MAGE : A Platform for Performative Speech Synthesis - New Approach in Exploring Applications Beyond Text-To-Speech

Speech is the richest and most ubiquitous modality of communication used by human beings. Voice production is one of the most expressive instruments of the human body and people have always been fascinated for artificial speech and singing production. Even though vocal behavior and expression is a very complex mechanism, we realize a highly interactive and social process. Until now, artificial voice production has been based on the Text-To-Speech (TTS) technologies, converting a static text into an intelligible and natural waveform, lately with great success. However the relevance of speech production is not only based on these properties, it is also linked to the context, to an ongoing process of interaction between speakers and to their social environment and background. Through this interdisciplinary redefinition of expressive speech, we see that the nature of voice production is primarily a realtime, dynamic gesture, involving the vocal organs, face and body. Speech is a performance, a gestural phenomenon that transmits messages with both information and emotions.