Fig. 4From: A neural network-supported two-stage algorithm for lightweight dereverberation on hearing devicesLog-energy spectrograms of clean, reverberant, and processed utterances. \(\textrm{T}_{\textrm{60}} = 0.68\textrm{s}\). HA scenario [ \(\varvec{\nu } = \varvec{d} + \varvec{e}\) ; \(\Delta = \tilde{\Delta } = 5\) ] Heavy speech distortions can be observed in the DNN-PF output, as highlighted in the red ellipseBack to article page