Evaluation of Deep-Learning-Based Voice Activity Detectors and Room Impulse Response Models in Reverberant Environments

TL;DR

This paper evaluates deep-learning-based voice activity detectors in reverberant environments using a large augmented training set with simulated room impulse responses, showing significant performance improvements over anechoic training.

Contribution

It introduces a comprehensive augmented training dataset with reverberant data and compares multiple RIR models and VADs, highlighting the impact of reverberant training on detection accuracy.

Findings

20% average increase in accuracy, precision, and recall with reverberant training

One RIR model outperforms others across all VADs

One VAD consistently outperforms others in all tested environments

Abstract

State-of-the-art deep-learning-based voice activity detectors (VADs) are often trained with anechoic data. However, real acoustic environments are generally reverberant, which causes the performance to significantly deteriorate. To mitigate this mismatch between training data and real data, we simulate an augmented training set that contains nearly five million utterances. This extension comprises of anechoic utterances and their reverberant modifications, generated by convolutions of the anechoic utterances with a variety of room impulse responses (RIRs). We consider five different models to generate RIRs, and five different VADs that are trained with the augmented training set. We test all trained systems in three different real reverberant environments. Experimental results show $20\%$ increase on average in accuracy, precision and recall for all detectors and response models, compared to anechoic training. Furthermore, one of the RIR models consistently yields better performance than the other models, for all the tested VADs. Additionally, one of the VADs consistently outperformed the other VADs in all experiments.