Microphone array power ratio for quality assessment of reverberated speech

Abstract

Speech signals in enclosed environments are often distorted by reverberation and noise. In speech communication systems with several randomly distributed microphones, involving a dynamic speaker and unknown source location, it is of great interest to monitor the perceived quality at each microphone and select the signal with the best quality. Most of existing approaches for quality estimation require prior information or a clean reference signal, which is unfortunately seldom available. In this paper, a practical non-intrusive method for quality assessment of reverberated speech signals is proposed. Using a statistical model of the reverberation process, we examine the energies as measured by unidirectional elements in a microphone array. By measuring the power ratio, we obtain a measure for the amount of reverberation in the received acoustic signals. This measure is then utilized to derive a blind estimation of the direct-to-reverberation energy ratio in the room. The proposed approach attains a simple, reliable, and robust quality measure, shown here through persuasive simulation results.