【作者】 季青； 张杰； 王生九； 吴镇扬；

【Author】 JI Qing,ZHANG Jie,WANG Sheng-jiu,WU Zhen-yang 1.Jiangsu Institute of Metrology and Science,Nanjing 210007,China 2.Department of Radio Engineering,Southeast University,Nanjing 210096,China

【机构】 江苏省计量科学研究院； 东南大学无线电工程系；

【摘要】 与头相关传递函数(Head-related Transfer Functions:HRTFs)的准确、有效建模对于空间听觉的分析研究以及虚拟听觉空间的生成起着关键的作用。本文提出将HRTF分解成与方位相关的DTF(Directional Transfer Function)和与方位无关的CTF(Common Transfer Function)两部分,并应用有限长单位冲激响应(Finite-Impulse-Response:FIR)来对其分别进行逼近建模。与其他的建模方法相比较,该模型有利于HRTF中与方位相关的特征的分析处理;并且在实际应用中,只需要较少的模型参数即可描述对应于特定声源方位的传输特征,减少了存储需求,更易于3D音频效果的实时实现。更多还原

【Abstract】 The efficient model of the head-related transfer function has an important role in the research of spatial hearing and the realization of the VAS(Virtual Auditory Space) system.This paper proposes a model in which the HRTF is divided into a cascade of a common transfer function(CTF) and a directional transfer function(DTF).They are respectively approximated by finite-impulse-response filters.In contrast to other modeling methods,this model has the advantage to analyze and process the direction-dependent features of HRTF. Besides,fewer parameters of this model are required to characterize the variations of the sound source location in applications,so fewer storage demands are required and it is easier in real time to implement the 3D sound effects.更多还原