The development of the audiometer to estimate the listening ability of speech formants.

点击翻译按钮获取中文摘要

In this research, we propose a new type of audiometer, which estimates speech listening ability of the hearing impaired. In previous experiments, we have shown a clear relation between speech listening ability and formant information existing in the hearing area. When we are using the single-resonant-analysis-type hearing aid, this relation was observed not only in sensory-neural hearing-impaired persons but also in normal hearing people whose hearing areas were made smaller by increasing HTL with colored masking noise. Based on these results, we consider that single resonant signals may be more suitable for measuring hearing profiles than sine waves. The new audiometer, which generates single resonant waves as well as sine waves, has been realized as a software tool for a personal computer. Using one of the tools in the audiometer, we can immediately estimate speech-listening ability from the compression characteristic of the hearing aid and the measured hearing area that is sandwiched between UCL and HTL.We examine the audiograms of normal hearing persons and hearing impaired persons. The results of the measurement with single resonant signals showed some difference from those with sine waves in some cases. The subjects seem to notice not the information at the resonant frequency, but another one at different frequency. Next, we conducted listening tests for vowels with controlling first and second formant peak levels of the samples. The results are examined by their vowel categories and category similarities. They showed a possibility to improve the vowel identification when we consider compression coefficients based on F1, F2 information for amplitude compression type hearing aids.