Suppr超能文献

可接受噪声水平下的语音识别

Speech Recognition at the Acceptable Noise Level.

作者信息

Gordon-Hickey Susan, Morlas Holly

出版信息

J Am Acad Audiol. 2015 May;26(5):443-450. doi: 10.3766/jaaa.14079.

Abstract

BACKGROUND

The acceptable noise level (ANL) has been proposed as a prehearing aid fitting measure that could be used for hearing aid selection and counseling purposes. Previous work has demonstrated that a listener's ANL is unrelated to their speech recognition in noise abilities. It is unknown what criteria a listener uses when they select their ANL. To date, no research has explored the amount of speech recognized at the listener's ANL.

PURPOSE

To examine the amount of speech recognized at the listener's ANL to determine whether speech recognition in noise is utilized as a factor for setting ANL.

RESEARCH DESIGN

A descriptive quasi-experimental study was completed. For all listeners, ANL was measured and speech recognition in noise was tested at ANL and at two additional signal-to-noise ratio (SNR) conditions based on the listener's ANL (ANL + 5 and ANL - 5).

STUDY SAMPLE

Forty-four older adults served as participants. Twenty-seven participants had normal hearing and seventeen participants had mild to moderately-severe, symmetrical, sensorineural hearing loss.

DATA COLLECTION AND ANALYSIS

Acceptance of noise was calculated from the measures of most comfortable listening level and background noise level. Additionally, speech recognition in noise was assessed at three SNRs using the quick speech-in-noise test materials.

RESULTS

A significant interaction effect of SNR condition and ANL group occurred for speech recognition. At ANL, a significant difference in speech recognition in noise was found across groups. Those in the mid and high ANL groups had excellent speech recognition at their ANL. Speech recognition in noise at ANL decreased with ANL category.

CONCLUSIONS

For listeners with mid and high ANLs, speech recognition appears to play a primary role in setting their ANL. For those with low ANLs, speech recognition may contribute to setting their ANL; however, it does not appear to be the primary determiner of ANL. For those with very low ANLs, speech recognition does not appear to be significant variable for setting their ANL.

摘要

背景

可接受噪声水平(ANL)已被提议作为一种助听前适配测量方法,可用于助听器的选择和咨询目的。先前的研究表明,听众的ANL与他们在噪声中的言语识别能力无关。目前尚不清楚听众在选择自己的ANL时使用了什么标准。迄今为止,尚无研究探讨在听众的ANL水平下能够识别的言语量。

目的

研究在听众的ANL水平下能够识别的言语量,以确定噪声中的言语识别是否被用作设定ANL的一个因素。

研究设计

完成了一项描述性准实验研究。对于所有听众,测量了ANL,并在基于听众ANL的ANL、另外两个信噪比(SNR)条件(ANL + 5和ANL - 5)下测试了噪声中的言语识别。

研究样本

44名老年人作为参与者。27名参与者听力正常,17名参与者患有轻度至中度重度、对称的感音神经性听力损失。

数据收集与分析

根据最舒适聆听水平和背景噪声水平的测量值计算噪声接受度。此外,使用快速噪声言语测试材料在三个SNR下评估噪声中的言语识别。

结果

对于言语识别,SNR条件和ANL组之间存在显著的交互作用。在ANL水平上,各小组在噪声中的言语识别存在显著差异。中高ANL组的听众在其ANL水平下具有出色的言语识别能力。噪声中的言语识别随着ANL类别而降低。

结论

对于中高ANL的听众,言语识别似乎在设定他们的ANL中起主要作用。对于低ANL的听众,言语识别可能有助于设定他们的ANL;然而,它似乎不是ANL的主要决定因素。对于ANL非常低的听众,言语识别似乎不是设定他们的ANL的重要变量。

相似文献

1
Speech Recognition at the Acceptable Noise Level.可接受噪声水平下的语音识别
J Am Acad Audiol. 2015 May;26(5):443-450. doi: 10.3766/jaaa.14079.
4
Understanding excessive SNR loss in hearing-impaired listeners.理解听力受损听众中过度的信噪比损失。
J Am Acad Audiol. 2013 Apr;24(4):258-73; quiz 337-8. doi: 10.3766/jaaa.24.4.3.
6
Intertester reliability of the acceptable noise level.可接受噪声水平的测试者间可靠性。
J Am Acad Audiol. 2012 Jul-Aug;23(7):534-41. doi: 10.3766/jaaa.23.7.5.

文献检索

告别复杂PubMed语法,用中文像聊天一样搜索,搜遍4000万医学文献。AI智能推荐,让科研检索更轻松。

立即免费搜索

文件翻译

保留排版,准确专业,支持PDF/Word/PPT等文件格式,支持 12+语言互译。

免费翻译文档

深度研究

AI帮你快速写综述,25分钟生成高质量综述,智能提取关键信息,辅助科研写作。

立即免费体验