All termsGlossary

What is voice activity detection (VAD)?

Voice activity detection is the process of deciding, moment by moment, whether an audio signal contains human speech or only silence and background noise.

Why it matters

A voice agent needs to know when a caller has started and stopped talking. If it judges too early, it cuts people off. If it waits too long, the conversation feels slow. Voice activity detection provides that signal. It also saves processing, because only the parts of a call containing speech need to be sent for transcription.

How it works

The audio is examined in very short frames, a small fraction of a second each. For every frame, a detector looks at features such as energy and pitch pattern, or uses a trained model, and marks it as speech or not speech. Short gaps are smoothed over so a brief pause inside a sentence is not treated as the end of speaking.

In practice

A caller is on a street when a voice agent asks for their postcode. A lorry passes before they answer. Voice activity detection marks the lorry as noise, so the agent keeps waiting. When the caller speaks, the detector marks the start, and after a short silence it marks the end. Only that stretch of audio goes to speech recognition.

Get Started

Let’s build AI agents, together.

Talk with us to see the top three use cases for AI agents at your business, and how quickly you could go live.

Live in 7 days · Works inside the tools your team already uses · Hands off with full context