调频广播中噪声和语音的检测方法

Method for Detecting Noise and Speech in FM Broadcasting

调频广播中,黑广播出现前会出现一段似高斯噪声,一些人的声音和这些噪声极为相似,若使用傅里叶变换等频域方法很难将这些噪声和语音区分开来。针对上述问题,提出了基于相关函数最大值的方差和基于短时能零比的方差两种方法来区分调频广播中的噪声和语音,通过计算每秒内语音信号的相关函数的最大值和短时能零比的方差,设定合适的阈值。实验结果表明,提出的方法能够有效地区分调频广播中的噪声和语音,同时还有利于提升广播或其他声音处理中声音的清晰度以及监测黑广播的出现。

In FM broadcasting,a period of Gaussian noise will appear before the appearance of black broadcasting.Some people’s voices are very similar to these noises.It is difficult to distinguish these noises from speech using frequency domain methods such as Fourier transform.In response to these problems,this paper proposes two methods to distinguish noises from speech in FM broadcasting based on the variance of the maximum value of the correlation function and the variance based on the short-term energy zero ratio.By calculating the maximum value of the correlation function of the speech signal per second and the variance of the short-term energy zero ratio,an appropriate threshold is set.The experimental results show that the method proposed in this paper can effectively distinguish between noise and speech in FM broadcasting,and also help to improve the clarity of the sound in broadcasting or other sound processing and monitor the emergence of black broadcasting.