How Bad Are Artifacts?: Analyzing the Impact of Speech Enhancement Errors on ASR

TL;DR

This paper analyzes how artifacts from speech enhancement impact automatic speech recognition, identifying artifacts as the main degradation source and proposing methods to mitigate their effect to improve ASR accuracy.

Contribution

It introduces an orthogonal projection-based decomposition to isolate noise and artifact components of SE errors and demonstrates that mitigating artifacts significantly enhances ASR performance.

Findings

Artifact components are the main cause of ASR degradation.

Adding scaled observed signals improves ASR by increasing signal-to-artifact ratio.

Mitigating artifacts leads to substantial ASR performance gains.

Abstract

It is challenging to improve automatic speech recognition (ASR) performance in noisy conditions with single-channel speech enhancement (SE). In this paper, we investigate the causes of ASR performance degradation by decomposing the SE errors using orthogonal projection-based decomposition (OPD). OPD decomposes the SE errors into noise and artifact components. The artifact component is defined as the SE error signal that cannot be represented as a linear combination of speech and noise sources. We propose manually scaling the error components to analyze their impact on ASR. We experimentally identify the artifact component as the main cause of performance degradation, and we find that mitigating the artifact can greatly improve ASR performance. Furthermore, we demonstrate that the simple observation adding (OA) technique (i.e., adding a scaled version of the observed signal to the enhanced speech) can monotonically increase the signal-to-artifact ratio under a mild condition. Accordingly, we experimentally confirm that OA improves ASR performance for both simulated and real recordings. The findings of this paper provide a better understanding of the influence of SE errors on ASR and open the door to future research on novel approaches for designing effective single-channel SE front-ends for ASR.