So wil tts turned out interesting his English is better than yuu but his japanese is a lot worse. I think the English samples are warping his japanese might split them up and have wil English tts and japanese tts. Thinking about cutting up some of his ex voice samples to make japanese particles it's kinda weird they didn't have them.
I present to you all a tts trained on yuu's ex voice samples all 2 mins of data. He is a little a rough but It's better than nothing and I have plans to improve him in the future. Please tell me how it sounds.
The training on big al was successful but he's starting to sound less like big Al and more like his voice provider. He now has trouble telling which vowels to use ex. oo in soon sounds like sun might look into training him on a phonetic transcription