Frank Seide and Alfred Mertins 1994
In Proc. IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP)