AUDITORY ANALYSIS AND SPEECH RECOGNITION

听觉分析和语音识别

• 批准号: 3216017
• 负责人: Judy R Dubno
• 金额: $ 23.45万
• 依托单位: UNIVERSITY OF CALIFORNIA LOS ANGELES
• 依托单位国家: 美国
• 财政年份: 1981
• 资助国家: 美国
• 起止时间: 1981-12-01 至 1991-11-30
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/3216017
• 关键词: audiometry; auditory discrimination; auditory feedback; auditory stimulus; auditory threshold; cues; human subject; perceptual maskings; psychoacoustics; sensorineural hearing loss; sound frequency; speech

The proposed research is designed to investigate the relationship between frequency analysis and speech recognition by hearing- impaired individuals. Abnormal auditory processing is often assumed to be responsible for reduced speech recognition observed in these listeners, without careful consideration of the concurrent effects of threshold elevation. It is possible that a strong association between these two conditions secondary to hearing impairment has not, as yet, been established because, with current techniques, evaluation of the frequency resolution of normal and impaired auditory systems is often performed without accounting for effects of threshold sensitivity, making an accurate estimate of functional auditory capabilities difficult. Experiment 1 will estimate the auditory filter width of normal-hearing and hearing- impaired listeners while controlling for differences in threshold sensitivity. Under conditions in which the frequency analysis of speech is likely to result in modifications of its spectral characteristics, speech recognition by normal-hearing listeners does not deteriorate. However, speech recognition by hearing-impaired listeners may not be as resistant to spectral modification. Experiment 2 will determine if consonant recognition by normal- hearing listeners is equivalent to performance or hearing-impaired listeners then the stimuli presented to normal listeners has been modified to approximate consonants processed by wide auditory filters. In addition, changes in frequency analysis may explain residual deficits in speech recognition which remain once threshold elevation and/or excessive masking is taken into account, or under difficult listening conditions. Experiment 3 will evaluate if diminished spectral contrasts in speech sounds resulting from widened auditory filters differentially affects normal-hearing and hearing-impaired listeners in their ability to understand speech in noisy environments. The results of these experiments will provide a clearer understanding of the dynamics of the auditory processing dysfunction imposed by sensorineural hearing loss and the magnitude of the effects on speech recognition for most hearing-impaired individuals.

本研究旨在探讨 频率分析和语音识别之间的联系 受损的个体。 异常的听觉处理通常 被认为是负责减少语音识别观察 在这些听众中，如果不仔细考虑并发的 阈值升高的影响。 有可能是一个强大的 这两种情况之间的联系， 目前尚未确定减值，因为 技术，评估正常和 受损的听觉系统通常在没有考虑到 阈值敏感性的影响，准确估计 功能性听觉障碍 实验1将 估计正常听力和听力的听觉滤波器宽度， 在控制听阈差异的同时， 灵敏度 在语音的频率分析被 可能导致其光谱特性的改变， 听力正常的听众的语音识别不 恶化 然而，听力受损的人的语音识别 收听者可能不会抵抗频谱修改。 实验2将确定是否辅音识别正常- 听力听众相当于表演或听力受损 听众然后刺激呈现给正常听众已经 修改为近似于宽听觉处理的辅音 filters. 此外，频率分析的变化可以解释 语音识别的残余缺陷， 高度和/或过度掩蔽被考虑在内，或 听力条件差。 实验3将评估是否 语音中的频谱对比度减弱， 加宽的听觉滤波器不同地影响正常听力， 听力受损的听众在他们的能力，理解讲话 在嘈杂的环境中。 这些实验的结果将 提供了一个更清晰的了解动态的听觉 感觉神经性听力损失造成的处理功能障碍， 对大多数人来说， 听力受损的人。