Contaminated speech training methods for robust DNN-HMM distant speech recognition

TL;DR

This paper revises contaminated speech training methods for robust distant speech recognition using DNN-HMM systems, proposing three novel techniques that significantly improve error rates in adverse acoustic conditions.

Contribution

It introduces asymmetric context windowing, close-talk supervision, and pre-training methods to enhance contaminated speech training for DNN-HMM based recognition.

Findings

15% error rate reduction with proposed methods

Effective on both real and simulated data

Works with small and large training sets

Abstract

Despite the significant progress made in the last years, state-of-the-art speech recognition technologies provide a satisfactory performance only in the close-talking condition. Robustness of distant speech recognition in adverse acoustic conditions, on the other hand, remains a crucial open issue for future applications of human-machine interaction. To this end, several advances in speech enhancement, acoustic scene analysis as well as acoustic modeling, have recently contributed to improve the state-of-the-art in the field. One of the most effective approaches to derive a robust acoustic modeling is based on using contaminated speech, which proved helpful in reducing the acoustic mismatch between training and testing conditions. In this paper, we revise this classical approach in the context of modern DNN-HMM systems, and propose the adoption of three methods, namely, asymmetric context windowing, close-talk based supervision, and close-talk based pre-training. The experimental results, obtained using both real and simulated data, show a significant advantage in using these three methods, overall providing a 15% error rate reduction compared to the baseline systems. The same trend in performance is confirmed either using a high-quality training set of small size, and a large one.