Comparative Study of Several Novel Acoustic Features for Speaker Recognition