Hello all,
I have trained a Mogolian speach synthesis model using HTS, and the result is very noisy.
I am using mgc and lf0 coefficients as my training feactures, I guess there was some parameter
was not appropriate when I extracting these feactures. Is there any parameter that I can change
to get a better result?
My speech analysis parameter is setted as bellow.
Thanks very much.