A recurrent encoder-decoder approach with skip-filtering connections for monaural singing voice separation | Publicación