The VAD parameters and default values in the source code is inconsistent with the description in README.md

[wikios]

Hello, I found that there is a description about VAD filter usage in README.md that may be inconsistent with the source code. I think "removes silence longer than 2 seconds" should probably use the argument min_speech_duration_ms rather than min_silence_duration_ms according to the source code.

In README.md :

VAD filter

... The default behavior is conservative and only removes silence longer than 2 seconds. See the available VAD parameters and default values in the source code. They can be customized with the dictionary argument vad_parameters:

segments, _ = model.transcribe(
   "audio.mp3",
   vad_filter=True,
   vad_parameters=dict(min_silence_duration_ms=500),
)

I referred to the "Attributes" in vad.py :

https://github.com/SYSTRAN/faster-whisper/blob/master/faster_whisper/vad.py , Line 21, 26

min_speech_duration_ms: Final speech chunks shorter min_speech_duration_ms are thrown out. min_silence_duration_ms: In the end of each speech chunk wait for min_silence_duration_ms before separating it

and utils_vad.py in snakers4/silero-vad :

https://github.com/snakers4/silero-vad/blob/master/utils_vad.py , Line 205, 213

min_speech_duration_ms: int (default - 250 milliseconds) Final speech chunks shorter min_speech_duration_ms are thrown out min_silence_duration_ms: int (default - 100 milliseconds) In the end of each speech chunk wait for min_silence_duration_ms before separating it

Please let me know if this understanding is correct, looking forward to a reply, thanks~

[Purfview]

Please let me know if this understanding is correct

Your understanding is incorrect. There is nothing wrong with VAD parameters nor description.