Fig. 17From: Learning-based robust speaker counting and separation with the aid of spatial coherenceComparison of separation performance in terms of a PESQ and b WER for the LibriCSS datasetBack to article page