What ASR backend did you guys use? I tried to input your audios from the demo page to Whisper Large v3, but the text from the enhanced audio missed some charter. In general, most of ASR model are trained with noisy speech dataset. I don't quiet understand...
What ASR backend did you guys use? I tried to input your audios from the demo page to Whisper Large v3, but the text from the enhanced audio missed some charter. In general, most of ASR model are trained with noisy speech dataset. I don't quiet understand...