多通道连续头相关传输函数的测量

Measurement of Multi-Channel Continuous Head-Related Transfer Function

头相关传输函数(HRTF)是虚拟声研究的重要参数,在沉浸式语音交互中有广泛的应用.在很多应用中常常需要高密度HRTF数据,而高密度HRTF的测量是一件十分费时繁琐的工作.采用自适应滤波器和连续转动方式测量高密度HRTF是一个有效途径.本文提出了基于广义多通道频域自适应滤波器的多通道连续HRTF测量方法,该方法适用于快速多通道系统辨识.实验表明,在不同的系统噪声、环境噪声以及不同转速情况下,本文方法比传统的归一化最小均方测量方法效果更好,测量误差显著降低.

Head-related transfer function（HRTF）is an important parameter for virtual sound research,which has been widely applied in immersive voice interaction.High density HRTF data are often required in many applications,but the measurement of high density HRTF is time-consuming and tedious.The adaptive filter and con-tinuous rotation method is effective in measuring high density HRTF.This paper presents a multi-channel continuous HRTF measurement method based on generalized multi-channel frequency-domain adaptive filter,and the method is suitable for fast multi-channel system identification.Experiments show that compared with traditional normalized least mean square measurement method,the proposed method can achievebetter effects and significantly reduce measurement errors under different system and environmental noises and at different rotating speeds.