Closed lcxsnow closed 3 years ago
I think the duration of the audio file is too short.
It caused that PESQ cannot find the correct alignment between degraded and reference audio.
But I have other files which are e same as these data.size. They can be PESQ.
Of course that can happen,as the duration may not the only reason.
But,The NoUtteranceError will be raised only if the alignment failed.
To note that,this is not a bug. It can be expected that audios cannot be aligned in some situations.
If you want to know more details,you can read the specification of PESQ (ITU P.862)