Fig. 1From: Speech dereverberation for enhancement and recognition using dynamic features constrained deep neural networks and feature adaptationIllustration of three processing stages: beamforming log-magnitude spectrum enhancement, feature enhancement. DS and MVDR are the delay and sum and minimum variance distortionless response beamforming, respectively. SS refers to spectral subtraction, while DNN stands for deep neural network-based speech coefficient mapping. CT is the cross transform feature adaptation. Three enhanced signals are produced for the speech enhancement task, and two enhanced features are produced for the ASR taskBack to article page