Vinay Kothapally(The University of Texas at Dallas - Center for Robust Speech Systems (CRSS)), Wei Xia(University of Texas at Dallas), Shahram Ghorbani(Center for Robust Speech Systems (CRSS), University of Texas at Dallas, Richardson, TX 75080), John H.L. Hansen(Univ. of Texas at Dallas; CRSS - Center for Robust Speech Systems), Wei Xue(JD AI Research) and Jing Huang(JD AI Research)
The reliability of using fully convolutional networks (FCNs) has been successfully demonstrated by recent studies in many speech applications. One of the most popular variants of these FCNs is the `U-Net', which is an encoder-decoder network with skip connections. In this study, we propose `SkipConvNet' where we replace each skip connection with multiple convolutional modules to provide decoder with intuitive feature maps rather than encoder's output to improve the learning capacity of the network. We also propose the use of optimal smoothing of power spectral density (PSD) as a pre-processing step, which helps to further enhance the efficiency of the network. To evaluate our proposed system, we use the REVERB challenge corpus to assess the performance of various enhancement approaches under the same conditions. We focus solely on monitoring improvements in speech quality and their contribution to improving the efficiency of back-end speech systems, such as speech recognition and speaker verification, trained on only clean speech. Experimental findings show that the proposed system consistently outperforms other approaches.