Investigating the influence of loudspeakers equalization on crosstalk cancellation performance for headrest applications

Immersive audio aims to create a sound field that allows listeners to perceive sound sources at specific spatial locations. While binaural headphone reproduction provides accurate spatial rendering, loudspeaker reproduction suffers from crosstalk that degrades this perception. Crosstalk cancellation (CTC) can mitigate the problem, though its effectiveness depends on the loudspeaker-listener configuration, speaker characteristics, listener anatomy, and room acoustics. In particular, headrest-based binaural systems, where loudspeakers are positioned close behind the listener’s head, are becoming increasingly popular due to their suitability for integration into transportation settings. A key advantage of these systems is the consistency of loudspeaker placement across different seat designs. In this context, this work investigates the influence of loudspeaker equalization on crosstalk cancellation performance in a two-loudspeaker headrest configuration assuming a fixed listening position. Crosstalk cancellation filters are derived from measured head-related transfer functions (HRTFs) using the fast deconvolution with regularization method. The proposed approach separates the loudspeaker electro-acoustic characteristics from the HRTFs during the filter design and allows the use of different loudspeakers for reproduction by means of an ad-hoc equalization function. The impact of loudspeaker equalization on CTC filter design and on the reproduction system is objectively assessed. Subjective tests demonstrate that, by properly equalizing the loudspeakers for HRTFs measurement and for sound reproduction, effective crosstalk cancellation can be maintained even when the loudspeaker setup changes, without the need to re-measure the HRTFs.