Suppr超能文献

频谱异步延迟对辅音浊音感知的影响。

Impact of spectrally asynchronous delays on consonant voicing perception.

作者信息

Ortmann Amanda J, Palmer Catherine V, Pratt Sheila R

机构信息

Department of Communication Science and Disorders, University of Pittsburgh, USA.

出版信息

J Am Acad Audiol. 2010 Sep;21(8):493-511. doi: 10.3766/jaaa.21.8.2.

Abstract

BACKGROUND

A possible voicing cue used to differentiate voiced and voiceless cognate pairs is envelope onset asynchrony (EOA). EOA is the time between the onsets of two frequency bands of energy (in this study one band was high-pass filtered at 3000 Hz, the other low-pass filtered at 350 Hz). This study assessed the perceptual impact of manipulating EOA on voicing perception of initial stop consonants, and whether normal-hearing and hearing-impaired listeners were sensitive to changes in EOA as a cue for voicing.

PURPOSE

The purpose of this study was to examine the effect of spectrally asynchronous auditory delay on the perception of voicing associated with initial stop consonants by normal-hearing and hearing-impaired listeners.

RESEARCH DESIGN

Prospective experimental study comparing the perceptual differences of manipulating the EOA cues for two groups of listeners.

STUDY SAMPLE

Thirty adults between the ages of 21 and 60 yr completed the study: 17 listeners with normal hearing and 13 listeners with mild-moderate sensorineural hearing loss.

DATA COLLECTION AND ANALYSIS

The participants listened to voiced and voiceless stop consonants within a consonant-vowel syllable structure. The EOA of each syllable was varied along a continuum, and identification and discrimination tasks were used to determine if the EOA manipulation resulted in categorical shifts in voicing perception. In the identification task the participants identified the consonants as belonging to one of two categories (voiced or voiceless cognate). They also completed a same-different discrimination task with the same set of stimuli. Categorical perception was confirmed with a d-prime sensitivity measure by examining how accurately the results from the identification task predicted the discrimination results. The influence of EOA manipulations on the perception of voicing was determined from shifts in the identification functions and discrimination peaks along the EOA continuum. The two participant groups were compared in order to determine the impact of EOA on voicing perception as a function of syllable and hearing status.

RESULTS

Both groups of listeners demonstrated a categorical shift in voicing perception with manipulation of EOA for some of the syllables used in this study. That is, as the temporal onset asynchrony between low- and high-frequency bands of speech was manipulated, the listeners' perception of consonant voicing changed between voiced and voiceless categories. No significant differences were found between listeners with normal hearing and listeners with hearing loss as a result of the EOA manipulation.

CONCLUSIONS

The results of this study suggested that both normal-hearing and hearing-impaired listeners likely use spectrally asynchronous delays found in natural speech as a cue for voicing distinctions. While delays in modern hearing aids are less than those used in this study, possible implications are that additional asynchronous delays from digital signal processing or open-fitting amplification schemes might cause listeners with hearing loss to misperceive voicing cues.

摘要

背景

一种可能用于区分浊音和清音同源词对的发声线索是包络起始异步性(EOA)。EOA是两个能量频段起始之间的时间(在本研究中,一个频段在3000 Hz进行高通滤波,另一个频段在350 Hz进行低通滤波)。本研究评估了操纵EOA对初始塞音发声感知的影响,以及听力正常和听力受损的听众对EOA变化作为发声线索的敏感程度。

目的

本研究的目的是考察频谱异步听觉延迟对听力正常和听力受损听众与初始塞音相关的发声感知的影响。

研究设计

前瞻性实验研究,比较两组听众操纵EOA线索的感知差异。

研究样本

30名年龄在21至60岁之间的成年人完成了本研究:17名听力正常的听众和13名轻度至中度感音神经性听力损失的听众。

数据收集与分析

参与者在辅音 - 元音音节结构中听浊音和清音塞音。每个音节的EOA沿连续体变化,使用识别和辨别任务来确定EOA操纵是否导致发声感知的类别转变。在识别任务中,参与者将辅音识别为属于两个类别之一(浊音或清音同源词)。他们还使用相同的刺激集完成了异同辨别任务。通过检查识别任务的结果对辨别结果的预测准确性,用d'敏感性测量来确认类别感知。根据识别函数和沿EOA连续体的辨别峰值的变化来确定EOA操纵对发声感知的影响。比较两组参与者,以确定EOA对发声感知的影响作为音节和听力状态的函数。

结果

对于本研究中使用的一些音节,两组听众在操纵EOA时均表现出发声感知的类别转变。也就是说,随着语音低频和高频频段之间的时间起始异步性被操纵,听众对辅音发声的感知在浊音和清音类别之间变化。由于EOA操纵,听力正常的听众和听力损失的听众之间未发现显著差异。

结论

本研究结果表明,听力正常和听力受损的听众可能都将自然语音中发现的频谱异步延迟用作发声区分的线索。虽然现代助听器中的延迟小于本研究中使用的延迟,但可能的影响是数字信号处理或开放式适配放大方案产生的额外异步延迟可能会导致听力损失的听众错误感知发声线索。

文献AI研究员

20分钟写一篇综述,助力文献阅读效率提升50倍。

立即体验

用中文搜PubMed

大模型驱动的PubMed中文搜索引擎

马上搜索

文档翻译

学术文献翻译模型,支持多种主流文档格式。

立即体验