Experimental study of structure to speech conversion

Most of the speech synthesizers have been developed as text (phoneme sequence) to speech converters and, in this framework, text input is a precondition for speech production. However, we can say that no child acquires spoken language by reading a given text out. Children are explained to acquire spoken language by imitating the utterances of their parents but they never imitate the voices of their parents. Developmental psychology claims that they extract a holistic and speaker-invariant sound pattern embedded in a given utterance, called word Gestalt, and realize the pattern acoustically using their short vocal tubes. In our previous studies, we mathematically defined this holistic and speaker-invariant pattern and used it for ASR. Here, we experimentally implement its inverse process, i.e. Gestalt-to-utterance conversion, on a computer.