Experiments of speech recognition in a noisy and reverberant environment using a microphone array and HMM adaptation

The use of a microphone array for hands-free continuous speech recognition in noisy and reverberant environment is investigated. An array of four omnidirectional microphones is placed at a distance of 1.5 m from the talker. Given the array signals, a time delay compensation (TDC) module provides a beam-formed signal that is shown to be effective as input to a hidden Markov model (HMM) based recognizer. Given a small number of sentences collected from a new speaker in a real environment, HMM adaptation further improves the recognition rate. These results are confirmed both by experiments conducted in a noisy office environment and by simulations. In the latter case, different SNR and reverberation conditions were recreated by using the image method to reproduce synthetic array microphone signals.