Investigation on crosstalk cancellation introducing loudspeakers directivity in HRTFs approximation

In the field of immersive audio, crosstalk cancellation is adopted preserving the spatial characteristics of the reproduced signals. It is based on the use of HRTFs that can be approximated through a free-field model. However, this method assumes that the physical head is removed and replaced with two monopole receivers at the position of the ears and thus the source characteristics is not considered. In this context, a new approximation for the HRTFs is introduced taking into account the source directivity, i.e., the loudspeaker is modelled as a rigid sphere with a membrane vibrating on a spherical cap. A deep investigation on the effect of introducing the source directivity in the free-field model is reported in this paper. In particular, several experimental results are shown, considering both objective and subjective measurements, to validate the effectiveness of the proposed solution in terms of both channel separation and audio quality preservation.