文章摘要
陈砚圃,张长青,谭薇,杨森斌.基于听觉感知的语音稀疏表示及压缩感知*[J].,2014,33(1):53-59
基于听觉感知的语音稀疏表示及压缩感知*
Sparse representation and compressed sensing for speech signal based on auditory perception
投稿时间:2013-04-03  修订日期:2014-01-02
中文摘要:
      本文针对语音信号稀疏表示及压缩感知问题,将听觉感知引入稀疏系数筛选过程,用掩蔽阈值筛选重要系数,以得到更符合听觉感受的语音稀疏表示。通过对一帧浊音信号分别采用掩蔽阈值和能量阈值方法进行系数筛选对比实验,结果表明掩蔽阈值法具有更好的稀疏表示效果。为验证听觉感知对语音压缩感知性能的影响,与能量阈值法对照对测试语音进行压缩感知观测和重构,通过压缩比、信噪比、主观平均意见分等主客观指标评价其性能,结果表明,掩蔽阈值法可有效地提高压缩比且保证重构语音具有较高的主观听觉质量。
英文摘要:
      This paper concerns the sparse representation and compressed sensing for speech signal, the auditory perception is brought into the selection of sparse coefficients to obtain a sparse representation which is more suitable to our ears. An experiment of sparse coefficients’ selection under masking thresholds method was done comparing with the energy threshold method, the result showed the masking thresholds method is better. In order to test the impact of the auditory perception model to compressed sensing for speech signal, the testing speeches were chosen to compress under compressed sensing framework by masking threshold method and energy threshold method, through the subjective and objective indicators, such as compression ratio, signal noise ratio and mean opinion score, a conclusion was made: the masking threshold method can lead a increasing of compression ratio while the quality of hearing for reconstructed signal is not decreased.
DOI:10.11684/j.issn.1000-310X.2014.01.008
中文关键词: 语音信号,稀疏表示,听觉感知,压缩感知
英文关键词: speech  signal,sparse  representation,auditory  perception,compressed  sensing
基金项目:(61072125)
作者单位E-mail
陈砚圃 西安通信学院 dzxxlab@163.com 
张长青 西安通信学院 zhangcq1108@163.com 
谭薇 西安通信学院  
杨森斌 西安通信学院  
摘要点击次数: 2391
全文下载次数: 1170
查看全文   查看/发表评论  下载PDF阅读器
关闭