A Study of Enhancement, Augmentation, and Autoencoder Methods for Domain Adaptation in Distant Speech Recognition

TL;DR

This paper reviews and experimentally compares various domain adaptation methods like enhancement, augmentation, and autoencoders to improve distant speech recognition from close-talking speech, highlighting their effectiveness and the performance gap.

Contribution

It provides a systematic evaluation of multiple domain adaptation techniques for distant speech recognition within a controlled experimental framework.

Findings

Different methods yield varying improvements under their assumptions

Quantifies the performance gap between close-talking and distant speech recognition

Sets a foundation for future adaptation research

Abstract

Speech recognizers trained on close-talking speech do not generalize to distant speech and the word error rate degradation can be as large as 40% absolute. Most studies focus on tackling distant speech recognition as a separate problem, leaving little effort to adapting close-talking speech recognizers to distant speech. In this work, we review several approaches from a domain adaptation perspective. These approaches, including speech enhancement, multi-condition training, data augmentation, and autoencoders, all involve a transformation of the data between domains. We conduct experiments on the AMI data set, where these approaches can be realized under the same controlled setting. These approaches lead to different amounts of improvement under their respective assumptions. The purpose of this paper is to quantify and characterize the performance gap between the two domains, setting up the basis for studying adaptation of speech recognizers from close-talking speech to distant speech. Our results also have implications for improving distant speech recognition.