| Project/Area Number |
04650291
|
|---|
| Research Category |
Grant-in-Aid for General Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
電子通信系統工学
|
|---|
| Research Institution | KUMAMOTO UNIVERSITY |
|---|
Principal Investigator |
WATANABE Akira KUMAMOTO UNIVERSITY, DEPARTMENT OF ELECTRICAL ENGINEERING AND COMPUTER SCIENCE, PROFESSOR, 工学部, 教授 (50040382)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
IKEDA Takashi KURUME NATIONAL COLLEGE OF TECHNOLOGY, ASSISTANT, 助手 (80222884)
UEDA Yuichi KUMAMOTO UNIVERSITY, DEPARTMENT OF ELECTRICAL ENGINEERING AND COMPUTER SCIENCE,, 工学部, 助教授 (00141961)
|
|---|
| Project Period (FY) |
1992 – 1993
|
| Project Status |
Completed (Fiscal Year 1993)
|
| Budget Amount *help |
¥2,100,000 (Direct Cost: ¥2,100,000)
Fiscal Year 1993: ¥700,000 (Direct Cost: ¥700,000)
Fiscal Year 1992: ¥1,400,000 (Direct Cost: ¥1,400,000)
|
|---|
| Keywords | telephonic speech / hard of hearing / hearing aid / single resonant analysis / formant extraction / DSP processing / 音声処理DSPボード / 補聴方式 |
|---|
| Research Abstract |
This research aims at development of a new hearing aid in which acoustic theory of speech production is applied to telephonic speech processing. We expect that this hearing aid is helpful for the sensory neural hearing impaired to listen telephonic speech. The research results are as follows :
1.System research of telephonic speech hearing aid by computer simulation
Based on the fact that telephonic signals convey speech information only in almost all cases, we have studied speech parameter estimation methods and those application to amplitude compression.
(1)Development of parameter estimation methods and single resonant analysis-synthesis method in telephonic speech
Since signal components near the formants are important to perceive speech clearly, we first developed formant estimation metod by inverse filter control for telephonic speech. Next, we achieved the analysis-synthesis method in which speech signals were separated into two single resonant components by the estimated formants a
… More
nd after multiplying compression coefficients, the two components were added again.
(2)Proposal of amplitude compression by effective levels
Instantaneous amplitude compression as an old method causes harmonic distortions for non-linear transformation. We proposed amplitude compression by the effective levels of signals. As a result of comparison between those methods by spectra and restored speech quality, it is verified that the proposed method, that is, the single resonant analysis-synthesis method with amplitude compression by the effective levels is more successful to restore distortionless speech.
2.Design and trial product of real-time hearing aid with DSPs
A general-purpose speech processing board which can be used as both a master and a slave has been designed and produced with a 32 bit floating-point DSP.The hearing aid is constructed with the three boards. The results of the real-time processing by the aid show to be quite same as them of computer simulation in the speech quality and the spectra. Less
|
