基于DSP的语音信号自适应滤波系统的设计与实现

Design and realization of the voice signal self_adaptive filtering system based on DSP

传统的自适应滤波噪声抵消法一般均采用双声道系统,其结构十分复杂,针对这个缺点,提出选取原始输入的延时信号作为参考噪声输入,设计一种利用DSP处理器对语音信号进行自适应滤波的的单声道系统。该系统利用音频芯片TLV320AIC23和TMS320VC5509 DSP共同搭建,实现了语音信号的采集、传输及噪声信号的自适应滤波处理。实验结果表明,基于DSP的语音信号自适应滤波系统能够有效地将有用信号与噪声分离,获得理想的语音播放效果。

The complex double channel system is always used in the traditional adaptive noise cancellation method. Aiming at the shortcoming, this paper proposes that the delay of original input signal is selected as the reference noise input, and then designs a single-channel system by using a DSP processor which can self-adaptively filter the voice signal. The system achieves the acquisition, transmission of voice signal and self-adaptively filters the noise signal, by seting up by TMS320VC5509 DSP and audio chip TLV320A1C23. The experimental results show that the self adaptive filtering system can effectively separate the useful signal and noise and get a perfect playback effect.