Glossary · also: VAD
Voice activity detection
Software that decides, from audio alone, when speech starts and stops.
What it means
Voice activity detection watches the loudness and character of incoming audio — an RMS gate, in the simplest form — and segments it: speech above a threshold begins a passage, sustained silence ends it.
Why it matters
VAD is what makes hands-free dictation possible: no key to hold, no button to press, just speaking and stopping. It is also the boundary between dictation and the always-on assistants, because a VAD gate that only listens while you speak is not surveillance.
How notchos does it
The gate runs inside the local audio tap: speech starts a segment, about a second and a quarter of silence ends it and inserts the text.
Fair questions
- Will it trigger on music or conversation nearby?
- The gate needs sustained speech-shaped audio above threshold; a moment of noise does not start a session.
- Is there a wake word?
- No — and deliberately none. You start speaking; the silence is what stops you.
Related terms
Twenty tools. One hover. $9.99, once.