![]() |
Second ESCA/IEEE Workshop on Speech SynthesisSeptember 12-15, 1994 |
![]() |
This paper describes the intonation accent placement rules of Bridge, a computer programme which generates spoken dialogues from the communicative goals of two model agents. l By modelling the entire process of speech generation, from ideation to acoustics, BRIDGE can produce all the information required by the intonation rules. There are three main stages in the intonation component of Bridge: the assignment of prosodic boundaries, the computation of accentability, and the application of a rhythm rule. This paper concentrates on accentability.
An important point is that Bridge's intonation rules make no use of syntactic structure: our working hypothesis is that syntax is not relevant to determining intonation.
Bibliographic reference. Monaghan, Alex I. C. (1994): "Intonation accent placement in a concept-to-dialogue system", In SSW2-1994, 171-174.