Fast and Accurate High-density Hrtf Mea- Surements Using Generalized Frequency- Domain Adaptive Filter

High-Density (HD) Head-related transfer function (HRTF) measurements are extremely timeconsuming and complicated. This problem can be solved by continuously rotate a dummy head and using on-line de-convolution of binaural signals to obtain the HD HRTFs. In this work, we use a generalized frequency-domain adaptive filter (GFDAF) for HD HRTF measurements, which has significant better convergence performance than the normalized least mean square (NLMS) method, so as to admit faster rotation speed. The performance is validated based on a well-established HD HRTF database. We simulate the binaural signals using time-varying convolution, and both microphone noise and measurement noise are considered. The HRTF errors are evaluated under different noise levels and different rotation speeds. The results show that the proposed method can always achieve better results than the traditional method when the rotation speed is faster.