符合普通话语境的宽动态范围压缩算法低频释放时间研究

发布时间：2018-08-01 15:54

【摘要】：目的:探讨在普通话语境下WDRC(宽动态范围压缩算法)低频释放时间对助听器佩戴者言语识别率的影响。方法:将低频释放时间设置为50、100、200、400、800ms五个组,分别在3种掩蔽噪声下进行实验,共形成15组实验条件。以听力障碍者作为实验对象,以普通话言语仿真声作为测试音,通过仿真平台输出,在特定的噪声条件下测试言语识别率。结果:WDRC低频释放时间设置为400ms时,听障者言语识别率最佳;Babble Noise(多说话人噪声)对听力障碍者的噪声下言语识别率有显著影响。结论:在普通话语境下低频区间时间常数与噪声类型都会影响听力障碍者的普通话言语识别率。
[Abstract]:Aim: to investigate the effect of low frequency release time of WDRC (wide dynamic range compression algorithm) on speech recognition rate of hearing aid wearer in Mandarin context. Methods: the low frequency release time was set to five groups of 50100200400800ms, and the experiment was carried out under three kinds of masking noise, and 15 groups of experimental conditions were formed. The speech recognition rate was measured by using the speech simulation sound of Putonghua as the test sound and the speech recognition rate under the specific noise condition through the output of the simulation platform by taking the hearing impaired as the experimental object and the Putonghua speech simulation sound as the test sound. Results when the low frequency release time of WDRC was set to 400ms, the best speech recognition rate of hearing impaired persons was Babble Noise (multi-speaker noise). Conclusion: in the context of Putonghua, the low frequency interval time constant and the type of noise will affect the speech recognition rate of hearing impaired people.
【作者单位】： 上海市徐汇区中心医院耳鼻咽喉科复旦大学附属中山医院徐汇医院;中国科学院声学研究所东海研究站;
【基金】：上海市卫生和计划生育委员会青年科研项目(No:2013-154)
【分类号】：R764.43
，

本文编号：2158077