concept · Web Audio API

Latency Benchmark

Compare baseLatency and outputLatency across three AudioContexts: one with the default 128-frame quantum, one with renderSizeHint: "hardware", and one with a custom 256-frame quantum. The reading that best aligns with your hardware buffer will show the lowest output latency.

renderSizeHint: "default"

renderQuantumSize

—

baseLatency

—

outputLatency

—

renderSizeHint: "hardware"

renderQuantumSize

—

baseLatency

—

outputLatency

—

renderSizeHint: 256

renderQuantumSize

—

baseLatency

—

outputLatency

—

baseLatency reflects the quantum itself — the time for one render block to process. outputLatency includes the hardware buffer delay on top of that. When the quantum aligns with the hardware buffer, no extra buffering occurs and outputLatency is minimised. The "hardware" hint is the easiest path to this alignment: just ask the browser to pick for you.

reading the metrics

const ctx = new AudioContext({ renderSizeHint: 'hardware' });

// Actual quantum chosen by the browser
console.log(ctx.renderQuantumSize); // e.g. 256

// Latency of one render quantum in seconds
console.log(ctx.baseLatency);       // renderQuantumSize / sampleRate

// Total output pipeline latency (quantum + hardware buffer)
console.log(ctx.outputLatency);     // ideally ≈ baseLatency when aligned

implementation reference

Need the exact API surface, compatibility boundaries, errors, lifecycle, and source links? Read the matching gendn reference ↗

reading the metrics

see also

implementation reference