Speech Recognition in Echoic Environments and the Effect of Aging and Hearing Impairment
Nai Ding,Jiaxin Gao,Jing Wang,Wenhui Sun,Mingxuan Fang,Xiaoling Liu,Hua Zhao
DOI: https://doi.org/10.1016/j.heares.2023.108725
IF: 3.672
2023-01-01
Hearing Research
Abstract:Temporal modulations provide critical cues for speech recognition. When the temporal modulations are distorted by, e.g., reverberations, speech intelligibility drops, and the drop in speech intelligibility can be explained by the amount of distortions to the speech modulation spectrum, i.e., the spectrum of tempo-ral modulations. Here, we test a condition in which speech is contaminated by a single echo. Speech is delayed by either 0.125 s or 0.25 s to create an echo, and these two conditions notch out the temporal modulations at 2 or 4 Hz, respectively. We evaluate how well young and older listeners can recognize such echoic speech. For young listeners, the speech recognition rate is not influenced by the echo, even when they are exposed to the first echoic sentence. For older listeners, the speech recognition rate drops to less than 60% when listening to the first echoic sentence, but rapidly recovers to above 75% with ex-posure to a few sentences. Further analyses reveal that both age and the hearing threshold influence the recognition of echoic speech for the older listeners. These results show that the recognition of echoic speech cannot be fully explained by distortions to the modulation spectrum, and suggest that the audi-tory system has mechanisms to effectively compensate the influence of single echoes. (c) 2023 Elsevier B.V. All rights reserved.