基于时域逆滤波的宽带脉冲声生成技术
Pulse Sound Generation Based on Time-Domain Inverse Filter

为了改善扬声器的频率响应,生成稳定、可重复的宽带脉冲声信号,提出了基于时域逆滤波的宽带脉冲波生成技术。首先利用最长序列(MLS)测量扬声器的单位冲激响应,然后设计基于最小二乘(LS)准则和最小均方误差(MMSE)准则下的FIR逆滤波器,最后将信号分别经过FIR逆滤波器预处理,再次通过扬声器激励,生成了理想的巴特沃斯宽带脉冲信号。实验证明2种逆滤波算法行之有效,可以用于生成声学测试需要的宽带脉冲声,为实现声学性能现场测量提供了一种解决方案。
In order to improve the frequency response of loudspeaker and generate a stable and repeatable broadband pulse sound, a new technique of pulse sound generation based on time-domain inverse filter is proposed. Firstly, Maximum Length Sequence (MLS) is utilized to measure the impulse response of the loudspeaker. Then, an inverse filter is designed with Least Squares (LS) rule and Minimum Mean Square Error (MMSE) rule respectively. Finally, by exciting the loudspeaker with pre-filtered voltage signal, the desired Butterworth pulse sound can be produced successfully. Experimental results and their analysis show preliminarily that both inverse filter methods are effective, which can be applied in pulse sound generation and provide a solution to real-time online acoustic measurement in the field