Diminishing Domain Mismatch for DNN-Based Acoustic Distance Estimation via Stochastic Room Reverberation Models

TL;DR

This paper introduces a novel approach combining geometric and stochastic modeling of room impulse responses to reduce domain mismatch, thereby improving the accuracy of DNN-based acoustic distance estimation.

Contribution

The paper proposes a new method that combines geometric and stochastic models to generate more realistic RIRs, reducing domain mismatch in DNN training.

Findings

Significant improvement in distance estimation accuracy.

Reduced domain mismatch between simulated and real RIRs.

Enhanced robustness of DNN-based acoustic distance estimation.

Abstract

The room impulse response (RIR) encodes, among others, information about the distance of an acoustic source from the sensors. Deep neural networks (DNNs) have been shown to be able to extract that information for acoustic distance estimation. Since there exists only a very limited amount of annotated data, e.g., RIRs with distance information, training a DNN for acoustic distance estimation has to rely on simulated RIRs, resulting in an unavoidable mismatch to RIRs of real rooms. In this contribution, we show that this mismatch can be reduced by a novel combination of geometric and stochastic modeling of RIRs, resulting in a significantly improved distance estimation accuracy.