Source code for "Visually aligned sound generation via sound-producing motion parsing" (Published at Neurocomputing)