• 👂 Lets your website communicate using 2 human senses instead of 1
• 🔥 Built with Vue Composition API
• ✅ Supports Vue 2 & 3 using vue-demi
• 🚚 Supports Nuxt 2 & 3 using @vueuse/sound/nuxt
• ⚡️ <1kb bytes (gzip) in your bundle! ~10kb loaded async.
• ✨ Built with TypeScript
• 🗣 Uses a powerful, battle-tested audio utility: Howler.js

If you want to take a quick look at the composable in effect, you should visit the 🌍 demo.

This package is a Vue version of the useSound React hook by joshwcomeau.

Package can be added using yarn:

yarn add @vueuse/sound

Or, use NPM:

npm install @vueuse/sound

This is the most basic example of how fast you can implement sounds in your app using @vueuse/sound.

<template>
  <button @click="play">Play a sound</button>
</template>

<script>
import { useSound } from '@vueuse/sound'
import buttonSfx from '../assets/sounds/button.mp3'

export default {
  setup() {
    const { play } = useSound(buttonSfx)

    return {
      play,
    }
  },
}
</script>

This example is shown in the demo.

This example is shown in the demo.

For the user's sake, browsers don't allow websites to produce sound until the user has interacted with them (eg. by clicking on something). No sound will be produced until the user clicks, taps, or triggers something.

useSound takes advantage of this: because we know that sounds won't be needed immediately on-load, we can lazy-load a third-party dependency.

useSound will add about 1kb gzip to your bundle, and will asynchronously fetch an additional package after load, which clocks in around 9kb gzip.

If the user does happen to click with something that makes noise before this dependency has been loaded and fetched, it will be a no-op (everything will still work, but no sound effect will play). In my experience this is exceedingly rare.

Consider the following snippet of code:

const playbackRate = ref(0.75)

const { play } = useSound('/path/to/sound', { playbackRate })

playbackRate doesn't just serve as an initial value for the sound effect. If playbackRate changes, the sound will immediately begin playing at a new rate. This is true for all options passed to the useSound composable.

The useSound composable takes two arguments:

• A URL to the sound that it wil load
• A config object (ComposableOptions)

It produces an array with two values:

• A function you can call to trigger the sound
• An object with additional data and controls (ExposedData)

When calling the function to play the sound, you can pass it a set of options (PlayOptions).

Let's go through each of these in turn.

When calling useSound, you can pass it a variety of options:

• volume is a number from 0 to 1, where 1 is full volume and 0 is comletely muted.
• playbackRate is a number from 0.5 to 4. It can be used to slow down or speed up the sample. Like a turntable, changes to speed also affect pitch.
• interrupt specifies whether or not the sound should be able to "overlap" if the play function is called again before the sound has ended.
• soundEnabled allows you to pass a value (typically from context or redux or something) to mute all sounds. Note that this can be overridden in the PlayOptions, see below
• sprite allows you to use a single useSound composable for multiple sound effects. See “Sprites” below.

[delegated] refers to the fact that any additional argument you pass in ComposableOptions will be forwarded to the Howl constructor. See "Escape hatches" below for more information.

When calling the composable, you get back a play function as the first item in the tuple:

const { play } = useSound('/meow.mp3')
//      ^ What we're talking about

You can call this function without any arguments when you want to trigger the sound. You can also call it with a PlayOptions object:

• id is used for sprite identification. See “Sprites” below.
• forceSoundEnabled allows you to override the soundEnabled boolean passed to ComposableOptions. You generally never want to do this. The only exception I've found: triggering a sound on the "Mute" button.
• playbackRate is another way you can set a new playback rate, same as in ComposableOptions. In general you should prefer to do it through ComposableOptions, this is an escape hatch.

The composable produces a tuple with 2 options, the play function and an ExposedData object:

const [play, exposedData] = useSound('/meow.mp3')
//                ^ What we're talking about

• stop is a function you can use to pre-emptively halt the sound.
• pause is like stop, except it can be resumed from the same point. Unless you know you'll want to resume, you should use stop; pause hogs resources, since it expects to be resumed at some point.
• isPlaying lets you know whether this sound is currently playing or not. When the sound reaches the end, or it's interrupted with stop or paused, this value will flip back to false. You can use this to show some UI only while the sound is playing.
• duration is the length of the sample, in milliseconds. It will be null until the sample has been loaded. Note that for sprites, it's the length of the entire file.
• sound is an escape hatch. It grants you access to the underlying Howl instance. See the Howler documentation to learn more about how to use it. Note that this will be null for the first few moments after the component mounts.

An audio sprite is a single audio file that holds multiple samples. Instead of loading many individual sounds, you can load a single file and slice it up into multiple sections which can be triggered independently.

There can be a performance benefit to this, since it's less parallel network requests, but it can also be worth doing this if a single component needs multiple samples. See the Drum Machine component for an example.

For sprites, we'll need to define a SpriteMap. It looks like this:

const spriteMap = {
  laser: [0, 300],
  explosion: [1000, 300],
  meow: [2000, 75],
}

SpriteMap is an object. The keys are the ids for individual sounds. The value is a tuple (array of fixed length) with 2 items:

• The starting time of the sample, in milliseconds, counted from the very beginning of the sample
• The length of the sample, in milliseconds.

This visualization might make it clearer:

We can pass our SpriteMap as one of our ComposableOptions:

const { play } = useSound('/path/to/sprite.mp3', {
  sprite: {
    laser: [0, 300],
    explosion: [1000, 300],
    meow: [2000, 75],
  },
})

To play a specific sprite, we'll pass its id when calling the play function:

<button
  @click="play({id: 'laser'})"
>

Howler is a very powerful library, and we've only exposed a tiny slice of what it can do in useSound. We expose two escape hatches to give you more control.

First, any unrecognized option you pass to ComposableOptions will be delegated to Howl. You can see the full list of options in the Howler docs. Here's an example of how we can use onend to fire a function when our sound stops playing:

const { play } = useSound('/thing.mp3', {
  onend: () => {
    console.info('Sound ended!')
  },
})

If you need more control, you should be able to use the sound object directly, which is an instance of Howler.

For example: Howler exposes a fade method, which lets you fade a sound in or out. You can call this method directly on the sound object:

<template>
    <button
      @click={sound.fade(0, 1, 1000)}
    >
      Click to win
    </button>
</template>

<script>
import { useSound } from '@vueuse/sound'

export default {
    setup() {
        const { play, sound } = useSound('/win-theme.mp3')

        return {
            sound
        }
    }
}
</script>

If you are using Vite, you should add the following to your defineConfig options in vite.config.js:

optimizeDeps: {
  exclude: ['vue-demi']
}

If you use Nuxt 2, you must have @nuxt/bridge setup in your project.

Once you installed it, add @vueuse/sound/nuxt dependency to your project.

Add @vueuse/sound/nuxt to the modules section of your nuxt.config:

defineNuxtConfig({
  modules: ['@vueuse/sound/nuxt']
})

Configure your sounds 🥁:

defineNuxtConfig({
  sound: {
    sounds: {
      back: {
        src: "/back.wav",
        options: {
          volume: 0.25
        }
      }
    }
  }
})

You can also automatically scan an generate typings for your sounds in public/sounds directory by using scan feature:

defineNuxtConfig({
  sound: {
    sounds: {
      scan: true
    }
  }
})

Then move your sounds into public/sounds directory, and get autocompletion on useSound({url}).

All the credit behind this idea goes to Josh W. Comeau.

The documentation of this package has only been updated for Vue Composition API instead of React Hooks.

If you like this package, consider following me on GitHub and on Twitter.