Multi-language Speech Synthesis for Voice Output Communication Aids using Eloquence

Many speech- and visually-impaired individuals can benefit from text-to-speech systems for unlimited vocabulary speech synthesis. The demands on these systems are especially severe in the case of voice output communication aids (VOCAs) for speech-impaired individuals, since the speech produced substitutes for a speaker's own voice. VOCA quality thus depends not only on intelligibility, but on the ability to express emotion and to provide voices of appropriate age, gender, and dialect. Eloquent Technology, Inc. (ETI) has addressed these concerns in Eloquence, a high-quality and flexible multi-voice text-to-speech program presently available for English. However, systems of equal quality and functionality for other languages are few, and text-to-speech software for major dialects often does not exist at all. In this project, ETI proposes to extend Eloquence to include both a North American and a European dialect of Spanish and French. In Phase I, ETI will demonstrate the feasibility of applying its innovative speech synthesis technology to a language other than English by adding Mexican Spanish to Eloquence. The resulting program for English and Spanish will be incorporated into a VOCA for testing by speech-impaied users.