摘要
在对自适应语音抗噪技术实时实现的可行性分析的基础上 ,提出了实时处理所需的基于TMS32 0 C32 (DSP)的硬件平台 ,采用两块 TL C32 0 44 CN(A/ D,D/ A转换器 )做成了与 TMS32 0 C32的接口板 ,实现了两路无相差信号的采集 ,形成了语音处理工作站样机。通过对几种自适应算法的分析比较 ,最后选择恰当的算法和滤波器结构实现了语音抗噪的实时处理 ,对信噪比的改善达 12 d B.
First,the feasibility of the real time adaptive noise cancellation in speech system is first described. Based on TMS320C32(DSP) and the interface between two pieces of TLC32044CN(A/D and D/A converter) and TMS320C32, a work of framework for the speech processing station is provided. Some practical adaptive noise cancellation algorithms are used in this station. In realistic noise environment, the improvement of SNR can reach more than 12dB without degrading the quality of speech.
出处
《桂林电子工业学院学报》
2001年第1期66-69,共4页
Journal of Guilin Institute of Electronic Technology
基金
电科院"九五"军事预研项目! (编号 :DY95 73 10 5 0 90 3 )
关键词
语音信号处理
语音抗噪技术
自适应抗噪
adaptive noise cancellation(ANC), real time implementation, adaptive algorithm
