Basic params: rate/pitch/volume
Voice fonts: name/gender/age
Prerecorded speech is used when you know all the phrases that the application will use
treated as special kind of voice font, using longest match to select phrases from dictionary
Prosody: making speech sound natural
W3C's Aural Style sheets (ACSS) -- style properties for rendering XML to speech
SABLE -- XML format for driving TTS engines, has been proposed for use with VXML
W3C will work on harmonizing ACSS and SABLE