| 摘要 |
PURPOSE:To provide a voice processing device which performs a voice process in real time, has a high formant emphasis effect and less unnatural residual noises and to provide a hearing aid which efficiently performs a formant emphasis and signal processes that put voice signal spectra into the audible region of a hard of hearing person. CONSTITUTION:The device is provided with a filter bank 360 which divides the input frequencies into plural bands, a gain determination means 390 which performs a frequency analysis of input voices for every plural sampling frequency, provides a higher gain for a higher energy band and gives a lower gain for a lower energy band so as to correspond to each frequency band of the filter bank, a multiplication means 370 which multiplies the output of the filter bank and the gain determined by the means 390 for every sampling frequency and a total sum means 380 which totals the results of the multiplications above. An FFT computation means 310 analyzes the input signals in terms of frames to perform a formant emphasis and decides a gain for each band and the input signals, which are divided into bands by the filter bank 360, are multiplied by the gains to accomplish the formant emphasis. |
