Prosody Generation by Means of a Syntactic Approach and its Application in a Text to Speech System