Fig. 2From: Dynamically localizing multiple speakers based on the time-frequency domainReal-life experiment setupBack to article page