System and Method for Characterizing Voiced Excitations of Speech and Acoustic Signals, Removing Acoustic Noise from Speech, and Synthesizing Speech
Greg C. Burnett, John F. Holzrichter, Lawrence C. Ng
U.S. Patent 7,089,177 B2
August 8, 2006
This system can characterize human or animate voiced excitation functions and acoustic signals in order to remove unwanted acoustic noise that often occurs when a speaker uses a microphone in common environments. The system can also synthesize personalized or modified human or other animate speech on command from a controller. A low-power electromagnetic (EM) sensor detects the motions of windpipe tissues in the glottal region of the human speech system before, during, and after a user produces speech. A voiced excitation function is derived from these tissue motion measurements. The excitation function provides information to enhance noise removal from human speech and allows accurate transfer functions of speech to be obtained. Previously stored excitation and transfer functions can be used to synthesize personalized or modified human speech. The EM sensor and acoustic microphone systems can be configured to enhance noise cancellation and enable multiple articulator measurements.