论文部分内容阅读
本文提出一种基于谱相关技术的语音信号检测方法。该方法以信号幅度谱的自相关函数为基础，通过峰值检测来达到自动检测语音的目的。在无线单工电台自动话音转接系统中的实验结果表明该方法能够达到满意的检测效果。与时域自相关方法以及倒谱方法相比，基于谱相关技术的检测方法不仅检测效果好，而且其对白噪声和周期性干扰的稳健性要高得多。
In this paper, we propose a speech signal detection method based on spectral correlation. Based on the autocorrelation function of the signal amplitude spectrum, the method achieves the goal of automatic voice detection through peak detection. Experimental results show that this method can achieve satisfactory detection results in the wireless simplex station automatic voice transfer system. Compared with the time-domain autocorrelation method and the cepstrum method, the detection method based on the spectral correlation technique has not only good detection effect but also much higher robustness to white noise and periodic interference.