Fairy tale storytelling system: Using both prosody and text for emotional speech synthesis

Abstract

It is rather obvious that prosody and text play crucial roles in successful emotional speech synthesis systems. However, the effects of prosody structure and word semantics are still studied independently despite their combined importance. In this paper, we propose an interactive fairy tale storytelling system that can express emotions too. For this system, we classify emotions into prosody-sensitive emotions and word-semantics-sensitive emotions, following the characteristics of individual emotions as analyzed from several experiments. As a result, we have found that joy is predominantly a word-semantics-sensitive emotion, which is why it is very hard to express the emotion with a prosody structure alone. © 2012 Springer-Verlag.