Fact · Languages and accents
The faster-whisper transcribe method has a language_detection_segments option that sets how many segments are considered for language detection.
Checked against its source on October 17, 2026
In the source's own words
“language_detection_segments: Number of segments to consider for the language detection.”
faster_whisper/transcribe.py (SYSTRAN faster-whisper, master branch), SYSTRAN, GitHub, SYSTRAN/faster-whisper, 2026, transcribe() parameters.
Open the original- What it measures
- In the same signature the default is language_detection_segments: int = 1, and the detection threshold defaults to 0.5.
- What it does not show
- The option changes which audio is used for detection. The code does not show how this affects accuracy.