数据采集与处理2012，Vol.27Issue(6)：646-651,6.
一种基于TMS320C6713的语音信号频域盲分离算法
Blind Speech Separation in Frequency Domain Based on TMS320C6713 DSK
摘要
Abstract
An implementation for speech blind source separation of frequency-domain in a real environment is introduced. Based on the TMS320C6713 DSK platform, the hardware resources are used, and the blind speech separation on the DSP platform is implemented. Meanwhile, software optimizations on the levels of compiling environment, algorithm and program are conducted. Experimental results show good performance on two-speaker mixed speech separation in a real room on offline and online scenarios.关键词
频域盲分离/非平稳信号/Ping-Pong缓存/软件优化Key words
frequency-domain blind source separation/ non-stationary signals/ Ping-Pong buffering/ software optimization分类
信息技术与安全科学引用本文复制引用
刘朝晨,吴强,殷超,李清石,许宏吉,李玉军..一种基于TMS320C6713的语音信号频域盲分离算法[J].数据采集与处理,2012,27(6):646-651,6.基金项目
国家重点基础研究发展计划("九七三"计划)(2010CB735906)资助项目 （"九七三"计划）
中国博士后科学基金(2011M501092,2012M511508)资助项目 （2011M501092,2012M511508）
山东省博士后创新项目专项资金(201103003)资助项目 （201103003）
山东大学自主创新基金(2012ZD035,2011GN062)资助项目. （2012ZD035,2011GN062）
