Prosody model in a Mandarin text-to-speech system based on a hierarchical approach | IEEE Conference Publication | IEEE Xplore