More predictable, but can be monotonous.

Higher values are more like the original voice, but can have artifacts.