提出一种联合主成分分析与空间傅里叶重构的方法,在对头相关传输函数(HRTF)数据进行有效压缩的同时,可以从有限的离散空间方向数据重构空间方向上连续的HRTF函数。以KEMAR人工头的HRTF数据为例计算表明.实现水平方向上连续重构所需的数据量仅为原来的20%。
A method of united principal component analysis and spatial Fourier reconstruction is proposed to analyze HRTFs(Head-Related Transfer Functions). By using the proposed method, it is able to compress HRTF data effectively, and reconstruct continuous HRTF. The results calculated from KEMAR HRTF database show that only 20% of the original data are necessary for a good reconstruction of continuous HRTF on the horizontal plane.