基于小波变换的列车广播声压自适应调整算法

随着社会交通水平的日益提高，对列车广播质量与稳定性要求也越来越高。列车噪声包括列车的振动噪声与乘客的语音噪声，由于乘客语音噪声特性与广播本身类似，因此在该噪声背景下采用传统信号处理方法进行去噪遇到很大困难。采用改进的自适应小波滤波器，对列车广播时的实际环境所采集的信号进行小波分解与重构，然后用一个以人耳主观舒适度为基础的SVM声功率参考模型，再通过前一时刻的广播、噪声声压级来修正当前时刻的广播声压级，在列车环境下达到了良好的语音去噪效果。

Self - adaptive Algorithm Based on Wavelet transform for Sound Pressure Level of the Broadcasting in Train

With the development of transporting, the requirement for the sound quality of the broadcasting in train be- comes more and more sophisticated. It is difficult to derive the voice of broadcasting from the background noise, due to the noise combined with the floor mechanical noise from the train and the voice from the passengers, of which acoustical charac- teristics are similar to the one from the broadcasting. Thus, a self - adaptive algorithm based on Wavelet transform according to the human auditory characteristics, for sound pressure level of the train broadcasting is introduced in this paper. By means of the wavelet transform and time - frequency analysis for the background noise and a SVM model from the subjective aural assessments, the current dynamical, self - adaptive and optimal sound pressure level for the broadcasting are depended on the last adjacent background noise sample. According to the computer simulation and a rough implement on the train, this algorithm results in a good control for the SPL, compared to the other common methods, but a practical time - fast version of the algorithm is still needed a further study.