Ming Xiao Shengli Xie Yuli Fu Statistically non-sparse decomposition of two underdetermined audio mixtures. 458-462 2008 IJCNN http://dx.doi.org/10.1109/IJCNN.2008.4633832 conf/ijcnn/2008 db/conf/ijcnn/ijcnn2008.html#XiaoXF08