Dept. of Information and Communication Sciences

The auditory filter of hearing impaired is wider than that of normal hearing people. Thus, the frequency selectivity decreases because of increased of masking effects. We have focused on this wider auditory filter shape and developed a method in which the critical band is compressed along the frequency axis. We also implemented a system based on the proposed algorithms using "SIMULINK." To test our method, we conducted two experiments. The first evaluated the Mean Opinion Score (MOS) and the second was an articulation test. The quality and intelligibility of speech sounds were improved in Exp.1. In Exp.2, the articulation score was also improved when frequencies were compressed. These results show the feasibility of frequency compression algorithm for the hearing impaired people.

A new method of processing speech for digital hearing aids was proposed. In this method, each critical band was compressed along the frequency axis. To implement this, we tested two algorithms based on a filter bank (FB) approach and the fast Fourier transform (FFT).

Abstract Normal hearing listeners are able to understand speech with different types of degradation, because speech has redundancy in the spectro-temporal domains. On the other hand, hearing impaired listeners have less such capability. Because of this, speech signal processing for hearing impairment needs to preserve important landmarks when enhancing a speech signal.

Abstract Reverberation is a large problem for speech communication and it is known that strong reverberation affects speech intelligibility. This is especially true for people with hearing impairments and/or elderly people. Several approaches have been proposed and discussed to improve speech intelligibility degraded by reverberation.

Abstract This study aimed to reduce ventilator noise in the speech signals of patients who use a ventilator. Two applications, spectral subtraction and adaptive filtering, were examined. For the first experiment, we employed a new technique based on spectral subtraction for ventilator noise reduction and then evaluated the result using mean opinion score. The results showed that the new technique was superior to the common approach of spectral subtraction in reducing ventilator noise.

Abstract In a large auditorium, perceiving speech may become difficult. One reason that reverberation degrades speech intelligibility is the effect of overlap-masking (Bolt and MacDonald, 1949; Nabelek and Robinette, 1978). Reverberation is a more critical issue for elderly people to perceive speech than it is for young people (Fitzgibbons and Gordon-Salant, 1999). Arai et al.

There are a number of ways to reduce the effects of reverberation and noise to improve speech intelligibility for elderly people. As an example of listening systems, Nábelek and Donahue [7] found that the word identification scores of elderly people in tests delivered through an assistive listening system were higher than corresponding scores for tests delivered through a public address system that was installed in a reverberant environment.

Abstract Elevation of the threshold of audibility occurs in hearing-impaired people, and these individuals have an expanded auditory filter (Glasberg and Moore, 1986). Threshold elevation is assumed to occur due to an increase in frequency components that pass the auditory filter; an assumption known as the “power spectrum model” of masking (Patterson and Moore, 1986).

Abstract Many individuals experience some degree of hearing loss as they age. In previous studies, Arai et al.(2001, 2002) reported that steady-state suppression of speech improves speech intelligibility in reverberant environments. Steady-state portions are defined as those having more energy, but which are less crucial for speech perception.

Abstract In an actual hall, reverberation degrades speech intelligibility, which is the result of overlap-masking occurred when segments of an acoustic signal are affected by reverberation components of previous segments. Arai et al.(2001, 2002) have been proposed a pre-processing technique which suppresses steady-state portion of speech in order to prevent the result of overlap-masking. However, the originally proposed technique is not suitable for real-time processing.

1. Introduction Pattern playback, a device that converts a spectrographic representation back to a speech signal, was developed by Cooper and his colleagues from Haskins Laboratories in the late 1940s [1] and has contributed tremendously to the rapid development of research in speech science [2–4]. By converting a spectrogram into a sound, we can test which acoustic cue projected on the spectrogram is important for speech perception.