AudioWorkletProcessor

The AudioWorkletProcessor interface of the Web Audio API represents an audio processing code behind a custom AudioWorkletNode. It lives in the AudioWorkletGlobalScope and runs on the Web Audio rendering thread. In turn, an AudioWorkletNode based on it runs on the main thread.

Constructor

AudioWorkletProcessor()

Creates a new instance of an AudioWorkletProcessor object.

Properties

port Read only

Returns a MessagePort used for bidirectional communication between the processor and the AudioWorkletNode which it belongs to. The other end is available under the port property of the node.

Methods

The AudioWorkletProcessor interface does not define any methods of its own. However, you must provide a process() method, which is called in order to process the audio stream.

Events

The AudioWorkletProcessor interface doesn't respond to any events.

Usage notes

Deriving classes

To define custom audio processing code you have to derive a class from the AudioWorkletProcessor interface. Although not defined on the interface, the deriving class must have the process method. This method gets called for each block of 128 sample-frames and takes input and output arrays and calculated values of custom AudioParams (if they are defined) as parameters. You can use inputs and audio parameter values to fill the outputs array, which by default holds silence.

Optionally, if you want custom AudioParams on your node, you can supply a parameterDescriptors property as a static getter on the processor. The array of AudioParamDescriptor-based objects returned is used internally to create the AudioParams during the instantiation of the AudioWorkletNode.

The resulting AudioParams reside in the parameters property of the node and can be automated using standard methods such as linearRampToValueAtTime. Their calculated values will be passed into the process() method of the processor for you to shape the node output accordingly.

Processing audio

An example algorithm of creating a custom audio processing mechanism is:

1. Create a separate file;
2. In the file:
   1. Extend the AudioWorkletProcessor class (see "Deriving classes" section) and supply your own process() method in it;
   2. Register the processor using AudioWorkletGlobalScope.registerProcessor() method;
3. Load the file using addModule() method on your audio context's audioWorklet property;
4. Create an AudioWorkletNode based on the processor. The processor will be instantiated internally by the AudioWorkletNode constructor.
5. Connect the node to the other nodes.

Examples

In the example below we create a custom AudioWorkletNode that outputs white noise.

First, we need to define a custom AudioWorkletProcessor, which will output white noise, and register it. Note that this should be done in a separate file.

// white-noise-processor.js
class WhiteNoiseProcessor extends AudioWorkletProcessor {
  process (inputs, outputs, parameters) {
    const output = outputs[0]
    output.forEach(channel => {
      for (let i = 0; i < channel.length; i++) {
        channel[i] = Math.random() * 2 - 1
      }
    })
    return true
  }
}

registerProcessor('white-noise-processor', WhiteNoiseProcessor)

Next, in our main script file we'll load the processor, create an instance of AudioWorkletNode, passing it the name of the processor, then connect the node to an audio graph.

const audioContext = new AudioContext()
await audioContext.audioWorklet.addModule('white-noise-processor.js')
const whiteNoiseNode = new AudioWorkletNode(audioContext, 'white-noise-processor')
whiteNoiseNode.connect(audioContext.destination)

Specifications

Browser compatibility

See also

• Web Audio API
• Using the Web Audio API