Post-processing for Bss Algorithms to Recover Spatial Cues

This paper addresses the problem of recovering spatial cues after microphone array processing by blind source separation. Based on the known demixing system determined by the blind source separation, we derive two spatialization algorithms. One algorithm exploits the inverse of the demixing system, while the other algorithm exploits the adjoint of the demixing system. Both algorithms are evaluated by objective and subjective measures. We therefore consider the recovered time difference of arrival and the subjective perception of the spatialized signals.