DEVELOPMENT OF AN IMPROVED TACTUAL VOCODER

WE WILL DEVELOP A DIGITAL TACTUAL VOCODER SIGNIFICANTLY IMPROVED OVER EXISTING COMMERCIALLY AVAILABLE DEVICES. CURRENT VOCODERS HAVE BEEN PROVEN PARTIALLY EFFECTIVE BUT HAVE SEVERAL SIGNIFICANT PROBLEMS. THESE PROBLEMS INCLUDE (1) THE FAILURE TO TRANSMIT ADEQUATELY INFORMATION ABOUT VOICE FUNDAMENTAL FREQUENCY AND (2) THE FAILURE TO TRANSMIT RAPIDLY CHANGING BRIEF COMPONENTS OF THE SPEECH SIGNAL, NAMELY FORMANT TRANSITIONS. BOTH VOICE FUNDAMENTAL FREQUENCY AND FORMANT TRANSITIONS CARRY A SIGNIFICANT PORTION OF THE SPEECH CODE. THE RESEARCH SEEKS TO TEST THE FEASIBILITY OF NEW APPROACHES TO THE PROBLEMS OF TRANSMISSION OF FUNDAMENTAL FREQUENCY AND RAPIDLY CHANGING BRIEF SPECTRAL EVENTS BY 1) DEVELOPING SIGNAL PROCESSING STRATEGIES THAT CAN EFFICIENTLY AND ACCURATELY EXTRACT BOTH FORMANT AND FUNDAMENTAL FREQUENCY INFORMATION IN REAL TIME AND 2) DESIGNING A NEW CLASS OF STIMULATORS CAPABLE OF TRANSMITTING FUNDAMENTAL FREQUENCY AS FREQUENCY OF VIBRATORY STIMULATION. THE CHARACTERISTICS OF THE VIBRATORS AND SIGNAL PROCESSING STRATEGIES WILL BE EVALUATED USING EXPERIMENTAL PSYCHOPHYSICAL METHODOLOGY WITH SIMULATED SPEECH STIMULI.