The relations between region of formants and their peak levels in identification of vowels with colored masking noise

This paper describes the relation between the region defined by the first and second formant frequencies and those peak levels in order to design a compression type of hearing aid. The experiments are conducted with the normal hearing subjects whose hearing areas are restricted by increasing HTL with masking noise. We present the synthetic speech signal with two resonant components to examine his/her vowel identification. In the previous experiments, we have shown a clear relation between speech listening ability and formant information existing in the hearing area. In the experiments 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 restricted with the noise.[1] After that, the new audiometer, which generated periodic single resonant signal as well as sinusoidal one, has been introduced as a software tool. This makes it possible to estimate his/her speech-listening ability from the hearing area measured by use of the speech-like signal and the compression characteristics of the hearing aid.[2] In order to extend this knowledge, we try to investigate the relations between the formant peak levels and the hearing of vowels. This information can be one of the parameters to control the compression conditions for hearing aids in listening to speech signals. The results of hearing tests conducted by normal hearing subjects controlling the second formant information will be shown in the paper.