双耳环形扬声器阵列的虚拟听觉研究

The research explores the soundfield synthesis at ear canal by the ring loudspeaker array of small size around pinna. The effect of the audiences’ pinnas of themselves is used to help construction of the soundfield at the left and right ear canals to avoid the problems of measurement and calculation of HRTF and get the better virtual auditory effect. The simplified models of human head, rigid sphere model and snowman model, are introduced to deduce the impulse response from virtual sound source to loudspeaker array according to the spherical harmonics theory. Without consideration of the effect of pinna, the sound of central point is deduced from loudspeaker array, which is also the position of ear cannal, is studied. Considering the effect of the pinna, the transfer function from loudspeaker array to the ear cannel is calculated by the boundary element method. The subject and object test and analysis can be exploited by comparing with the HRTF data. To get the better spatial hearing, image source method (ISM) and spherical harmonics decomposition are combined to obtain the impulse response including reflected sound from source to the loudspeaker array in the room. A loudspeaker array platform around pinna is developed to verify the theoretical research and have comparison with conventional HRTF method to validate the project.

本课题探索分别环绕双耳外耳廓两个小型环形扬声器阵列在耳道处的声场合成问题，研究利用个体自身外耳廓的作用在耳道处分别重建左右耳道的声场，避免HRTF的测量或者数值计算面临的问题，从而获得更优异的虚拟听觉效果。采用简化的人头模型，如刚性球体模型或雪人模型，根据球谐函数理论推导虚拟声源到双耳扬声器位置的冲击响应，在不考虑耳廓影响时，推导阵列中心位置即耳道位置的合成声场，在考虑耳廓影响时，利用边界元的数值计算方法获得双耳扬声器到耳道处的传递函数，通过主观评测和客观分析两方面与测量所得HRTF数据进行比较研究。为获得更好的空间听觉效果，利用镜像声源法与球谐波分解合成理论结合推导得到房间环境下声源到双耳扬声器的冲击响应。利用双耳扬声器阵列验证平台对上述理论研究进行主观评测，与传统HRTF方法进行比较，验证本项目的可行性。

本课题探索虚拟听觉中的声场合成问题，研究利用个体自身外耳廓的作用在耳道处分别重建左右耳道的声场，避免HRTF的测量或者数值计算面临的问题，扩展头外声像，获得更优异的虚拟听觉效果。采用简化的人头模型，在不考虑耳廓影响时，推导阵列中心位置即耳道位置的合成声场，在考虑耳廓影响时，利用边界元的数值计算方法获得双耳扬声器到耳道处的传递函数。采用稀疏测量建模的方法实现双耳处的虚拟声场合成，提高声场的方向辨识准确性。与传统HRTF方法进行比较，验证本项目的可行性。