Informed Group-Sparse Representation for Singing Voice Separation