BAST-Mamba

Human listeners localize sound almost entirely from two cues, the tiny time gap between a sound reaching one ear before the other, and the level difference caused by the head blocking sound from the far ear. Prior deep learning attempts…