I am working on my FYP to build a text to speech in Urdu language. i currently building my data set. i want to know which audio features is best for machine learning. i try MFCC but when i reconstruct its audio the result is very bad.
I know how to do it for one feature. but i want to use more features in my data set but i do not know how to use two or more features output and convert it into audio.
Means the how output audio features will reconstruct into audio file.
any python library?
please help me its urgent