论文部分内容阅读
语音信号实时去噪净化处理是语音通信系统的关键技术。传统的小波阈值去噪方法由于阈值的单一性,在去除噪声的同时,造成语音信号损伤。针对这一问题,提出了一种基于小波变换的动态阈值法语音信号净化方法。能根据噪声和信号的变化而动态调节阈值;采用改进的阈值函数对小波系数缓变地压缩,仿真实验结果表明,能有效去除信号中的噪声和较好保留语音细节。
The real-time denoising of voice signal is the key technology of voice communication system. Because of the unity of the threshold, the traditional wavelet threshold denoising method can remove the noise and cause the speech signal to be damaged. In response to this problem, a dynamic threshold method of speech signal purification based on wavelet transform is proposed. According to the change of noise and signal, the threshold can be dynamically adjusted. The wavelet coefficient is slowly compressed by the improved threshold function. The simulation results show that it can effectively remove the noise in the signal and retain the detail of speech.