基于DSP的语音降噪实时实现

Real Time Realization of Noise Reduction for Speech Enhancement Based on DSP

语音教室中使用的各种语言学习系统,语音音质是衡量系统好坏的重要指标,提高语音音质的办法一方面要在硬件上下功夫,另一方面必须考虑在软件中进行降噪处理。文中根据对DNR(动态降噪)系统原理的分析,在电路降噪系统原理方框图的基础上,给出软件流程,总结了在DSP(数字信号处理器)中具体实时实现的算法,并在基于DSP的数字化语言学习系统实际应用中得到检验。实践证明,该降噪算法在语言学习系统中效果显著,而且有着实现方便、算法简单的优点。

Through analyzing the basic principle of dynamic noise reduction （DNR）, the algorithm implementation in DSP （digital signal processor） is presented in this paper. And the application to speech enhancement in the digital language learning system based on DSP is also introduced. This DNR arithmetic has a prominent effect on the learning system and has the advantage of easy realization.