Reactive Speech Synthesis: Actively Managing Phonetic Contrast along an H&H Continuum.
Roger K. Moore, Mauro Nicolao · ICPhS · 2011
Notwithstanding the tremendous progress that has taken place in the science and technology of textto-speech synthesis, state-of-the-art systems still exhibit a rather limited range of speaking styles as well as an inability to adapt to the listening conditions in which they operate. Such behaviour is typical of human speech production and, in 2006, Moore proposed a new type of ‘reactive’ speech synthesizer that would monitor the effect of its output and modify its characteristics in order to maximize its communicative intentions. This paper presents an investigation into reactive speech synthesis based on the hypothesis that the neutral vowel represents a low energy attractor for a human speech production system, and that interpolation/extrapolation along the key dimension of hypo/hyper-articulation can be motivated by energetic considerations of phonetic contrast. The results indicate that the intelligibility of synthesized speech can be manipulated successfully, thus confirming the potential of reactive speech synthesis.