The 64 files trackXX.NIST correspond to the speech examples described in
Appendix E of Robert Donovan's PhD thesis "Trainable Speech Synthesis". 
The thesis is available by anonymous ftp to svr-ftp.eng.cam.ac.uk in 
pub/reports/donovan_thesis.ps.Z

Each speech file is a 16kHz 16-bit waveform file with a 1024 byte NIST header.