>>42034585The voices are consistent, problem is the way Japanese is spoken is radically different from English. Here's one of my PPT attempts, it starts well but just goes full murrican, but the voice is kinda there
https://files.catbox.moe/c8dlr5.mp3And of course it sounds nothing like her English attempts because those attempts sound nothing like a native English speaking person, that's kinda the point, you won't be able to replicate such a backwards way of speaking unless the model itself is specifically aimed at it.
