Relationship Between Psychoacoustic Thresholds and Speech Recognition for Aided L

心理声学阈值与辅助学习语音识别之间的关系

• 批准号: 7915965
• 负责人: Marc A. Brennan
• 金额: $ 4.14万
• 依托单位: UNIVERSITY OF WASHINGTON
• 依托单位国家: 美国
• 财政年份: 2010
• 资助国家: 美国
• 起止时间: 2010-01-01 至 2011-12-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/7915965
• 关键词: Acoustics; Affect; Clinical; Cueing for speech; Cues; Detection; Goals; Guidelines; Hearing; Hearing Aids; Individual; Lead; Measures; Noise; Outcome; Protocols documentation; Psychoacoustics; Signal Transduction; Speech; Testing; Variant; base; computerized data processing; evidence base; hearing impairment; improved; interest; public health relevance; speech recognition; vocal cord

DESCRIPTION (provided by applicant): Compression hearing aids are the most common remedy for individuals with hearing loss. User benefit depends upon the ability of compression hearing aids to provide the acoustic cues necessary for speech recognition. The detection of amplitude modulations (variations in intensity) and gaps (silent intervals) are essential for speech recognition. Compared to linear amplification, compression might degrade amplitude modulation detection thresholds but improve gap detection thresholds. The recognition of consonant place cues depends upon the ability to detect amplitude modulations in speech whereas the ability to detect consonant manner cues depends upon the ability to detect gaps in speech. In order to improve the outcomes of hearing aid users, it is critical that we determine how compression compared to linear amplification affects amplitude modulation detection thresholds, gap detection thresholds, and speech recognition. Understanding those relationships can lead to improved hearing aid fitting protocols, improved signal processing for hearing aids, and improved evidence based clinical guidelines. Surprisingly, the ability to detect gaps or amplitude modulations has not been evaluated for fitting hearing aids. The first aim is to measure the ability of hearing aid users to detect amplitude modulations and gaps while listening through a hearing aid. The subjects will be tested under two different hearing aid conditions. The first condition is linear amplification, which should not distort modulation or gap cues. The second condition is compression, which is expected to reduce modulation cues but improve gap cues. Additionally, the project will evaluate whether a broadband noise or narrowband noise carrier better predicts speech recognition. Different carrier signals, representing the carrier (vocal folds) and the modulator (vocal tract) in speech, will be used. The second aim is to compare psychoacoustic thresholds with speech recognition. Speech recognition will be measured with voiceless consonants because subjects with hearing loss have difficulty recognizing voiceless consonants. In order to focus on the temporal cues of interest, the speech signal will be vocoded to minimize spectral cues. Differences in amplitude modulation, gap detection thresholds, consonant place identification, and consonant manner identification between compression and linear amplification will be calculated for each subject. The differences in amplitude modulation thresholds will be compared to differences in consonant place identification. The differences in gap detection thresholds will be compared to differences in consonant manner identification. The goal is to lay the groundwork for developing hearing aid fitting protocols based on modulation and gap detection thresholds. PUBLIC HEALTH RELEVANCE: Hearing aid users continue to recognize less speech than normal hearing listeners and this might be because it is unclear how hearing aids alter important speech cues. This study will measure the ability of hearing aid users to detect acoustic cues essential for speech recognition and then relate those results to speech recognition. This may lead to improved speech recognition for hearing aid users.

描述（由申请人提供）：压缩式助听器是听力损失人士最常见的补救措施。用户的利益取决于压缩助听器提供语音识别所需的声音提示的能力。幅度调制（强度变化）和间隙（无声间隔）的检测对于语音识别至关重要。与线性放大相比，压缩可能会降低幅度调制检测阈值，但会提高间隙检测阈值。辅音位置线索的识别取决于检测语音中的幅度调制的能力，而检测辅音方式线索的能力取决于检测语音中的间隙的能力。为了改善助听器用户的结果，我们必须确定压缩与线性放大相比如何影响调幅检测阈值、间隙检测阈值和语音识别。了解这些关系可以改进助听器验配方案、改进助听器的信号处理以及改进基于证据的临床指南。令人惊讶的是，检测间隙或幅度调制的能力尚未针对验配助听器进行评估。第一个目标是测量助听器用户在通过助听器收听时检测幅度调制和间隙的能力。受试者将在两种不同的助听器条件下进行测试。第一个条件是线性放大，它不应扭曲调制或间隙线索。第二个条件是压缩，预计会减少调制线索但改善间隙线索。此外，该项目还将评估宽带噪声或窄带噪声载波是否能更好地预测语音识别。将使用不同的载波信号，代表语音中的载体（声带）和调制器（声道）。第二个目标是将心理声学阈值与语音识别进行比较。语音识别将用清辅音来测量，因为听力损失的受试者很难识别清辅音。为了关注感兴趣的时间线索，将对语音信号进行声码处理以最小化频谱线索。将为每个受试者计算压缩和线性放大之间的幅度调制、间隙检测阈值、辅音位置识别和辅音方式识别的差异。幅度调制阈值的差异将与辅音位置识别的差异进行比较。间隙检测阈值的差异将与辅音方式识别的差异进行比较。目标是为开发基于调制和间隙检测阈值的助听器验配协议奠定基础。 公共卫生相关性：助听器用户识别的语音仍然比正常听力听众少，这可能是因为尚不清楚助听器如何改变重要的语音提示。这项研究将测量助听器用户检测语音识别所必需的声学线索的能力，然后将这些结果与语音识别相关联。这可能会改善助听器用户的语音识别能力。