MP3音频压缩对声纹识别的影响分析  

Analysis of the Influence of MP3 Audio Compression on Voiceprint Recognition

作　　者：袁晓波 Yuan Xiaobo(National Radio and Television Administrator 203 Station,Inner Mongolia 010000,China)

机构地区：[1]国家广播电视总局二〇三台,内蒙古010000

出　　处：《广播与电视技术》2021年第10期127-131,共5页Radio & TV Broadcast Engineering

摘　　要：传统声纹识别需要采用无损音频数据,对大数据广播监测带来较大存储压力。本文提出了采用MP3替代无损数据进行声纹识别的方法,采用多相位音频匹配算法提升了声纹识别能力,将压缩对声纹识别的影响显著减小,并且用实际信号分析了MP3音频压缩对声纹识别的影响。Traditional voiceprint recognition needs lossless audio data, which brings great storage pressure to big data broadcast monitoring. In this paper, MP3 is used to replace lossless data for voiceprint recognition. Polyphase audio matching algorithm is used to improve the ability of voiceprint recognition, which reduces the impact of compression on voiceprint recognition significantly. The paper analyzes the impact of MP3 audio compression on voiceprint recognition with actual signals.

关 键 词：广播监测 声纹识别 多相位匹配 音频压缩 

分 类 号：G202[文化科学—传播学]