SASNet Stanford Audio Separation Network. A deep learning method based of wavUnet for audio source separation.