Direct-to-reverberant Energy Ratio Estimation Based on Interaural Coherence and a Joint ITD/ILD Model

This paper proposes a novel algorithm to estimate the direct-to-reverberant energy ratio (DRR) using hearing aid microphones. The algorithm is based on the interaural magnitude-squared coherence of signals and is able to take both phase and level differences of microphones signals in the binaural configuration into account. We employ a spherical head model to approximate binaural cues. The proposed algorithm uses the common assumption of an ideally diffuse reverberation sound field. We test our approach on signals based on simulated and on measured binaural room impulse responses. Results show improved performance of the proposed algorithm as compared to other coherence-based DRR estimation methods.

1. INTRODUCTION

DRR is an important measure to describe an acoustic transmission path and is beneficial in speech enhancement algorithms, e.g., for dereverberation [1], [2], [3], [4], [5] or for speaker distance estimation [6], [7], [8], [9]. Typically, DRR can be estimated from the measured room impulse response (RIR) which practically requires an intrusive measurement within the room or a blind estimation procedure. A common assumption made for DRR estimation is to discard the energy of the early reflections and to consider only the energy of the direct sound and the late reverberation which is modeled as a diffuse sound field. Therefore, several studies have introduced the coherent-to-diffuse ratio (CDR) [10], [11] as an estimator of DRR since a fully coherent source implicitly represents the direct component.