采用双麦克风工作模式提高语音采集效果.为解决传统CIS算法在信噪比低的情况下语音识别率差的问题,在CIS算法设计前端增加了基于LMS算法的自适应滤波器.通过Matlab仿真,语音中的噪声得到很大程度上消除.为了降低运算量、减少硬件资源和功耗,通过FFT运算在频域实现带通滤波功能.在硬件实现中,与刺激芯片联合仿真,刺激幅度与刺激时间均满足要求.%Double microphone working mode to improve speech acquisition effects is put forword in this paper. In order to solve the problem of that the traditional CIS algorithm is difficult to get a good rate of speech recognition under the low signal-to-noise ratio envirment, an adaptive filter based on LMS algorithm is designed in front of CIS portion. Through Matlab simulation, the noise in the voice is largely eliminated. In addition, in order to reduce the amount of computation, hardwire resoures and power waste, band pass filter is realized by FFT operation. In hardwire, the stimulus amplitude and stimulus duration meet the requirements in the stimulation with stimulating chip.
展开▼
