Fig. 4
Speech spectrograms before and after postprocessing. a Clean speech. b Speech enhanced by DARCN. c DARCN with postfiltering based on SPP-MMSE. d DARCN with postfiltering based on SPP-proposed-1. e DARCN with postfiltering based on SPP-proposed-2. f DARCN with postfiltering based on SPP-proposed-3