Fig. 3From: Deep encoder/decoder dual-path neural network for speech separation in noisy reverberation environmentsThe illustration of SISNR and SOSISNRBack to article page