factory method const analyserNode = audioCtx. The Web Audio API will up-mix or down-mix the number of channels as required check the Web Audio spec for details.įor a list of all audio nodes, see the Web Audio API homepage. For example, mono audio has one channel, while stereo audio has two channels. For example, a volume control ( GainNode) should be the last node so that volume changes take immediate effect.Įach input and output has a given amount of channels. Therefore, if you need to have interactive audio, keep the graph as small as possible, and put user-controlled audio nodes at the end of a graph. For example, if your graph has a latency of 500ms, when the source node plays a sound, it will take half a second until that sound can be heard on your speakers (or even longer because of latency in the underlying audio device).

The more nodes in a graph, the higher the latency will be. The exact processing done varies from one AudioNode to another but, in general, a node reads its inputs, does some audio-related processing, and generates new values for its outputs, or lets the audio pass through (for example in the AnalyserNode, where the result of the processing is accessed separately). In addition, there are processing nodes which have inputs and outputs.

On the other hand, a destination node has no outputs instead, all its inputs are directly played back on the speakers (or whatever audio output device the audio context uses). This graph is contained in an AudioContext, and each audio node can only belong to one audio context.Ī source node has zero inputs but one or multiple outputs, and can be used to generate sound. Each AudioNode has inputs and outputs, and multiple audio nodes are connected to build a processing graph.

Nodejs audiodevice