European Conference on Speech Technology
Edinburgh, Scotland, UK
The high degree of standardization that can be achieved in the intonational synthesis by rule of isolated utterances may not be applicable to that of longer texts, because the listener may become aware of the unnatural recurrence of invariably the same melodic recipe. Two experiments were run in which listeners evaluated the naturalness of a prose text that had been intoned with varying degrees of melodic standardization.
Bibliographic reference. Collier, Rene / Terken, Jacques (1987): "Intonation by rule in text-to-speech applications", In ECST-1987, 2165-2168.