Second Biennial Hearing Aid Research and Development Conference
September 22-24, 1997 National Institutes of Health Bethesda, Maryland
Multi-microphone Directional Hearing Aid using Hybrid Adaptive Beamforming: Simulation Results and Plans for Real-world TestingG.L. Gibian, W.S. Koroljow, and S.T. Shaw, Planning Systems, Incorporated, McLean, Virginia
Planning Systems, Inc. (PSI) has developed signal processing technology intended to provide better speech comprehension in noisy environments. The technology uses multiple microphones and combines the advantages of two previous approaches, each with its own advantages under different listening conditions. A feasibility study employed computer simulation of room acoustics and non-real-time processing, and demonstrated the expected performance of PSI's signal processing method, termed the Hybrid Adaptive Beamformer (HAB). In the feasibility study [J. Acoust. Soc. Amer. 99:2517(A)], signals received at 5 microphones in a 4-inch line were simulated. Speech was separated from 12-talker babble originating from two different directions, using a frequency range of approximately 250 to 4000 Hz. Improvements in signal-to-noise ratio were assessed using intelligibility-weighted gain, as well as by listening to the signals before and after processing. Examples will be played from a CD.
PSI's HAB combines the advantages of optimal fixed-weight beamforming for diffuse noise, and adaptive beamforming to cancel specific interferers. Simulation results demonstrated intelligibility-weighted gains of: (1) 25 to 30 dB in an anechoic environment, like previous studies with adaptive beamformers; (2) 5 to 11 dB in strong reverberation, as good or better than previous fixed-weight beamformers; and (3) 12 to 18 dB in moderate reverberation for a jammer at 120 degrees and target-jammer ratios of -10 to 0 dB, exceeding performance of both adaptive and fixed-weight beamformers in previous studies.
PSI plans to build several wearable models of the HAB using DSP chips and compare their performance against conventional aids for hearing-impaired subjects. Some characteristics of the devices will be discussed, as well as tentative plans for test procedures and environments.
[Work supported by NIH.]
Return to Conference Abstracts List
Top |
|
|