数字语音降噪系统实现研究  

Research on the Implementation of Digital Speech Noise Reduction System

作　　者：宋飞 范焜 张昊宇 朱泳翔 SONG Fei;FAN Kun;ZHANG Haoyu;ZHU Yongxiang(School of Electronic Engineering,Xi′an Aeronautical Institute,Xi′an 710077,China)

机构地区：[1]西安航空学院电子工程学院,西安710077

出　　处：《西安航空学院学报》2023年第5期54-59,共6页Journal of Xi’an Aeronautical Institute

基　　金：西安航空学院大学生创新创业训练计划项目(S202211736086)。

摘　　要：在短波语音通信场景下,基于梅尔倒谱距离的语音端点检测与谱减法降噪和有效的滤波器滤波相结合研究了数字语音降噪系统设计方案。经过对实际采集的带噪语音处理分析,该方案既能够高效实现对强背景噪声的降噪处理,又能够有效抑制“音乐噪声”的影响,算法简洁,资源利用率较高,适合于在可移动的短波接收机的嵌入式芯片中实现。In the short-wave speech communication scenario,speech endpoint detection based on MEL cepstrum distance is combined with spectral subtraction and effective filter filtering to study the design of digital speech noise reduction system.Through the analysis of the actual acquisition of noisy speech processing,the scheme can not only realize the noise reduction of strong background noise,but also effectively suppress the influence of music noise.with simple algorithm and high rate of resource utilization,it is suitable for implementation in the embedded chip of the mobile short-wave receiver.

关 键 词：梅尔倒谱距离 语音端点检测 谱减法降噪 短波语音通信 

分 类 号：TN924[电子电信—通信与信息系统]