web audio API player

About this project

🎶 An opensource javascript (typescript) audio player for the browser, built using the Web Audio API with support for HTML5 audio elements

This player can be added to any javascript project and extended in many ways, it is not bound to a specific UI, this player is just a core that can be used to create any kind of player you can imagine

To get started I recommend checking out the guide to building a simple audio player UI in this readme, or head straight to the player options and functions documentation, but also have a look at the code of the working simple player example that is part of this repository

If you want to help improve the documentation or contribute to this project by improving and fixing it, then first check out the TODOs section below, maybe there is something in the list you want to help with

Any contribution, even things not listed on the TODO list are of course welcome. To get started check out the section "contributing" section below

If you found a bug or want to request a new feature please go to the issues page and if you have a question please use the dicussions page

Demo

I use the player on my personal blog chris.lu, there is a "jukebox" in the header on the top right where you can try out the player. The source code for my blog is on GitHub in the chris.lu repository, the player UI is a React component in which I use the web audio API player

installation

web audio API player is published to the npm registry so you can install it with either npm or yarn

with npm:

npm i web-audio-api-player

or with yarn:

yarn add web-audio-api-player

examples

the best way to get started is to check out the examples folder, check out the source of simple player example if you want to see how to build a fully working player with UI elements of a basic audio player

documentation

guide to building a simple audio player UI

in this chapter I will try to explain how to set up the most important parts of a player, but I also recommend you have a look at the simple player example which is an HTML / Javascript client and has an express.js server, to demonstrate how to build an UI, you can explore and run the example locally if you want to know more about how to use this package and see a working example

after having installed the package you need to import it, like so:

import { PlayerCore, ICoreOptions, ISoundAttributes } from 'web-audio-api-player'

what you must import is the PlayerCore, the other two ICoreOptions and ISoundAttributes are optional, I import those two because I write my code using typescript and want the player and sound / song options types

first we define some options for our player core:

const options: ICoreOptions = {
    soundsBaseUrl: '/assets/songs/',
    loopQueue: true,
}

[!NOTE]
the soundsBaseUrl is the first option we set, it will tell the player what the full URL for the songs source is (for example https://www.example.com/songs/) or if the player and songs are hosted on the same domain the path is enough, loopQueue by default is set to false, I enable it here, this means that at the end of a queue (a playlist) the player won't stop but instead go back to the first song and play that song

for a full list of all available player options check out the player options chapter

next we initialize the player using our options object and get a player instance in return:

const player = new PlayerCore(options)

now we are going to create our first song:

const firstSongAttributes: ISoundAttributes = {
    source: [
        {
            url: 'mp3/song1.mp3',
            codec: 'mp3',
        },
        {
            url: 'ogg/song2.ogg',
            codec: 'ogg',
            isPreferred: true,
        }
    ],
    id: 1,
}

the only two attributes that are mandatory are the source array and the sound id, the source only needs one entry but for demonstration purposes I added two here, the first one is the song encoded as an mp3 and the second source is the same song but this time it has is encoded using the ogg codec, a third source option is isPreferred, it tells the player that if the browser has support for both codecs but that it should preferably use ogg over mp3, the id can be any numeric value, it can be useful if you have additional song data stored somewhere, for example if you have the related band name info, the songs music genre and so on, for example stored in a database and want to display that data in the UI while the song is being played

[!TIP]
for a full list of all available sound attributes check out the sound attributes chapter

after we have set the attributes for our first song we pass these attributes to the player queue:

const firstSong = player.addSoundToQueue({ soundAttributes: firstSongAttributes })

if you want to you can add callbacks via the songs attributes, these callbacks will get triggered by the player when an internal event happens to let your code adapt the UI based on them, I'm going to use those callbacks with a console.log inside to demonstrate their use as I add a second song to queue:

const secondSongAttributes: ISoundAttributes = {
    source: [
        {
            url: 'mp3/song2.mp3',
            codec: 'mp3',
        },
        {
            url: 'ogg/song2.ogg',
            codec: 'ogg',
            isPreferred: true,
        }
    ],
    id: 2,
    onLoading: (loadingProgress, maximumValue, currentValue) => {
        console.log('onLoading (loadingProgress, maximumValue, currentValue): ', loadingProgress, maximumValue, currentValue)
    },
    onPlaying: (playingPercentage, duration, playTime) => {
        console.log('onPlaying (playingPercentage, duration, playTime): ', playingPercentage, duration, playTime)
    },
    onStarted: (playTimeOffset) => {
        console.log('onStarted (playTimeOffset): ', playTimeOffset)
    },
    onPaused: (playTime) => {
        console.log('onPaused (playTime): ', playTime)
    },
    onStopped: (playTime) => {
        console.log('onStopped (playTime): ', playTime)
    },
    onResumed: (playTime) => {
        console.log('onResumed (playTime): ', playTime)
    },
    onEnded: (willPlayNext) => {
        console.log('onEnded (willPlayNext): ', willPlayNext)
    },
    onSeeking: (seekingPercentage, duration, playTime) => {
        console.log('onPlaying (seekingPercentage, duration, playTime): ', seekingPercentage, duration, playTime)
    },
}

after we have set the attributes for our second song we pass these attributes to the player queue too, which means we now have a queue with two songs:

const secondSong = player.addSoundToQueue({ soundAttributes: secondSongAttributes })

some player options can be changed even after initialization, for example if you want to adjust the volume, you could do this:

let volume = 90

player.setVolume(volume)

or you want to player to pause playing (or mute depending on what action you chose) the current song when the browser / app gets put into the background, then you can enable the option like this:

player.setVisibilityWatch(true)

to change what happens when the visibility API detects that the player is hidden, you can use the following setter:

player.setVisibilityHiddenAction(PlayerCore.VISIBILITY_HIDDEN_ACTION_PAUSE) // or PlayerCore.VISIBILITY_HIDDEN_ACTION_MUTE

or you want the queue to make a loop when the last song in the player queue (your playlist) finishes playing, then you would enable / disable it like this:

player.setLoopQueue(true)

[!NOTE]
all of these setters have a corresponding getter, so if you want to now what the current value is, for example if you want to know what the current volume is set to:

const volume = player.getVolume(volume)

now it is time to build your player UI, if you want a good example of such an UI check out the simple player example

first thing we need is an play button (of course you can use any element you want, you just need to attach an onclick to it), in this example we will use an HTML button element:

<button id="playButton" class="button">
    <span id="play-icon">&gt;</span>
</button>

and then you listen for the onclick, when the onclick gets triggered you tell the player to start playing (if nothing is defined it will play the first song in the queue by default):

const playButton = document.getElementById('playButton');
playButton.addEventListener('click', (event) => {
    event.preventDefault();
    player.play()
})

here is another example from a react component I use for my blog chris.lu source on github:

<button onClick={onClickPlayHandler} className={styles.play}>
    <FontAwesomeIcon icon={faPlay} size="2x" color='white' />
</button>

and here is the click handler I have in my react component, which tells the player to play the first song from the queue:

const onClickPlayHandler = () => {
    player.play()
}

One last tip, when you want to change the position of the song, for example when someone uses the range slider of your player UI, then it is best to not stop (or pause) the song and then use play() to resume playing at a certain position, instead the easiest way is just to call the setPosition method of the player:

const onChangePositionHandler = (positionInPercent: number): void => {
    player.setPosition(positionInPercent)
}

player options

[!TIP]
if you use typescript, import the ICoreOptions interface along with the playerCore, this makes it a lot easier to see what player options are available and what the type of each value is

• volume: [number] (default: 80) the current playback volume
• loopQueue: [boolean] (default: false) after the last sound in the queue has finished to play should the player do a loop and continue to play by playing the first sound or stop playing
• soundsBaseUrl: [string] (default: '') the base url for the location of the sounds
• playingProgressIntervalTime: [number] (default: 200) the interval in milliseconds at which the player should trigger a sounds onPlaying callback which will tell you the playing progress in percent, this value is a minimum value because the player uses the requestAnimationFrame internally, meaning that if the browser is very busy it might take a bit longer than the defined interval time before the progress value is being reported, this helps to prevent that your UI uses resources that are needed more urgently somewhere else
• playNextOnEnded: [boolean] (default: true) when a sound or song finishes playing should the player play the next sound that is in the queue or just stop playing
• stopOnReset: [boolean] (default: true) when the queue gets reset and a sound is currently being played, should the player stop or continue playing that sound
• visibilityWatch: [boolean] (default: false) tells the player that if a sound is playing and the visibility API triggers a visibility change event, then the sound should get paused or muted (depending on what action is set to get executed on visibility is hidden), uses the Page Visibility API internally
• visibilityHiddenAction: [typeVisibilityHiddenAction] (default: VISIBILITY_HIDDEN_ACTION_PAUSE) chose what action should get executed on visibility is hidden, if set to VISIBILITY_HIDDEN_ACTION_PAUSE the sound will get paused on visibility hidden and will start playing again when the visibility API triggers an event that tells the player that it is visible again, if set to VISIBILITY_HIDDEN_ACTION_MUTE the song will continue to play but the player will get muted on visibility hidden and unmuted when it becomes visible again
• unlockAudioOnFirstUserInteraction: [boolean] (default: false) this tells the player to attempt to unlock audio as soon as possible, so that you can call the player play() method programmatically at any time, if you don't want to the player to handle this part and prefer to do it manually then you can use the player function called manuallyUnlockAudio(), for more info about this check out the chapter "locked audio on mobile"
• persistVolume: [boolean] (default: true) if this value is set to true the player will use the localstorage of the browser and save the value of the volume (localstorage entry key is WebAudioAPIPlayerVolume), if the page gets reloaded or the user comes back later the player will check if there is a value in the localstorage and automatically set the player volume to that value
• loadPlayerMode: [typePlayerModes] (default: PLAYER_MODE_AUDIO) this is a constant you can import from player, currently you can choose between two modes, PLAYER_MODE_AUDIO which uses the audio element to load sounds via the audio element and PLAYER_MODE_AJAX to load sounds via the web audio API, for more info about the modes read the player modes explained chapter
• audioContext: [AudioContext] (default: null) a custom audiocontext you inject to replace the default audiocontext of the player
• addAudioElementsToDom: [boolean] (default: false) when audio elements get created, they are by default offscreen (not added to the DOM), if you want the audio elements to be added to the DOM set this option to true
• volumeTransitionTime: [number] (default: 100) value in milliseconds, by default the volume if changed will not go instantly from the old value to the new value but there is transition where it fades from current to new value, use this option to increase or decrease the transition duration

player functions

[!TIP]
all player functions return a promise, I recommend using a try catch and await the promise or call promise.catch to fetch eventual errors thrown by the player, like so:

async function foo(): Promise<void> {
    try {
        await player.play()
    } catch (error) {
        console.error(error)
    }
}
foo()

or like so:

function bar(): void {
    player.play().catch((error) => {
        console.error(error)
    })
}
bar()

• play() [optional property IPlayOptions] (default {}) starts playback of a sound, returns a promise that when resolved returns the current sound

IPlayOptions {
    whichSound: accepted values: song ID (number or string) OR one of these 4 constants: PlayerCore.PLAY_SOUND_NEXT, PlayerCore.PLAY_SOUND_PREVIOUS, PlayerCore.PLAY_SOUND_FIRST, PlayerCore.PLAY_SOUND_LAST
    playTimeOffset: the time at which you want the sound to start (in seconds), usually the song would start at zero but if you set this it will start at playTimeOffset
}

[!NOTE]
the playTimeOffset (if set) will always get honored, so if you want to resume after a pause don't set the playTimeOffset, if playTimeOffset is set the song will start at the specified position, if no playTimeOffset is set the player will use the songs playTime value, which is 0 for a song that gets played for the first time or a value > 0 for a song that was paused

• pause() pauses playback, returns a promise that when resolved returns the current sound
• stop() stops playback, returns a promise that when resolved returns the current sound
• next() used to play the next sound from the internal queue, returns a promise that when resolved returns the current sound
• previous() used to play the previous sound from the internal queue, returns a promise that when resolved returns the current sound
• first() used to play the first sound from the internal queue, returns a promise that when resolved returns the current sound
• last() used to play the last sound from the internal queue, returns a promise that when resolved returns the current sound
• addSoundToQueue() [ISoundsQueueOptions] (default { soundAttributes, whereInQueue? }) adds a new sound to the queue, it returns a sound (ISound) object
  • addSoundToQueue has a single parameter soundsQueueOptions is an object of type ISoundsQueueOptions, it has two properties soundAttribtes (object) and whereInQueue (string)
  • the soundAttribtes property you probably want to set is source, source is either a single ISoundSource object or an array of ISoundSource objects and the second property you may want to set is id (string or number), if you set it you will be able to play sounds by id, however if you don't set it the player will create an id
  • an ISoundSource object has two required properties url (string) and a codec (string), the third property isPreferred is optional, it can be used to tell the player which source is preferred (if you set two sources, one with the codec mp3 and another source for that same song but with the codec ogg and both codecs are supported then the player will take the one that is marked as preferred and if none is marked as preferred it will take the first one that has a codec which is supported)
  • the other property whereInQueue is optional and tells the player to put the sound at the beginning or the end of the queue, use the constants PlayerCore.WHERE_IN_QUEUE_AT_START or PlayerCore.WHERE_IN_QUEUE_AT_END, if you don't specify a constant the sound will be put at the end of the queue
  • for some sample code check out building a simple player guide or for a working example have a look at simple player example source code in the file bootstrap.ts, or use this basic example to get started:

const mySoundAttributes = {
    source: [{ url: 'https://example.com/mySound.mp3', codec: 'mp3' }],
}

player.addSoundToQueue({ soundAttributes: mySoundAttributes })

• resetQueue() (or reset()) removes all sounds from the queue
• getQueue() returns the an array of sounds currently in the queue
• setVolume(volume: number) sets to the volume in percent, so a number ranging from 0 to 100
• getVolume() returns the current volume
• setLoopQueue(loppQueue: boolean) to let the player know it should loop the queue, meaning that after the last song in the queue finished playing it will play the first sound in the queue, set to false to just stop at the last song in queue
• getLoopQueue() get the current boolean value to which loopQueue is set
• mute() sets the volume to 0
• unMute() resets the volume to its previous value
• isMuted() a boolean value, true if the volume is currently muted else false
• setPosition(soundPositionInPercent: number) used to change the position of the song that is currently playing in percent, so a number ranging from 0 to 100, returns a promise
• setPositionInSeconds(soundPositionInSeconds: number): used to change the position of the song that is currently playing in seconds, the number should be smaller than the duration of the song, returns a promise
• setVisibilityWatch(visibilityWatch: boolean) a boolean value to change the visibilityWatch option of the player, if true the player will be paused / muted when the visibility API MDN visibility API triggers an event that tells the player that the browser / app is in the background, will start playing again or get unmuted when the visibility API notices that the browser is in the foreground again, if false visibility changes will get ignored by the player
• getVisibilityWatch() get the current boolean value that is set for the visibilityWatch option
• setVisibilityHiddenAction(visibilityHiddenAction: typeVisibilityHiddenAction) change what action will get executed when the visibility API detects that the player is hidden / visible again, see player option for more details
• getVisibilityHiddenAction() get the current action that is set for the visibilityHiddenAction option
• disconnect() disconnects the player and destroys all songs in the queue, this function should get called for example in react when a component unmounts, call this function when you don't need the player anymore to free memory
• getAudioContext() get the current audioContext that is being used by the player MDN audiocontext
• getCurrentSound() returns the current sound, can be useful if you want to do things in your UI like getCurrentSound().getLoop() to check if the loop feature is on or off for the current song
• manuallyUnlockAudio() this method can be used on mobile to unlock audio, you need to call this function inside an event handler that got triggered by the user, so for example an "onClick" event could call this function to unlock audio, calling this function programmatically without any user interaction will not unlock audio, an alternative if you don't want to implement this yourself is to enable the player option called unlockAudioOnFirstUserInteraction, for more info about this check out the chapter "locked audio on mobile"
• loadSound(sound: ISound, afterLoadingAction?: typeAfterLoadingAction) this method can be used to load a given sound, the second parameter afterLoadingAction is optional, if the afterLoadingAction is set to PlayerCore.AFTER_LOADING_SEEK then after the loading is done the sound will seek to a position defined by sound.seekPercentage or if set to PlayerCore.AFTER_LOADING_PLAY then the song will start playing after the loading has finished, if not set the sound will just get loaded, this is a helper function you don't need to trigger it manually if you have no use case that requires it, if you use play() instead then loadSound will get called automatically (with afterLoadingAction set to play) and if you use setPostion before the sound started playing then loadSound will get called automatically (with afterLoadingAction set to seek), loadSound is useful if you want to get the sound duration before the song is actually playing (as long as a sound is not loaded it's duration is not defined)

sound attributes

[!TIP]
if you use typescript, import the ISoundAttributes interface along with the playerCore, this makes it a lot easier to see what sound attributes are available and what the type of each value is

sound options:

• source: [(ISoundSource)[] | ISoundSource] (optional if an AudioBuffer or ArrayBuffer is provided instead else mandatory) a single sound source or an array of sound sources, an ISoundSource object consists of 3 values:
  • url [string] is the base url defined in the player options + the path defined here or you add the full url here, the URL will get used by the player to load the sound when needed
  • codec [string] the codec that got used to encode the sound, this allowed the player to check if that codec is supported by the browser and it also allows the player to decide which source to use if multiple sources have been defined
  • isPreferred [boolean] (optional) the player will use the first source that has a codec this is supported by the browser, if more than one codec is supported it will take the source that is marked as isPreferred
• id: [number | string] (optional, if none is set the player will generate one) unique id for the sound, can be used as a reference to link sound data which is not part of the sound object itself to an external source, for example if you have sound info stored in a database, set the sound id to the database id and you have a link between the two, it also allows you to call the player.play function using the sound id as argument to play that sound
• loop: [boolean] (optional, default false) if the sound playback should loop when it reaches the end of sound
• audioBuffer: [AudioBuffer] (optional) if you want to inject your own custom AudioBuffer to be used instead of the default one the player will create
• arrayBuffer: [ArrayBuffer] (optional) if you want to inject your own custom ArrayBuffer to be used instead of the default one the player will create
• duration: [number] (optional) if you know the duration of the sound and want to tell the player about it early, in PLAYER_MODE_AJAX the player will need to wait for the sound to be fully loaded until it can determine the duration
• seekPercentage: [number] (optional) a value in percent used by the player loadSound function, if the second parameter of loadSound is set to PlayerCore.AFTER_LOADING_SEEK then after loading the sound will seek to the seekPercentage value

sound callbacks:

• onLoading(loadingPercentage, total, loaded): [function] (optional) a callback function that will get triggered at intervals during the loading process of a sound, the interval duration can be changed using the player option "playingProgressIntervalTime", the callback has three parameters, loadingPercentage is the percentage that has been loaded so far (number ranging from 0 to 100)
  • if the player mode is PLAYER_MODE_AUDIO, then total is an integer telling you the total song duration in seconds (MDN HTMLMediaElement duration), loaded is an integer that tells you the amount that already got loaded (buffered end) in seconds (MDN HTMLMediaElement timerange end), if you prefer to use the raw values I recommend you use the sound.audioElement.buffered and read them periodically yourself
  • if the player mode is PLAYER_MODE_AJAX, then total is an integer telling you the total bytes that will get loaded (MDN ProgressEvent total), loaded is an integer that tells you the amount of bytes that already got loaded (MDN ProgressEvent loaded)
• onPlaying(playingPercentage, duration, playTime)): [function] (optional) a callback function that will get triggered at intervals while the sound is playing, the callback has three parameters, playingPercentage is the percentage that has been played so far (number ranging from 0 to 100), duration is the total song duration, playTime is the current time in seconds that have been played
• onEnded(willPlayNext): [function] (optional) a callback function that will get triggered when the end of the sound is reached, returns one parameter willPlayNext which is a boolean, true if there is a next song in the internal queue that will get played or false if no next sound will get played
• onStarted(playTimeOffset): [function] (optional) a callback function that will get triggered when the sound playback starts, returned value is the playTimeOffset of the song, usually playTimeOffset is zero unless you explicitly set it to be something else
• onStopped(playTime): [function] (optional) a callback function that will get triggered when the sound playback is stopped (the difference between pause and stop is that stop will free the resources needed to play a song), returns one parameter playTime which is the current sound position in seconds
• onPaused(playTime): [function] (optional) a callback function that will get triggered when the sound playback is being paused (use pause instead of stop if there is a reason to assume that the playback will be resumed at anytime, if this can't be assumed it is recommended to call stop), returns one parameter playTime which is the current sound position in seconds
• onResumed(playTime): [function] (optional) a callback function that will get triggered when the sound playback gets resumed after if got set to pause, returns one parameter playTime which is the current sound position in seconds

sound functions

• getCurrentTime() returns the current playTime
• getDuration() returns the duration of the sound
• setDuration(duration: number) you can set the duration of a sound manually, this avoids having to wait for the sound to be loaded to know it's duration, might be useful depending on the kind of UI you are building
• setLoop(loop: boolean) allows you to change the initial value for "loop sound", by default loop is false, use this to change the loop value after the sound got added to queue
• getLoop() returns a boolean telling you what the current value for "loop sound" is

locked audio on mobile

All mobile browsers prevent playing sounds (songs) if no user gesture has happened yet. This means that on mobile you can NOT play sounds (songs) programmatically (this is also the reason why the autoplay attribute on an audio element does not auto play a song on mobile and also the reason videos will only autoplay if they are muted)

[!NOTE]
If the user clicks on a play button and call player.play() then audio will play just fine, this chapter is about audio not playing when calling player.play() before the user interacted with the page (app)

If you attempt play a sound (song) on mobile programmatically (before a user interaction) then the mobile browser will throw a NotAllowedError error:

The request is not allowed by the user agent or the platform in the current context, possibly because the user denied permission (No legacy code value and constant name).

so if you encounter this error on iOS (iPhone) and android mobile devices, then it means that you tried to play the sound programmatically before the browser got "unlocked" by a user interaction

[!NOTE]
in the past iPad tablets would throw an error too, however newer versions are considered a desktop device and do not need throw an error

There is however a trick to unlock audio on mobile, the trick is to listen for events like a user clicking on something in your page and use that interaction to play a silent sound for a brief moment, after that audio will be unlocked and you will be able to trigger the play function at any time programmatically to play the song you want (even if it is not a direct action initiated by the user)

the web-audio-player has two options to unlock audio on mobile:

• solution 1: there is a player option called unlockAudioOnFirstUserInteraction, set it to true when initializing the player and the player will add user interaction listeners to the html document, on the first user interaction the player catches, it will attempt to unlock audio, after audio is unlocked you will be able to call the players play() function programmatically and it will not throw an error anymore

• solution 2: there is a player function called manuallyUnlockAudio() that you can use to attempt to unlock audio on mobile, this function MUST be played inside an event handler that got triggered by a user interaction, events that you can use are for example "keydown" (excluding the Escape key and possibly some keys reserved by the browser or OS), "mousedown", "pointerdown" or "pointerup" (but only if the pointerType is "mouse") and "touchend"

player modes explained

[!NOTE]
You might have read (like I did) a lot of outdated web audio articles which stated the web audio element lacks a lot of features the web audio API and that hence it is not suited to create complex audio software or for example be used in games where you might want to add effects and filters to sounds. This is not true anymore and especially not true for this library. Yes the audio element if used as a standalone lacks a lot of features. But this library does combine the web audio element with the web audio API, meaning that no matter what mode you chose the sound will be converted to an AudioSourceNode.

If you use this library, you have two player modes you can chose to use, the main difference is how the sound (song) gets loaded:

PLAYER_MODE_AJAX will use an XMLHttpRequest the source will be an AudioBufferSourceNode

PLAYER_MODE_AUDIO will use the HTML audio element, then the player will use createMediaElementSource method of the AudioContext internally to create an MediaElementAudioSourceNode

so which PLAYER_MODE should I use

If you build something like a music player, it is probably best to use the PLAYER_MODE_AUDIO as you might want to start playing the sound (song) as quickly as possible and don't care if it has fully loaded. This mode is ideal for big files that don't get loaded all at once (streaming). The audio mode (via the audio element) has support for partial content (http code 206) this means with this mode the song will start playing as soon as enough data has been buffered even though the song has not been fully loaded yet (it will load more data from the server in the background as the song progresses). The loading progress callback will return a percentage, which represents the amount of data that got loaded so far, which means it might not represent the loading state of the full song. If you want to display what parts of the song have been loaded more accurately (display the time range(s) that got loaded) I recommend using a 2D canvas element and to get the ranges that have been loaded, I recommend you use the audioElement property of a song to get the song.audioElement that is loading the song and then read the audioElement.buffered value(s).

You can use the PLAYER_MODE_AJAX if for example you want to build something where it doesn't matter that the song will only play after it has been fully loaded. However in this mode you can (pre-)load and maybe also cach sounds by yourself (you can inject an array buffer that you loaded yourself (via an XMLHttpRequest or using fetch) or even an already decoded audio buffer) by setting the sound.arrayBuffer or sound.audioBuffer. Use this mode if you prefer to have a smooth loading animation, because its loading progress callback is straight forward, when the loading progress callback gets triggered by the player, you can use the percentage value and pass it to a progress bar. To display the loading progress you could for example use a HTML progress element, you can find such an example in the simple player example.

advanced usage

You can create and then inject your own AudioContext

You can inject your own using the audioContext player option, if you want to reuse an existing one your app already created

This is especially useful if you want to add your own nodes to the AudioGraph (audio routing graph). For example you may want to add an AnalyserNode or a pannerNode, delayNode or any other node that is available in the web audio API.

read more

• W3C Recommendation, 17 June 2021
• Web Audio API: Editor’s Draft, 18 July 2023
• MDN "Web Audio API" page
• MDN "The Embed Audio element" page
• MDN webaudio examples repository on github
• Support tables for audio features, caniuse: web audio API / audio element / formats ...

development: build

install the latest nodejs (if you haven't already) nodejs

install or update to the latest git version git scm downloads (During installation at the step "choosing the default editor used by Git", if like me you are using visual studio code you might want to chose the new option "use visual studio code as Git's default editor") (also if like me you are on windows, at the step "adjusting your PATH environment", ensure the second radio button option is selected "git from the line and also from 3rd-party software" to ensure git is added to the windows PATH, which will allow you to use git with any command line tool like windows powershell for example)

git clone this repository to get a local copy

git clone [email protected]:chrisweb/web-audio-api-player.git

open your favorite command line tool and go to the root directory of this repository

update npm to latest version

npm install npm@latest -g

install the dependencies

npm i

to build the distributions

npm run build

in development you can use watch to rebuild every time you edit a typescript file

development: watch

npm run watch

development: linting

to lint the typescript files

npm run lint

Changelog

check out the releases page on github

contributing (PRs welcome)

if you wish to contribute to this project, please first open a ticket in the github issues page of this project and explain briefly what fix or improvement you want to provide (remember the github ticket number you will need it for the commit message later on), if you want to help but are not sure what would be useful then check out the todos list

go the github page of this project and hit the fork button

follow the instructions in the previous section "development: build", but instead of cloning this projects repository, clone your own fork of the project to get a local copy that you can edit in your IDE (VSCode)

git clone https://github.com/YOUR_GITHUB_USER/web-audio-api-player.git

when you are done coding, commit your local changes (if your commit is related to a ticket start your commit message with the "#TICKER_NUMBER", this will "link" the commit to the ticket)

git commit -m "#TICKER_NUMBER commit message"

now go to the github page of your fork and hit the pull request button

TODOs (ideas for future improvements)

things I intend to add some day or if you want to help but are not sure what to do, check out this list and just pick one idea you like or that would you think is most useful

if you are interested in helping out 😊 by working on one of the following TODOs, please start by reading the "contributing" chapter above

• add an option to the sound to allow setting a custom media element for audio mode!? similar to arraybuffer and soundbuffer for ajax mode
• the player has a big limitation, the audio mode uses a single audio element, this means that playing a song and loading the next one will not work as both would share the same media (audio) element, the single media element makes it easier on mobile to unlock audio early and then play whatever song later, it would be better if each song had it's own element but the update should not break mobile unlocking
• for audio mode, add option for songs "crossOrigin" to be set to "use-credentials" HTML attribute: crossorigin instead of "anonymous", "anonymous" could be kept as default value, also do the same for XHR calls (ajax mode) XMLHttpRequest: withCredentials property
• add a song feature to fade out (current song) / fade in (next song) maybe two new options fadeInDuration and fadeOutDuration would be enough, if the values are set we do a fade
• add onDurationChange / onLoopChange callbacks to song(s)!?
• add sound option to set the initial gain value of a sound, for now it is always 1 (1 = no change to loudness), (optional method that lets you define a modifier (or coefficient) per song by which the gain will be changed), useful if some songs are louder than others and you want to normalize the volume of all songs in a playlist to be similar
• add a shuffle songs feature, I thought about adding a PLAY_SHUFFLE(D) option for the play() function but then only the first song would be a random pick, so the player itself needs a shuffle mode which also shuffles the next songs, you need to be able to turn it on / off at any time
• today we have sound callbacks for events like "playing", "started"..., but would player callbacks be useful, we could add callbacks for things like an "onVolumeChange" event? HTMLMediaElement Events ... the AudioBufferSourceNode does not have events (other than onended), even the playbackState got removed: playbackState
• would it be useful to have player events, so that callbacks can be added, directly the player itself, for things like "volumechange" and maybe act as alias for the events of the current song, like "on playing current song"?
• add an example where dev custom audiocontext gets connected to an AnalyserNode or oscilator node and then passed to player (to which we then connect the song gainnode and then connect that audiocontext to the destination)
• instead of having sound properties like "isReadyToPLay, isBuffered, isBuffering" it would be better to use the SOUND_STATE_XXX constants
• add (stereo) panning (maybe add an example of how to do it by injecting an audiocontext that has the panningnode attached already) MDN: StereoPannerNode
• use service worker to cache ajax requests when the player mode is ajax (audio buffer) MDN: Service Worker API
• allow to add an array of sounds to queue all at once
• allow to add sound to queue after a sound by id (not just at beginning or end, as it is as of now)
• for volume (gain) allow values beyond 0 to 1 to amplify wave or invert it?
• add improve UI style of the "simple" example(s) (or any new example) and then add a screenshot of it to the readme to show what can be achieved
• opt in feature to use the browser notification system to alert which song is being played (?)
• option to preload the first song or more songs, would have to work for both player modes, would be most beneficial for ajax mode as song(s) can only play when fully loaded, need to check if preloading ArrayBuffers (or the already decoded audio buffer instead of array buffer) can be saved into indexeddb (the first song, maybe also the next song, or reuse the buffer (array) of the current song if loop is active), eventually it would be best to let the developer set the amount of preloaded ArrayBuffers he wants to store, but then I would also need a mechanism to remove them from "cache" at some point, maybe clear cache by least used or by date, maybe do such preloading work using this browser feature: "requestidlecallback" (when available)
• cache songs for offline mode? indexeddb is probably not very big (save directly to the filesystem?), check if it is doable, because saving a playlist of songs might exhaust the free space the browser has (depending on the playlist size), maybe using a service worker to make such an operation in the background?
• some methods return a promise others don't, use promises for all to make it more consistent?
• add feature to be able to abort the XHR request, we could abort the loading of the sound if the user clicks play and then pause (or stop / next / previous) before the end of the loading (buffering) process
• add new loadPlayerMode similar to XMLHttpRequest but that uses fetch to get the sound / song data from server, fetch as of now is still young, it is not yet possible to get the loading progress (https://stackoverflow.com/a/69400632/656689)
• Keep an eye on this ticket, if tracking errors via reporting API gets implemented some day it would be nice to use that new feature
• use suspend (audioContext) if for some time no sound was played? (and then resume needed?) ... to free device resources. As suspend returns a promise, does this mean suspending and resuming takes time? If so, how much time does it take, based on that information we can decide after how much time it makes sense to suspend the ctx to save device resources, should this be an event of the player so that user can attach a callback to show a dialog like netflix (still watching?)
• use web workers, especially for the decoding of the ArrayBuffer into an AudioBuffer, to not block the main thread while decoding?
• add support for more codecs? Note: the player should check which codecs are supported by the browser and compare that list with the ones defined in the sound sources, then the player should use the first codec that is supported and that is marked as "isPreferred", if none is marked as "isPreferred" use the first sources codec that is supported
• write code tests!!! (goal ofc 100% coverage), add tests coverage badge
• add saucelabs (or similar) browser testing (and their badge browser compatibility table badge in readme) to add a test suite for all player features
• use github actions for a CI/CD workflow
• add live demo (via github pages?) for people to see how the player works
• create a react example (vite server?)
• create a vue.js example
• create an example using the (browser) fileReader, something like:

var fileInput = document.querySelector('input[type="file"]');
fileInput.addEventListener('change', function(event) {  
  var reader = new FileReader();
  reader.onload = function(event) {
    playerCore._decodeSound(this.result);
  };
  reader.readAsArrayBuffer(this.files[0]);
}, false);

notes about problems I encountered during development

web audio api typings notes

As of the 25.05.2019 the web audio api typings seem to be included in lib.d.ts, so removing them from package.json:

  "dependencies": {
    "@types/webaudioapi": "0.0.27"
  },

Unfortunately (as of 06/2019) the defined window does not have AudioContext:

• check out github ticket
• the current dom.d.ts on github

This is fixed, as of now (20.02.2023) the AudioContext is now defined properly

License

MIT