public final class ConvolverNode

Convolves its input with an impulse response, for reverb and cabinet or room simulation: the Web Audio ConvolverNode.

The response is a 1, 2 or 4 channel AudioData at the context’s sample rate. One channel is applied to each input channel; two are a left and a right response; four are true stereo (left to left, left to right, right to left, right to right). The output is mono only for a mono response on a mono input, otherwise stereo.

OfflineAudioContext audio = new OfflineAudioContext(2, 44100, 44100);
OscillatorNode dry = audio.createOscillator();
GainNode wetGain = audio.createGain();
AudioData impulse = audio.createBuffer(2, 22050, 44100); // decoded or synthesised response
ConvolverNode hall = audio.createConvolver();
hall.setBuffer(impulse);
dry.connect(hall).connect(wetGain).connect(audio.getDestination());

Convolution is uniformly partitioned in the frequency domain, one 128 frame block at a time, so there is no added latency. Work grows linearly with the response’s length; a few seconds of stereo reverb is practical, a whole cathedral on a low-end phone may not be.