threejs-impl-audio

Create one AudioListener attached to the camera, then create a THREE.Audio instance and load a file with AudioLoader. Call play() only inside a user interaction handler after resuming the AudioContext, since browsers suspend audio until a user gesture.

Create a PositionalAudio with the shared listener, set its buffer via AudioLoader, configure refDistance, rolloffFactor, and a distance model, then add it as a child of the target mesh. Volume and panning update automatically as the camera moves.

Browsers suspend the AudioContext until a user gesture occurs, so play() called before interaction fails silently. Resume listener.context inside a click or keypress handler, and check sound.isPlaying before calling play() again.

Use 'inverse' for realistic gradual falloff that never reaches zero, 'linear' when you need a hard cutoff at maxDistance, and 'exponential' for steep close-versus-far contrast. Note that maxDistance only affects the linear model.

Yes, use setMediaElementSource() with an HTML5 audio element so the file streams rather than fully loading into an AudioBuffer. Control playback through the HTML element's play() and pause() methods, not sound.play().

Wrap an Audio instance with AudioAnalyser using a power-of-two fftSize, then call getFrequencyData() each frame to get a Uint8Array of bin magnitudes from 0 to 255. Use those values to scale meshes or drive other visual properties.