Physical Interpretation of Word Gestalt Based on Invariant Representation of Speech and Its Application to Implement of Infants' Vocal Imitation * 1

In this paper we propose a new framework of speech generation by imitating “infants’ vocal imiation”. Most of the speech synthesizers take a phoneme sequence as input and generate speech by converting each of the phonemes. However infants usually acquire speech generation ability without text or phoneme. As developmental psychology states, from the utterances of their parents, they acquire the holistic sound pattern of words, called word Gestalt, and reproduce it with their vocal tubes. In our previous studies, the word Gestalt was defined physically and a method of extracting it from an utterance was proposed. In this paper, a method of converting the word Gestalt back to speech is proposed and evaluated. Unlike a reading machine, our proposal simulates infants’ vocal imiation.