Melonjs Audio Agent Skill

Audio

Two setup rules

import { audio, loader } from "melonjs";

audio.init("mp3,ogg");                      // 1. BEFORE preloading any audio

await loader.preload([
    { name: "cling", type: "audio", src: "data/sfx/" },   // 2. a PREFIX, not a file
]);

• audio.init(format) must run before any audio asset is preloaded, and

lists the formats you actually ship. Preloading first throws

`target audio extension(s) should be set through me.audio.init() before

calling the preloader. It returns false` when no audio backend is

available, so it is worth checking. The format list defaults to "mp3".

• src is a prefix, not a file. The loader appends the asset name plus

each configured extension and uses the first that loads — so

data/sfx/ + cling + mp3 → data/sfx/cling.mp3. Passing a full filename

gives you a 404. A data:audio/... URL is the one exception: it is used

verbatim, prefix and extension skipped.

audio.hasAudio() says whether the browser can play audio at all, and

audio.hasFormat("ogg") whether it can play one codec — useful for choosing

the init list or hiding a sound toggle on a device with no audio.

Order matters — melonJS tries the listed formats left to right, so put the

preferred one first. Two formats ("webm,mp3" or "mp3,ogg") is the usual

belt-and-braces for codec coverage. Accepted tokens: mp3, mpeg, opus,

ogg, oga, wav, aac, caf, m4a, m4b, mp4, weba, webm, dolby,

flac.

Effects versus music

audio.play("cling");                    // fire and forget
audio.play("cling", false, null, 0.5);  // loop, onend, volume
const id = audio.play("engine", true);  // looping; keep the id to stop it

audio.playTrack("theme");               // background music
audio.stopTrack();

Use playTrack for music, not play. playTrack(name, volume?) always

loops, and it records the name as the current track — which is what

app.pauseOnBlur / resumeOnFocus / stopOnBlur act on, via

state.pause(true) → pauseTrack(). stopTrack(), pauseTrack(),

resumeTrack() and getCurrentTrack() all read that same single slot, so there

is only ever one current *track name*.

playTrack stops the track it replaces (20.9.0 and later; before that it

only assigned the slot, and the previous bed kept looping underneath where

stopTrack() could no longer reach it). So a scene just states its own music

and nothing cancels anything:

class PlayScreen extends Stage {
    onResetEvent() { audio.playTrack("battle", 0.6); }
}

Calling it with the track already playing restarts it, the same as play()

on any other clip. A scene that wants the bed to carry on across a screen

change should not call it again, rather than rely on it doing nothing.

stopTrack() is for where nothing follows: a scene whose successor has no

music of its own, or a game going quiet. state.change destroys the outgoing

stage before it resets the incoming one, so a stopTrack() in

onDestroyEvent and a playTrack() in the next onResetEvent are in that

order and never cancel each other.

To end a track on an event rather than on a screen change, say a run ending

while the player stays on the same screen, audio.fade(name, from, 0, ms)

takes it down gently. A fade to zero is silence, not a stop, so the voice is

still live until something stops it or replaces it.

Every per-clip function throws audio clip <name> does not exist if the clip was

never loaded, and play additionally throws when a named sprite region is

missing.

Audio sprites

One file carrying many effects, with named regions in milliseconds:

await loader.preload([{
    name: "sfx", type: "audio", src: "data/sfx/",
    sprite: { jump: [0, 450], hit: [2000, 250], music: [4000, 12000, true] },
}]);

audio.play("sfx", { sprite: "jump" });

A loop with no gap

An MP3 encoder pads the end of its file. Chrome and Firefox trim the padding when

they decode it, Safari keeps it, so a looping MP3 has a short silence at every

loop there (27 ms on a 12 s clip at 22 kHz, measured). Declare a region of the

loop's exact length and loop on that: the voice never reaches the padding.

await loader.preload([{
    name: "ambience", type: "audio", src: "data/sfx/",
    sprite: { loop: [0, 12000] },
}]);
audio.playTrack("ambience", { sprite: "loop", volume: 0.6 });
// or, as a plain clip: audio.play("ambience", { sprite: "loop", loop: true })

Ogg Vorbis carries no such padding.

The optional third element marks a region as looping: it loops unless play() is

told loop: false (before 20.11 it only looped when loop: true was passed as

well). play()'s second argument

takes either the original loop boolean or a PlayOptions object

(sprite, loop, onend, volume, plus the spatial follow, at and

stopWithTarget) — both forms work. Object fields win over the positional

onend / volume arguments when both are supplied.

Clip state

audio.duration("theme");        // seconds — 0 while still loading
audio.duration("sfx", id);      // duration of the region THAT instance plays
audio.playing("theme");         // is anything playing?
audio.playing("theme", id);     // is that instance playing?
audio.state("theme");           // "unloaded" | "loading" | "loaded"

stop(), pause() and resume() take an optional instance id too; omit it

and the whole group is affected. audio.stop() with no arguments at all stops

every sound. audio.seek(name) reads the position, audio.seek(name, s, id?)

writes it; audio.rate(name, r, id?) changes speed and pitch together

(0.5..4.0, 1 is normal) and audio.fade(name, from, to, ms, id?) ramps

the volume (0..1, over milliseconds).

Volume and muting

audio.setVolume(0.5);           // global, 0..1
audio.getVolume();              // read it back
audio.mute("cling");            // one clip  — mute(name, id?, shouldMute = true)
audio.unmute("cling");
audio.muteAll();                // everything (audio.disable() is an alias)
audio.unmuteAll();              // (audio.enable() is an alias)
audio.muted();                  // global mute state

audio.unload(name) / audio.unloadAll() free the decoded buffers; unload

also clears the current-track pointer if it named that clip.

Losing focus mutes everything

When the window loses focus the engine mutes all audio, and unmutes it on

focus. That covers sound effects, procedural tone / noise and the mixer

effects, not just the music track that state.pause() pauses. Set

pauseOnBlur: false in the application settings if a clip has to keep playing

in the background.

Spatial audio

audio.position("engine", x, y, z);
audio.orientation("engine", x, y, z);
audio.stereo("engine", -1);     // -1 left … 1 right
audio.panner("engine", { coneInnerAngle: 90, /* … */ });

position(name, x, y = 0, z = -0.5, id?) and

orientation(name, x, y, z, id?) both default id to "the whole group".

Called with just the clip name, each of the four returns the current value —

stereo gives 0 and position gives [0, 0, 0] before they have ever been

set.

Let the engine place the sound: follow and at

The shortest path to positional audio is a play option. Both take **world

pixels**, the same numbers as pos.x / pos.y, and the engine converts to

audio space for you:

// tracks the renderable every frame, and stops when it is destroyed
audio.play("ufo", { follow: this, stopWithTarget: true });

// pinned to a fixed world point
audio.play("portal", { at: { x: 1200, y: 480 } });

follow and at are mutually exclusive: passing both throws, rather than

quietly picking one. follow re-reads the target's absolute position each

frame, so a renderable nested in a moving container is placed correctly. stopWithTarget is what you

want for a sound that belongs to an entity: without it a looping clip on a

destroyed renderable keeps playing from wherever it last was. audio.unfollow(id)

detaches a sound and leaves it playing where it is.

Move the listener, don't do the subtraction yourself

audio.setListener(player);                  // tracked every frame
audio.listener(x, y, z);                    // or place it by hand
audio.listener();                           // → [x, y, z] in world pixels
audio.listenerOrientation(fx, fy, fz, ux, uy, uz);
audio.setListener(null);                    // stop tracking

With a listener in the world, source positions are **absolute world

coordinates**, not offsets. The older idiom — passing `source.pos.x -

player.pos.x` every frame with the ear stuck at the origin — still works, but

only if you leave the listener at (0, 0, 0); mixing the two double-counts the

player's position.

setListener places the ear at the target's absolute position after the world

updates. A plain renderable contributes position only, so a side-scroller keeps

the default "facing into the screen" and gets left-right panning plus distance

falloff from the one call.

A Camera3d target also contributes its orientation, taken from its own

basis, so turning the camera swings the stereo image with it. That is the whole

of what a posed 3D scene needs:

audio.setListener(app.viewport);                 // in a 3D scene this is a Camera3d
audio.play("drone", { follow: saucer });         // scene coordinates, directly

Y is down. Every one of these calls takes and returns melonJS world

coordinates, where y grows downward; the Y-up flip Web Audio wants happens

inside the engine. Pass pos.y as it is.

The defaults are tuned for pixels

audio.setSpatialDefaults() / getSpatialDefaults() set the panner attributes

a sound placed through follow or at starts from, and the shipped values

assume a 2D game measured in pixels:

| attribute | default |

|---|---|

| refDistance | 240 |

| maxDistance | 10000 |

| panningModel | "equalpower" |

| distanceModel | "inverse" |

With those, a placed sound is at full volume out to 240 px and then halves

each time the distance doubles: about half at 480, a quarter at 960, roughly

an eighth at 2000. maxDistance is where it stops getting quieter.

They apply to placed sounds only. A plain audio.play() followed by a

bare audio.position() keeps WebAudio's own defaults, where refDistance is

1 METRE and a source 100 px out is already near silent. That path is

unchanged, which is what keeps old code working.

Widen or tighten the field once, globally:

audio.setSpatialDefaults({ refDistance: 400, maxDistance: 20000 });

audio.panner(name, attrs, id?) still overrides per clip or per voice, and it

merges over the current values rather than replacing them, so writing one field

no longer resets the rest. Called without an id it writes the group

defaults, which is what a clip-wide tuning call looks like.

For a 3D scene where front-to-back matters, panningModel: "HRTF" is worth the

cost; for a 2D one it mostly muddies the stereo image.

stereo and position: last writer wins

Both drive the same per-instance panner node, and the engine swaps the node to

match whichever you called last: stereo() after position() gives you

plain left-right panning, and position() after stereo() gives you 3D

falloff. Neither silently does nothing any more, and the order you call them in

no longer matters.

Still pick one per voice and stay with it. A 2D game usually wants stereo()

alone; reach for position() (or follow / at) when distance falloff

matters.

Spatial audio is WebAudio-only. On a clip loaded with stream: true (or

html5: true) stereo / position / orientation return early and do

nothing — which is exactly the case for the long music tracks people reach for

first. follow and at are equally inert on a streamed clip.

Procedural sound

For beeps, blips and noise without shipping files. Both take **one options

object**, and duration (in seconds) is required:

audio.tone({ freq: 440, duration: 0.2, gain: 0.3, pan: -0.5 });
audio.tone({ freq: [440, 660, 880], duration: 0.4 });   // layered partials
audio.noise({ duration: 0.5, filter: { type: "lowpass", frequency: 800 } });

tone also takes wave (default "sine"), attack (0.005) and pitchSlide

(1). noise takes type ("white" | "pink" | "brown", default

"white"), attack, pan, filter and filterSweep. gain defaults to 0.1

on both.

Sequencing: delay, not setTimeout

Both take delay — seconds before the sound starts, scheduled on the **audio

clock**. That is what builds a multi-part sound: a stinger's second note, an

explosion's double-tap.

audio.tone({ freq: 392, duration: 0.12 });
audio.tone({ freq: 587, duration: 0.18, delay: 0.11 });
audio.tone({ freq: [784, 1176], duration: 0.75, delay: 0.23 });

Do not reach for setTimeout here. A timer fires on the main thread, so a busy

frame slips the note late and the sequence audibly loses time — the fault is

easy to blame on the sound design rather than the scheduling. delay is

sample-accurate because the whole envelope is scheduled up front.

Both are silent no-ops when there is no WebAudio context — they bail on

getAudioContext() === null rather than throwing. Neither needs the clip to be

loaded; neither goes through the file-playback path at all.

audio.getAudioContext() and audio.getMasterGain() are the escape hatches for

custom Web Audio work; both return null when audio is unavailable, so guard.

Connect a custom graph to the master gain rather than ctx.destination if you

want it to respect setVolume / muteAll.

Effects on all game audio: reverb, echo, compressor

There is no built-in reverb, delay, distortion or compressor, and no per-clip

filter on file playback. There does not need to be one for the common case:

every buffered clip and every tone / noise passes through the master gain

on its way to the speakers, so an effect inserted after it applies to the

whole mix.

const ctx = audio.getAudioContext();   // creates the context if needed
const master = audio.getMasterGain();
if (ctx && master) {                   // null when audio is unavailable
    const comp = ctx.createDynamicsCompressor();
    master.disconnect();               // master -> speakers becomes
    master.connect(comp);              // master -> compressor -> speakers
    comp.connect(ctx.destination);
}

Any Web Audio node or chain works the same way: a ConvolverNode with an

impulse response for reverb, a DelayNode with a feedback gain for echo, a

BiquadFilterNode for an underwater low-pass, a WaveShaperNode for

distortion. Do it once, at startup; the master gain lives as long as the

context.

• setVolume / muteAll still work, because they act on the master gain,

which is now upstream of the effect.

• Streamed clips are not affected. A stream: true (html5: true) clip

plays through an <audio> element and never reaches the master gain — which

is usually the music. Preload the music buffered if it must go through the

effect.

• It is all or nothing. The per-instance nodes are not public, so an effect

on one clip only (muffling one sound behind a wall) is not possible through

this route; noise has its own filter for generated sounds.

• To remove it, reverse the wiring: master.disconnect(); master.connect(ctx.destination);.

Autoplay: sound is silent until the first gesture

Browsers create the audio context in the suspended state and keep it there

until a user interaction. Nothing errors — audio simply does not play, which

reads as a bug. melonJS installs one-shot touchstart / touchend / click /

keydown unlock listeners on document, so the first real gesture anywhere

unlocks it and fires the clip's on.unlock callback. Design for it: start music

on the first click or key press, not on load. The usual way to guarantee one is

a gate screen — a splash, a "press any key", or a health/photosensitivity notice

— whose dismissal *is* the gesture, so the menu behind it can start its music

unconditionally. Note the unlock listens for keydown too, so a keyboard-only

player is covered.

The context can stop again, long after it was unlocked

Unlocking is only the first half. A context that has been running can be

stopped later: melonJS suspends it itself after thirty idle seconds to save

power, and browsers stop it for their own reasons, a backgrounded tab among

them. Recovery is automatic and you do not have to write it: a user gesture

or the tab coming back to the front revives it, and a deliberate idle

suspension is left alone so the power saving still works.

The symptom worth memorising, because it points straight at the cause:

if tone and noise are audible while loaded clips are silent, the context

is healthy and the engine's playback state is wrong. The generators only need

a live context, where a clip goes through the engine's state machine and can

be parked waiting for a resume. It is never "the context died".

Both routes ask the same owner to resume, so hearing a tone tells you the

hardware is back; it tells you nothing about whether the clips waiting on the

engine's state have been let go.

That asymmetry was a real Safari-only bug: thirty idle seconds before the

first click left the engine waiting on a suspend the browser never confirmed,

so every clip queued for the life of the page while generated sound played

normally. Fixed in 20.9, and the lesson generalises: a clip that will not play

is a state question, not a hardware one.

Playback also starts asynchronously: play() returns an id while the

instance is still starting, so playing() can legitimately be false on the

very next line.

Load-time options

{ name: "sfx", type: "audio", src: "data/sfx/",
  sprite: { jump: [0, 450] },
  pool: 5,            // finished instances kept for reuse — NOT a concurrency cap
  rate: 1.0,          // initial playback rate, 0.5..4.0
  mute: false,
  loop: false,
  autoplay: false,
  preload: true,      // or "metadata"
  format: "mp3",      // explicit hint when the URL has no usable extension
  stream: true,       // long music: stream rather than decode into memory
  on: { play() {}, end() {}, stop() {}, unlock() {} },
}

pool (default 5) is a reuse cap on finished voices, not a limit on

simultaneous playback: when the number of ended voices exceeds it, the surplus

are disconnected and dropped. Raise it for a clip triggered in rapid bursts, not

to allow more overlap.

Use stream: true for long music tracks — it plays through an HTML5 <audio>

element instead of decoding the whole file into memory. (html5: true is the

same switch under its backend name.) on accepts the full lifecycle set:

play, pause, stop, end, fade, seek, rate, volume, mute,

unlock.

Outside a preload, audio.load(asset, onload?, onerror?) takes the same

descriptor and loads one clip at runtime — for a level's sounds fetched when

the level starts. The loader calls the same function, so the same rules apply.

Re-preloading a manifest that already contains a loaded clip is a no-op —

audio.unload(name) first if you genuinely want to reload it.

Credentials use the same global switch as every other asset type —

loader.setOptions({ withCredentials: true }), not an audio-specific option.

Buffered audio fetches through the shared loader path, so it honours it exactly

as an image or a JSON file does. A stream: true clip does not: it plays

through an <audio> element, which needs a crossorigin attribute rather than

fetch credentials. Preload a clip buffered if it sits behind a login.

Failing clips and setStopOnAudioError(false). The default is to fail the

preload when a clip cannot be loaded after its retries: the error reaches the

loader, so await loader.preload(...) rejects and your catch runs, and the

reason is logged. Turn the flag off and the engine mutes audio, warns

(failed loading <name>, disabling audio) and lets the load *continue* — the

preload finishes and the game reaches the next scene without sound.

Before 20.4 both halves were broken: with the flag off it disabled audio and

failed the preload anyway, and a clip whose bytes arrived but would not decode

(a missing file served as HTML by an SPA rewrite, say) reported nothing at all

either way — blank screen, empty console.

Symptom → cause

| symptom | cause |

|---|---|

| audio assets 404 | src given as a full filename instead of a directory prefix |

| target audio extension(s) should be set... thrown at preload | audio.init() not called before preloading |

| audio clip X does not exist | the clip was never preloaded, or the name is mis-cased — every per-clip call throws this |

| nothing plays until the user clicks | browser autoplay policy — expected, design around it |

| playing() is false right after play() | playback starts asynchronously |

| music keeps playing after leaving a scene | used play instead of playTrack, or no stopTrack in teardown |

| two music tracks playing at once, and stopTrack() only silences one | melonJS before 20.9.0: playTrack did not stop the previous track, so the scene that started it had to stopTrack() on the way out |

| the music restarts when returning to a screen | playTrack with the track already playing restarts it; don't call it again if the bed should carry on |

| stereo / position do nothing | the clip was loaded with stream: true — spatial is WebAudio-only |

| one format works, another does not | that codec is not in the audio.init() list |

| audio behind a login fails, or the preload hangs | loader.setOptions({ withCredentials: true }) — the same switch every asset type uses |

| a streamed clip ignores withCredentials | stream: true plays through an <audio> element; preload it buffered instead |

| audio.tone(440, {...}) does nothing / errors | tone takes a single options object with a required duration |

| tone / noise are silent with no error | no WebAudio context — getAudioContext() returned null |

| a positioned sound is silent, or only audible right next to the player | placed by hand with audio.position(), which keeps WebAudio's metre defaults (refDistance: 1) — use follow / at, or set refDistance in pixels with audio.panner() |

| a positioned sound does not follow the player | nothing is placing it — audio.play(name, { follow: entity }), and audio.setListener(player) once |

| a followed sound pans to the wrong side as the player moves | the listener was moved and positions are still relative to the player — with a listener in the world, pass absolute world coordinates |

| a looping sound outlives the entity that owns it | follow without stopWithTarget: true |

| a voice panned with stereo() loses its 3D placement, or the reverse | one panner node per voice and last writer wins — pick one per voice and stay with it |

| a reverb / compressor on the master gain does not affect the music | the music is stream: true, which bypasses the master gain — preload it buffered |

| a looping MP3 drops out for a moment at every loop, in Safari only | Safari keeps the encoder's padding at the end of the file: loop a sprite region of the clip's exact length (see "A loop with no gap") |

| sound effects keep playing after the window loses focus | pauseOnBlur and stopOnBlur are both off, so the engine is deliberately leaving the game running in the background |

Related skills

• melonjs-getting-started — preloading and asset ordering