Crosstalk cancellation system (CCS) is a technique for spatial sound reproduction by using two loudspeakers to deliver binaural audio signals to a listener’s ears. However, it is well-known that it’s quite sensitive to the listener’s head movement. So researchers of Institute of Acoustics, Chinese Academy of Sciences (IACAS) carried out a series of studies and put forward a new method to optimize this system.
The researchers propose a multi-position weighted method. Based on the former research result, the researchers optimize the criterion function. Defining -d = x (d=1,2,…,D) the head movement of d (cm) to the left away from the center. Similarly, d = x means head moves d (cm) to the right. A new criterion function is proposed through adhibiting Lagrange multiplier:
where 
correspond to the impulse responses from loudspeakers to ears at the head positions of d (cm) on the left and right respectively. By minimizing 
the optimal filter coefficients of multiple positions can be given by:
Through experiments, the non-uniform λd, with d = 2, 4 are selected to optimize overall channel separation performance and to provide increased robustness for head movement in low and medium frequencies. Experimental results demonstrate the validity of the proposed algorithm.
This research work was supported by the National Natural Science Foundation of China, the project of Guangdong Province Science and Technology program and Guangzhou City Higher Educational Science and Technology program.
