Fact · Speech recognition
2seconds
The default voice activity filter in faster-whisper removes only silences longer than 2 seconds.
Checked against its source on October 17, 2026
In the source's own words
“The default behavior is conservative and only removes silence longer than 2 seconds.”
faster-whisper: README (GitHub, SYSTRAN), SYSTRAN, GitHub (SYSTRAN/faster-whisper), 2026, VAD filter.
Open the original- What it measures
- The filter uses the Silero VAD model. Passing vad_filter=True turns it on, and the README says it is enabled by default for batched transcription.
- What it does not show
- Parameters can be changed with vad_parameters. The default is deliberately conservative, so short pauses remain in the transcript.