| Project/Area Number |
05650358
|
|---|
| Research Category |
Grant-in-Aid for General Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
情報通信工学
|
|---|
| Research Institution | Kumamoto University |
|---|
Principal Investigator |
UEDA Yuichi Kumamoto Univ., Faculty of Eng., Associate Prof., 工学部, 助教授 (00141961)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
IKEDA Takashi Kurume national college of tech., Res.associate, 助手 (80222884)
WATANABE Akira Kumamoto Univ., Faculty of Eng., Professor, 工学部, 教授 (50040382)
|
|---|
| Project Period (FY) |
1993 – 1994
|
| Project Status |
Completed (Fiscal Year 1994)
|
| Budget Amount *help |
¥1,900,000 (Direct Cost: ¥1,900,000)
Fiscal Year 1994: ¥1,000,000 (Direct Cost: ¥1,000,000)
Fiscal Year 1993: ¥900,000 (Direct Cost: ¥900,000)
|
|---|
| Keywords | Hearing impairement / Digital hearing aid / Tactile display method / Speech perception / Vowel normalization / Digital signal processor / Amplitude compression / Sensorineural hearing impaired / 補聴器 / ディジタル信号処理 / 聴覚障害者 / 感覚代行 / 音声信号処理 / 振動触覚 / 視覚パタン / 音声情報 |
|---|
| Research Abstract |
Brief summary is as follows.
Hearing aid : We proposed a composite compression filtering method which implemented the spectral compression in the waveform processing. The syllabic preference tests and the evaluation tests on suond quatilty were carried out by the sensorineural hearing impaired, where a digital hearing aid to realize the method in real-time has been used. As result, the favorable results were obtained in comparison with the conventional multiband compression.
On the speech feature extractions : A spectral normalization, its vowel vector representation and a speech features extraction from spectral figures were proposed in order to trasmit the phonemic information via non-auditory sensory systems. The effects of its normalization was confirmed by the identification test of vowels uttered by many speakers, where a vibro-tactile display has been used. The speech features were used in the visual discrimination test and then its effects became clear.
Prosody and its electrotactile image : Te identification tests on the accent types of words were carried out using an electrotactlile representation. It showed that it was possible to transmit the prosodic image similar to auditory one by taking account of both a conversion the voice pitch into the electrical pulse train and an addition of the speech intensity.
A speech feature extraction system : A real-time speech feature extraction system was developed with dual DSP chips and the above mentioned various processings were realized.
On this study, we have investigated the speech perception using an individual system, that is, a hearing aid, a tactile device and so on. The effects in the case using those systems simultaneously could not be become clear. In future, we will evaluate such effects systematically, based on the experimental results, the proccesing methods and the systems obtained in this study.
|
