论文部分内容阅读
针对单参考输入自适应噪声抵消器在多噪声源情况下除噪声效果不理想的问题 ,根据多参考输入自适应噪声抵消器模型 ,在推导多参考输入最小均方 (ML MS)算法的基础上 ,给出其改进算法 (IML MS) .结合语音信号的特点 ,给出了多参考输入自适应语音消噪算法 (VML MS) .利用调整算法执行过程的算法改进思想 ,突破了传统的修正收敛系数的算法改进模式 .最后对几种算法的性能进行了计算机仿真比较 .
Based on the multi-reference input adaptive noise canceler model and the multi-reference input minimum mean square (ML MS) algorithm, a single reference input adaptive noise canceller is not ideal in the case of multiple noise sources. (IML MS) is proposed.According to the characteristics of speech signal, a multi-reference input adaptive speech noise reduction algorithm (VML MS) is proposed.It improves the thinking of the algorithm by using the algorithm of adjusting algorithm and breaks through the traditional correction convergence Coefficient algorithm to improve the model.Finally, the performance of several algorithms were compared by computer simulation.