Abstract: A speech recognition system includes an input subsystem consisting of two or more simple microphones, each with a relatively large acceptance angle. The simple microphones are positioned so that the acceptance angles overlap by a relatively small amount. If signals received by each of the simple microphones exceed predetermined thresholds, then a switch is closed to permit a speech recognition subsystem to receive a signal for processing. An additional simple microphone, with a relatively wide acceptance angle centered on the overlap of the other microphones may be used to receive the signals that are fed to the speech recognition subsystem for processing.