This is impressive. I wish there were a voice clone option.

With so few parameters, I imagine a voice fine-tune might be readily tractable.