STUDIES OF VOWEL PERCEPTION

元音感知的研究

• 批准号: 3405116
• 负责人: WINIFRED STRANGE
• 金额: $ 12.31万
• 依托单位: UNIVERSITY OF SOUTH FLORIDA
• 依托单位国家: 美国
• 财政年份: 1985
• 资助国家: 美国
• 起止时间: 1985-09-01 至 1988-08-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/3405116
• 关键词: auditory discrimination; human subject; perception; speech

This proposal has as its goal a more complete specification of the information that serves to identify vowels in spoken syllables of American English (AE). Recent research has given strong support to the view that listeners use dynamic (time-varying) information available throughout the entire syllable to identify vowels. This research will investigate the locus and the interrelations of perceptually informative portions of the syllable, using the technique of electronic modification of natural speech stimuli. Modified stimuli are generated by altering, deleting and rearranging portions of the digitized waveforms of spoken syllables. The perceptual relevance of different combinations of these sources of information may be inferred from the perceptual error patterns in identification of the intended vowels by untrained listeners. Modifications include stimuli in which: 1) portions of syllables originally containing different initial and final consonants are recombined, 2) portions of syllables originally containing different vowels are recombined, and, finally, 3) portions of syllables produced by different speakers are recombined. In conjunction with these perceptual studies intensive acoustic analysis of the original stimulus materials will be pursued. By comparing the acoustic configurations of original syllables and electronically modified syllables which do and do not preserve vowel identity, the relational parameters that support vowel perception can be derived.

该提案的目标是更完整地规范 用于识别美国口语音节中元音的信息 英语（AE）。 最近的研究有力地支持了这样一种观点， 监听器使用整个系统中可用的动态（时变）信息， 整个音节来识别元音。 这项研究将调查 轨迹和相互关系的感知信息部分的 音节，使用自然语音的电子修改技术 刺激。 修改的刺激是通过改变、删除和 重新排列语音音节的数字化波形的部分。 的 这些来源的不同组合的知觉相关性 信息可以从感知错误模式推断， 未受过训练的听众对预期元音的识别。 修饰包括刺激，其中：1）音节的部分 最初包含不同的首辅音和尾辅音， 重组，2）原来含有不同元音的音节部分 重新组合，最后，3）由 不同的扬声器被重新组合。 结合这些感知研究，对 将继续使用原有的刺激材料。 通过比较 原始音节和电子修饰音节的结构 保留和不保留元音身份的关系参数 可以导出支持元音感知。