Phonetic segment switching based on differential-encoding of the vocal tract resonances Brad H. Story and Kate Bunton