Abstract
The present paper deals with a numerical simulation and a hearing experiment concerning age-related deterioration of hearing in noise. In the simulation, power values of the formants at threshold were calculated for a vowel presented in noise by employing an auditory filter model. Results of the simulation indicated that the deterioration of frequency selectivity made perception of vowel formants difficult in noise, and to narrow the formant bandwidths could reduce the difficulty. In the experiment, nonsense lists of syllables, that were composed of five synthesized Japanese vowels, were presented to young and old subjects in noise. As results of the experiment, to narrow the formant bandwidths or to fluctuate the pitch frequency (viz. to differentiate the frequency among phonemes) reduced the S/N required by old subjects to discriminate the lists; each reduction was approximately 2dB in average. These findings imply that to narrow the formant bandwidths and to fluctuate the pitch frequency are effective to compensate age-related deterioration of hearing in noise.
