一种有利于故障扬声器时频特征提取的预处理方法

A pretreating method available to time-frequency characteristics extraction of fault loudspeaker

采用时域预加重处理方法对扬声器响应信号进行处理,并对其进行短时傅里叶变换(STFT),从变换后的扬声器响应信号时频图中可以得到更为明显的扬声器故障特征,为故障扬声器的时频特征提取提供了一种有效的预处理方法,实验结果显示故障识别率达98%以上。

The loudspeaker response signal is pretreated by the pre emphasis in time domain,and short-time Fourier transformed （STFT）. Time-frequency characteristic of fault loudspeaker is obtained and more evident than before pretreatment, which provides a more effective method to recognize and identify the fault loudspeaker. Experimental results show that： the fault recognition rate can reach 98 % or more.