Computational cost reduction using coincident boundary microphones for convolutive blind signal separation

It is demonstrated that an acoustic mixture system can be configured to accomplish the instantaneous mixture model conditions. The system in achievable using coincident boundary microphones, which is a suitable disposition to uniform the propagation channel delays and to reduce the number of reflections characterising the system impulse response. With the use of coincident microphones, convolutive BSS algorithms provide optimal results with reduced number of taps, thus, decreasing computational cost.