This file contains the main part of DSR front-end. Speech samples * are read from input waveform file frame by frame. Feature * extraction is performed for each frame by calling basic * functions of the basic FE function package (see FEfunc.