Integrating remote microphone signals with hearing-aid processing

Abstract

Remote microphones (RM) are used with hearing aids to provide speech signals free from noise and room reverberation. However, because the RM signal bypasses the acoustics of the head and pinna, it can provide an unnatural signal that is lateralized at the center of the head rather than localized at the talker. A structural binaural model is combined with synthesized early room reflections to provide the missing externalization cues. The effectiveness of this simulation has been demonstrated over headphones using linear amplification to compensate for the hearing loss. In a hearing aid, however, the RM processed signal is combined with the acoustic inputs at the hearing-aid microphones, and the processed hearing-aid output is combined with the reverberated speech that is transmitted through the vent or open fitting. The hearing-aid processing implemented in the experiments described in this presentation includes both linear amplification and wide dynamic-range compression. The externalization, spatial soundfield characteristics, and sound quality provided by the entire signal-processing system are evaluated for normal-hearing and hearing-impaired listeners.