语音增强技术在实时音频处理中的应用研究  

Research on the Application of Speech Enhancement Technology in Real-Time Audio Processing

作　　者：姚晓龙 张雪 YAO Xiaolong;ZHANG Xue(China Unicom Software Research Institute Harbin Branch,Harbin 150036,China)

机构地区：[1]中国联通软件研究院哈尔滨分院,黑龙江哈尔滨150036

出　　处：《电声技术》2024年第11期73-75,共3页Audio Engineering

摘　　要：随着现代通信技术的快速发展,语音增强技术在实时音频处理中的应用备受关注。研究一种基于深度学习的语音增强应用方案,通过WaveCRN模型对语音信号进行时频特征建模与增强。实验结果表明,该方案在多种复杂噪声环境下显著提升语音清晰度和信噪比,优于传统算法,为实时音频处理系统的构建提供有效的技术支持。With the rapid advancement of modern communication technologies,speech enhancement applications have drawn significant attention.A speech enhancement scheme based on deep learning is studied.The time-frequency characteristics of speech signals are modeled and enhanced by WaveCRN model.Experimental results demonstrate that this approach significantly improves speech clarity and signal-to-noise ratio in complex noise environments,outperforming traditional algorithms,and providing effective support for constructing real-time audio processing systems.

关 键 词：语音增强 实时音频处理 深度学习 

分 类 号：TN912[电子电信—通信与信息系统]