Formal expressive indiscernibility underlying a prosodic deformation model
Résumé
We are here concerned by the setting up of a model and a formalism for expressive speech synthesis under the paradi gm of a corpus-based appr oach. Our objective is to apply prosodic expressive forms, acquired from natural human-reading recordings, on a new textual matter. We outline a general model for speech expressiveness. Then we deal with some formal aspects of expressive representation. We point out the core transformational aspects and the indiscernibility criteria allowing comparisons be- tween forms. We finish by some interpretational issues of such an approach.