US8024174B2 - Method and Apparatus for Training a Prosody Statistic Model and Prosody Parsing, Method and System for Text to Speech Synthesis