基于CS理论的“非盲式”带宽扩展算法研究  

Non-blind Bandwidth Extensionalgorithm Based on Compressive Sensing

作　　者：吕焱[1] 

机构地区：[1]北京工业大学实验学院,北京101101

出　　处：《计算机与数字工程》2014年第5期779-781,858,共4页Computer & Digital Engineering

摘　　要：提出一种基于压缩感知的"非盲式"带宽扩展算法,并利用"盲式"带宽扩展模块以进一步改善算法的性能。主客观测试结果表明,基于压缩感知的"非盲式"带宽扩展算法能够实现音频信号的高频重建,且在满足信号完美重建的条件下,重建音频质量要优于AMR-WB+中的带宽扩展算法。A new “non-blind” bandwidth extension algorithm based on CS is proposed and a “blind” bandwidth exten- sion method as one part is used to improve the performance of this algorithm. It proves by subjective and objective evaluation that the “non-blind” algorithm based on CS can basically restore the high-frequency information. Besides when it meets the condition of perfect reconstruction, the reconstruction quality of high frequency by this algorithm is better than that by the bandwidth extension algorithm in the AMR-WB＋.

关 键 词：压缩感知 带宽扩展 音频编码 AMR-WB+ 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]