FunASR icon indicating copy to clipboard operation
FunASR copied to clipboard
Published 4 months ago • verified publisher icon modelscope

Fsmn_vad参数speech_to_sil_time_thres

Open tesfayetong opened this issue 1 year ago • 1 comments

❓ Questions and Help

为啥speech_to_sil_time_thres设置的越大，切割出的音频会越多呢？

tesfayetong avatar Jul 30 '24 09:07 tesfayetong

这个参数的意思是声音到静音的最长时间阈值么？为什么我设置的越小，反而切分出的结果也越少了呢，照理说阈值减小了应该越多？

tesfayetong avatar Jul 31 '24 03:07 tesfayetong