SYSTRAN / faster-whisper

Faster Whisper transcription with CTranslate2
MIT License
12.6k stars 1.05k forks source link

fix:The timestamps for the second and subsequent segments are incorrect #980

Closed caiwuu closed 4 weeks ago

caiwuu commented 2 months ago

When VAD detects multiple segments of speech in an audio clip, the timestamps from the second segment onward are incorrect, as shown in the image below This is incorrect: image This is correct: image