This chapter describes how an artificial device, able to produce acoustic signals from articulatory motion, can learn to speak, i.e. coordinate its articulatory movements in such a way that it utters meaningful sequences of sounds belonging to a given language. This complex learning procedure, accomplished within a few years by the human child, is simulated in four major steps: (a) a babbling phase, where the device builds up a model of the forward kinematics, i.e. the articulatory-to-audio-visual mapping; (b) an imitation stage, where it tries to reproduce a limited set of sound sequences by audio-visual-to-articulatory inversion including a normalization procedure; (c) a "shaping" stage, where phonemes are associated with sensori-motor representation; and finally, (d) a "rhythmic" phase, where it learns the appropriate coordination of the activations of these sensori-motor targets. This artificial device has thus an ear which delivers both the control signals and the identification of pe...

This paper presents a methodology for evaluating the performance of synthetic prosody. The methodology focusses on the functional equivalence of natural and synthetic prosodic contours in terms of both distinctiveness and coherence. A perceptual experiment was designed to test the performance in terms of distinctiveness for a simple phrasing task. Results demonstrate the importance of a phonological representation that dynamically encodes fine details of the syntactic structure. Introduction Prosody shares at least two major functions with other linguistic processes (lexicon, morpho-syntax. . . ): segmentation and hierarchical structuration. Depending on the type of speaking style, prosody is more or less related to syntax. In a framework of isolated utterances read in French, both structures are close, and the generation of an adequate mapping between these two structures is therefore a fundamental aspect of prosody modelling. The prosodic module of a text-to-speech system has thus t...