Abstract
A new approach to improving speech intelligibility in noise is being developed for the hearing impaired. The approach is based on sinusoidal modeling of speech, in which the speech waveform is divided into over‐lapping segments, the FFT computed for each segment, and the N‐highest peaks are identified. The speech is then resynthesized using sinusoids having the amplitude and phase of the selected peaks, and the remaining spectral information is discarded. Using a small number of sinusoids results in a simplified speech signal; the most intense speech components of any given segment are reproduced, while the less intense speech and noise components are not. The question under consideration is how the speech intelligibility varies as the number of sinusoids is reduced. Normal‐ and hearing‐impaired listeners were asked to judge the intelligibility of sentences in quiet and at a 5‐dB speech‐to‐babble ratio; the results to be presented compare unprocessed speech with speech reproduced using 16, 8, and 4 sinusoids. [Work supported by NIH.]
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
