Boosting Distant Speech Recognition Using Multiple Microphones: Frontend Approaches