Method, apparatus, and manufacture for two-microphone array speech enhancement for an automotive environment

摘要

In a speech enhancement system for an automotive environment, signals from a first and second microphone of an array (fig. 2) are decomposed into subbands (352), processed (354 via eg. adaptive beamforming or decorrelation filtering) and then output based on an acoustic event detection (AED, 355) and current system mode (351, eg. driver-speech enhancement only) having determined whether the driver, passenger or neither is speaking. Each subband of this acoustics events detection output signal is then combined (356). Aspects of the invention include attenuating the signal from the passenger microphone when the driver is speaking, or vice versa, and calibrating the microphones.