Programmed by Chanwoo Kim for the ASRU 2009 package
(chanwook@cs.cmu.edu; chanwcom@gmail.com)  

SPB_D and SPB_R programs are included in this package:

Input should be at 16 kHz, and we assume that the input is uncompressed
raw PCM file format with 1024 byte header. (2 bytes per data)

1) SPB-D
SPB_D(szOutFeatFileName, szInFileName, 0)

2) SPB-D-LSMS
SPB_D(szOutFeatFileName, szInFileName, 1)

3) SPB-R
SPB_R(szOutFeatFileName, szInFileName, 0)

4) SPB-R-LSMS
SPB_R(szOutFeatFileName, szInFileName, 1)

In SPB-R and SPB-R-LSMS, if you change bFeat option to 0, then it resynthesizes
speech instead of doing feature extraction.

July 14, 2009