CISE Research Instrumentation: Speech Synthesis from Fluid Dynamic Principles, with Application to Human-Computer Interaction

9818313

Flanagan, James L.

Krane, Michael

Rutgers University - Busch Campus

CISE Research Instrumentation: Speech Synthesis from Fluid Dynamic Principles,

with Application to Human-Computer Interaction

This research instrumentation enables research projects in:

- Synergistic Multi-modal Communication in Collaborative Multi-user Environments, and

- Speech Synthesis Based upon Fluid Dynamics Principles

To support the aforementioned projects, this award contributes to the purchase of image acquisition, networking, visualization, and speech processing equipment at Rutgers University CAIP Center.

Advanced human-machine interfaces employ multiple modalities through which a human user can communicate effectively with computers and complex information systems. Interactive speech (in the form of automatic recognition and synthesis) is a natural modality, which improves the effectiveness of human-machine interfaces. However, advances in speech technology have been limited due to incomplete understanding of the physical mechanisms of human speech generation. A particular deficit has been the lack of naturalness in speech synthesis. Equipment to be instrumented under this proposal will directly enable research on the synthesis of natural quality speech using a more complete description of the fundamental physics of speech generation (NSF contract #IRI-98-00999). The instrumentation will also support research on multi-modal workstations with which human users interact using sight, sound, and touch (NSF contract #IRI-96-18854). Additionally, the proposed instrumentation will enhance the effectiveness of the speech component in multi-modal interfaces through visual tracking. And, the speech interface will directly benefit from a machine voice that exhibits improved human-like quality.