Certainly it's still far from being able to deceive a human into thinking synthesized speech of any speaker saying anything is real, but it has definitely and clearly capture a certain quality to each of those voices. Really cool project and I'm sure it portends even more awesome work in the area.