【作者】 王嵩； 鲍长春；

【Author】 WANG Song~(1,2) BAO Chang-chun~1 (1.School of Electronic Information and Control Engineering,Beijing University of Technology,Beijing 100124,China; 2.School of Information Engineering,Beijing Institute of Petro-Chemical Technology,Beijing 102617,China)

【机构】 北京工业大学电子信息与控制工程学院； 北京石油化工学院信息工程学院；

【摘要】 本文提出了一种由恒包络正弦和指数衰减包络正弦组成的音频信号建模方法,它可分别对信号的稳态和暂态分量建模。暂态分量的起始时刻由延迟参数确定,减少了合成信号的预回声。模型参数基于迭代分析合成法进行估计,通过采用递归滤波器组、折半搜索等快速计算方法,减少了计算复杂度。音频信号建模结果表明,该方法能有效表征原始信号。更多还原

【Abstract】 This paper presents a new audio signal modeling,which is made by constant-envelope sinusoids and exponentially damped sinusoids.These two sinusoids are used to model stationary components and transient components of signals,respectively. The onset times of transient components are characterized by a delay parameter,which mitigate the pre-echo artifact in synthesized signal.An efficient method based on iterative analysis/synthesis for estimating model parameters is proposed.By using fast algorithms including recursive filter banks and bisection search,the computational cost is reduced.The experimental results for modeling audio signals indicate that the original signal is represented very well with this method.更多还原