Fig. 4From: Progressive loss functions for speech enhancement with deep neural networksConvolutional block. Basic block for composing P-CNN and P-ResNet architecturesBack to article page