遗传算法用于HRTFs共用声学极点的极零点模型逼近

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。本文通过应用新型的面向多目标参数优化的遗传算法(Genetic Algorithm:GA)进行了 HRTFs共用声学极点的极零点模型(Common-Acoustical—Pole and Zero:CAPZ)逼近。实验结果表明,GA较改进的Prony设计方法获得了更优的效果。

CAPZ model approximation of HRTF using genetic algorithm

During the study of spatial hearing and synthesis of virtual auditory space, it is requisite to consider how to properly model the acoustical characteristics corresponding to certain positions. As a popular model, CAPZ has practical and theoretical application in this field. Here, in the light of hereditary characters of GA, associating with the similarity of neighboring positions, we applied the new kind of stochastic global search method that mimics the metaphor of natural biological evolution, and managed to fit CAPZ model of HRTFs. Results show that GA is better than the traditional Prony methods.