Modelling Prominence and Emphasis Improves Unit-Selection Synthesis

We describe the results of large scale perception experiments showing improvements in synthesising two distinct kinds of prominence: standard pitch-accent and strong emphatic accents. Previously prominence assignment has been mainly evaluated by computing accuracy on a prominence-labelled test set. By contrast we integrated an automatic pitch-accent classifie r into the unit selection target cost and showed that listeners preferred these synthesised sentences. We also describe an improved recording script for collecting emphatic accents, and show that generating emphatic accents leads to further improvements in the fict ion genre over incorporating pitch accent only. Finally, we show diffe rences in the effe cts of prominence between childdirected speech and news and fict ion genres.