Add WV-MOS

[lars76]

Add WV-MOS from https://arxiv.org/pdf/2203.13086 Code is here: https://github.com/AndreevP/wvmos/tree/main

Also relevant is: https://www.arxiv.org/pdf/2407.12707 On “TTS Arena” UTMOSv1 has only a weak correlation with the leaderboard, while WVMOS has much better results. I haven't tested UTMOSv2 and WVMOS yet. But UTMOSv1 does not necessarily lead to a correct voice quality evaluation in my experiments.

[Satellite30]

I have already used these model in my work. My work is clean the different language audio data. The WV-MOS will predict the score lke -0.1, 0.123...。i listen these audios, many of them are noise or over-noise reduction。if i use 0.1 as the interval, such as 2.0, 2.1, 2.2。the differences between them is hard to seperate. if you choose 2.0, 3.0 to listen, you can get a better distinguish results. Besides，in the multilingual audio clips, some language always get a lower score comparing to the american english, like chinese, Cantonese，even British English。
However, The UTMOSv1 model seems has some different with the WV-MOS。it seems to seperate the score more distinguishable. Of coure，it also have the different score in the different language. Besides, I haven't checked if the UTMOSv1 model will output negative numbers.