基于单稳随机共振的语音增强处理

Enhancement of Voice Signal Based on Monostable Stochastic Resonance

针对双稳系统结构参数不利于调节的问题,该文提出了一种基于单稳态随机共振系统进行语音增强的方法。该方法通过调节只有一个系统结构参数的单稳态随机共振系统的结构参数,即可使微弱信号与噪声在非线性系统中达到匹配的状态,从而噪声的能量向信号转移,实现信号增强的目的。通过实验仿真发现,利用本方法输出信噪比提高了4.4d B,因此该方法对语音信号的增强有一定的效果,且系统结构参数更易于调节。

Due to the bistable stochastic resonance system has two structural parameter which is not easy to be adjusted, a method based monostable stochastic resonance which has only one structural parameter to process voice signal has been proposed. In or- der to achieve the optimal output and enhance the voice signal we just need to monitor the only structural parameter. The simula- tion indicates that the voice signal which submerged in noise has been boosted and the rate of Signal to Noise （SNR） also has been raised 4.4dB.