Abstract:
A speech enhancer for eg. a vehicle (fig. 2) takes two audio signals from microphones 11, determines (via eg. a trained classifier 201) that the signals contain a speech component and determines (via eg. a source estimator/correlator 205) a location associated with that speech, and then outputs an audio signal 33 such that the speech is enhanced for a different location (eg. the drivers ear) than that associated with the speech via eg. a spatial processor 207 and a selective multichannel amplifier 209. Location parameters include Direction of Arrival (DOA) and delay time.