Atsunobu Murase

Determination of filtering parameters for dichotic‐listening binaural hearing aids

Sensorineural hearing‐impaired people have difficulty in hearing sounds not only because of the increased hearing threshold, but also because of their reduced dynamic range of hearing (loudness recruitment), as well as the large and extensive masking resulting from the reduced frequency selectivity, especially masking by which middle‐frequency and high‐frequency components are masked by the intense low‐frequency component, the so‐called upward spread of masking. Dichotic listening, listening to complementary filtered speech signals given to the two ears, has been proposed to cope with these problems. We introduce a dichotic listening technique for binaural hearing aids and the relationship between the effect of this technique and auditory characteristics of hearing‐impaired people. The results of listening tests imply that some relationship exists between the width of the auditory filter and the effect of this technique. Moreover, intelligibility test results obtained using the low‐frequency‐boosted sound...

Effects of dichotic listening for hearing impaired people

Methods applying the dichotic listening algorithm have been proposed and examined for application in hearing aids. Murase et al. (2005) proposed an algorithm in which dichotic listening was simply implemented by a set of complementary high‐pass and low‐pass filters. One filter is for either the right or left channel; the other filter is used for the other channel. This study examined two unsolved problems with this algorithm. One is the reason why this algorithm is effective. The other is probable degradation of sound localization that occurs because interaural level differences and interaural time differences is only slightly available with this signal processing. This problem is considerable when the sounds come from frontal incidence. Regarding the first problem, intelligibility tests were performed using the low‐frequency‐boosted sounds. The intelligibility scores under the dichotic listening conditions should be higher than those under the diotic listening conditions if the effectiveness of this algo...

Effect of two‐band dichotic listening for hearing impaired listeners

The increase of the upward spread of masking is a phenomenon that is typically observed among sensorineural hearing‐impaired listeners. To resolve this problem, dichotic listening, by which an input speech spectrum is split into two complementary parts and is presented dichotically, seems effective to reduce masking between contiguous frequency bands. This study examines effects of simple two‐band dichotic listening with a cut‐off frequency around and between the typical first and second formant frequencies of the preceding vowel. We measured speech intelligibilities in both quiet and noisy environments (S/N 4 and 0 dB). Three types of vowel‐consonant‐vowel nonsense monosyllables, of which preceding vowels were /a/, /i/, and /u/, were used as speech stimuli. Results showed that this dichotic processing was effective, especially in relatively high S/N conditions. Moreover, the best dividing frequency was dependent on the preceding vowel. When /a/‐consonant‐vowel was used, the best dividing frequency was 1....