Extraction and Representation of Prosody for Speaker, Speech and Language Recognition