This paper describes an auto-seeking microphone array consisting of novel multiple processing stages for speech acquisition. The array automatically detects the speech in the presence of noise and tracks the speech sources. A combination of dual-beamforming and post-beamformer enhancement significantly improves the quality and intelligibility of desired speech at the output of the system. Applications for such a system include hands-free telephony, teleconferencing and special situations where speech signals must be picked up in a noisy acoustic environment and the microphones are hidden (as, e.g., in a forensic covert recording system).