| Project/Area Number |
62550291
|
|---|
| Research Category |
Grant-in-Aid for General Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
電子機器工学
|
|---|
| Research Institution | Osaka Prefecture University |
|---|
Principal Investigator |
YONEDA Shojiro Univ. of Osaka Pref. College of Eng., Dep. of E. E. Prof., 工学部, 教授 (40081286)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
SASAKI Itsuo Univ. of Osaka Pref. College of Integrated Artsand Science, Dept. of Instrument, 総合科学部, 助手 (60094504)
|
|---|
| Project Period (FY) |
1987 – 1989
|
| Project Status |
Completed (Fiscal Year 1989)
|
| Budget Amount *help |
¥2,000,000 (Direct Cost: ¥2,000,000)
Fiscal Year 1989: ¥300,000 (Direct Cost: ¥300,000)
Fiscal Year 1988: ¥800,000 (Direct Cost: ¥800,000)
Fiscal Year 1987: ¥900,000 (Direct Cost: ¥900,000)
|
|---|
| Keywords | switched-capacitor / real-time processing / Discrete Fourier Transform / 64 sampling points / Fundamental period / Inverse Discrete Fourier Transform / Completely parallel processing / time-dependent sinusoidal output / 基本周期 / 音声エネルギ / セグメンテ-ション / 離散フーリエ変換 / 離散フーリエ逆変換 / 音声信号処理 / ピッチ周波数検出 / 母音音声認識 / 実射間処理 / 並列金スペクトル出力 |
|---|
| Research Abstract |
The processings from the first-stage to the analysis and the recognition of the speech's signal are processed by the computer-technology in general. This brings the more quantities of the memory, the calculation and the super high-speed processings to the computer. Recently, these trends will have been increasing more and more, because the unknown new factors have been consequently found by the computer technology. As a result, the complete real-time processing and the small-sized realization using computer technology may be impossible.
Then, the methods of the complete real-time and parallel processings are proposed by using the switched-capacitor Discrete Fourier Transform and its Inverse Transform (DFT-IDFT), as follows ;
1) The output of DFT is obtained as all the time-functioned higher-harmonics including the amplitude's and the phase's informations at the same time. 2) IDFT is obtained simply only by adding the above DFT's outputs. 3) The fundamental period of the unknown speech's signal are obtained also in the real-time. 4) From the speech-signal's DFT, the two-dimensional DFT have been developed also in order to approach to the basic of the futer image procession. 5) In order to use the many DFT's output in parallel and in same-time , the switched-capacitor neural network have been also proposed aiming at a real-time processing of the recognition.
|
