sample rate.

Import The Processed Audio Into Motion Canvas

See src/scenes/visualiser.tsx for example usage. src/index.ts provides an audiowaveformAmplitudes function to more easily use the data in Motion Canvas; this function takes an audiowaveform JSON and provides an array of amplitudes between 0 and 1, which can then be used to drive whatever animation.

For example, I can create an opacity signal:

import { audiowaveformAmplitudes } from 'motion-canvas-audio-visualiser';
import audioJSON from "../data/test.mp3.json";
const amplitudes = audiowaveformAmplitudes(audioJSON);
const opacity = createSignal(amplitudes[0]);

And use that opacity for a ring around an image:

export default makeScene2D(function* (view) {
  ...

  view.add(
    <Circle
      width={256}
      height={256}
      stroke={createSignal(() => new Color(0x23a559).alpha(opacity()))}
      lineWidth={32}
    >
      <Img src={thumb} width={256} radius={128} />
    </Circle>
  );
  ...
};

Which can then be updated according to how long each timestep is based on the sample rate:

for (let i = 0; i < amplitudes.length; ++i) {
    yield* opacity(
      amplitudes[i] * 3,
      1 / 5
    );
  }

In that example, I am using 1 / 5 for the duration, because each timestep takes 1/5 seconds, and I have multiplied the amplitude by 3 such that it will be maximally opaque whilst only achieving a third of the maximum amplitude.