Acoustic and perceptual effects of WDRC amplification

描述(由申请人提供)：WDRC处理使用快速、自动的增益调整，将更多的语音提示放在听者的可听范围内。大多数助听器(80%)现在支持WDRC，因此随着时间的推移，这项技术将提供给美国620万助听器佩戴者中的大多数。然而，WDRC处理并不总是比传统放大更好地改善语音识别，这可能是因为语音信号的时间特性的失真。我们对严重丧失听力的听众特别感兴趣，由于更广泛的听觉过滤器，他们可能更依赖于时间线索进行识别。严重受损的听众在没有可接受的放大效果的情况下无法进行交流，因此重要的是不要扭曲有用的信息。因此，这个项目的长期目标是描述宽动态范围压缩(WDRC)放大对语音识别时间线索的影响，并评估这些变化对严重感音神经性听力损失的听者的影响。 第一个具体目标是比较WDRC和线性放大在一组严重感音神经性听力损失的听众中的语音识别分数。受试者将戴上主助听器，使用实验室条件下测量的可接受的临床参数和语音识别。我们还将使用特征分析来分析辅音错误模式。我们的假设是，与WDRC放大匹配严重丧失的听众将比轻度至中度丧失的听众表现得更差，错误模式将表明依赖于时间线索的声音混淆。第二个具体目标是描述WDRC对语音识别所必需的时间线索的影响。将对处理后的语音进行识别测量，以限制频谱信息。放大语音的声学分析将与语音识别分数结合使用，以直接检查WDRC的时间线索失真程度，并确定可接受的幅度压缩范围。拟议的项目还将测试量化幅度包络变化的实用性，以及这些措施对预测WDRC放大的语音识别的有效性。第三个具体目标是描述光谱线索知觉是如何随着压缩通道数量的增加而变化的。我们将结合元音识别使用声学分析。我们的假设是，具有严重损失的听众对频谱细节的辨别能力较差，在较大数量的压缩通道中表现出较少的退化。

DESCRIPTION (provided by applicant): WDRC processing uses rapid, automatic gain adjustment to place more speech cues within the listener's audible range. Most hearing aids (> 80%) now offer WDRC, so over time this technology will be provided to most of the 6.2 million hearing aid wearers in the United States. However, WDRC processing does not always improve speech recognition over conventional amplification, perhaps because of distortions to the temporal properties of the speech signal. We are particularly interested in listeners with severe loss, who because of broader auditory filters may depend more heavily on temporal cues for recognition. Severely impaired listeners cannot communicate without acceptable amplification, so it is important not to distort usable information. Therefore, the long-term objectives of this project are to describe the effects of wide-dynamic range compression (WDRC) amplification on temporal cues for speech recognition and to assess the consequences of those changes for listeners with severe sensorineural hearing loss. The first Specific Aim is to compare speech recognition scores for WDRC versus linear amplification in a cohort of listeners with severe sensorineural hearing loss. Subjects will be fit with a master hearing aid using accepted clinical parameters and speech recognition measured under laboratory conditions. We will also analyze consonant error patterns, using a feature analysis. Our hypothesis is that listeners with severe loss fit with WDRC amplification will perform more poorly than listeners with mild-to-moderate loss, and that error patterns will indicate confusion of sounds that depend on temporal cues. The second Specific Aim is to describe the effects of WDRC on temporal cues essential for speech recognition. Recognition will be measured for speech processed to limit spectral information. Acoustic analysis of amplified speech will be used in conjunction with speech recognition scores to directly examine the extent of temporal cue distortion from WDRC and to determine an acceptable range of amplitude compression. The proposed project will also test the utility of quantifying amplitude envelope changes and the usefulness of these measures to predict recognition of WDRC-amplified speech. The third Specific Aim is to describe how spectral cue perception changes with increasing numbers of compression channels. We will use acoustic analysis in combination with vowel recognition. Our hypothesis is that listeners with severe loss who are less able to discriminate spectral details will show less degradation from larger numbers of compression channels.