Fact · Speech recognition
1millisecond
The Silero VAD project says one audio chunk of 30 ms or more takes less than 1 millisecond to process on a single CPU thread.
Checked against its source on October 16, 2026
In the source's own words
“One audio chunk (30+ ms) takes less than 1ms to be processed on a single CPU thread.”
Silero VAD (GitHub repository), Silero Team, GitHub (snakers4/silero-vad), 2026, README.
Open the original- What it measures
- A voice activity detector decides which stretches of audio contain speech so a recognizer can skip silence.
- What it does not show
- The authors' own figure, hardware-dependent.