基于参数心理声学模型的正弦波分析  

Sinusoid Analysis Based on Parametric Psychoacoustic Model

作　　者：周宏[1] 陈健[1] 

机构地区：[1]上海交通大学电子工程系,上海200030

出　　处：《上海交通大学学报》2002年第9期1250-1253,共4页Journal of Shanghai Jiaotong University

摘　　要：提出了基于临界频带率、依赖于强度的扩展函数、非线性迭加和前向掩蔽效应的参数心理声学模型 .将该模型与 MP算法配合 ,可以去除感知上无关的正弦波 ,控制模型噪声的分布 .提出两种对感知上重要的正弦波进行排序和选择的方法 ,它们对噪掩比的改善作用最大 .实验证明 ,将这两种方法与参数心理声学模型相结合 ,可以有效地减少所需的正弦波数量 ,大大提高正弦波跟踪的效率 .A parametric psychoacoustic model was proposed, based on critical band rate, level-dependent spreading function, nonlinear superposition and forward masking. Combined with the MP algorithm the model can wipe off perceptually irrelevant sinusoids and control the distribution of intrinsic model noise. Two methods which can order and select perceptually most important sinusoids were proposed. The experiments verify that with these new technologies, the amount of sinusoids can be greatly reduced and the tracking efficiency can be significantly improved. Consideration with respect to the reduction of amount of computation during iteration was given.

关 键 词：参数心理声学模型 噪掩比 正弦波 低码率音频编码 MP算法 语音信号处理 

分 类 号：TN912.32[电子电信—通信与信息系统]