Multilanguage Telephone Corpus
                           Release Version 1.1

              Center for Spoken Language Understanding


UPDATED: 29 MArch 2002


This document describes the file naming conventions used
for this distribution and gives a brief description of the
various file formats used.

File Naming Convention
----------------------
File naming follows the following convention:

	KOcall-70-G.nlang.wav

The first field ("KOcall") is the prefix indicating the
Language in which this utterance was made, the second field
("70") represents a unique ID number for the speaker, and
the third field ("nlang") indicates the type of utterance.
Note that the "G" has no bearing on the contents of the
file.

The .wav and .ptlola files are divided into sub-directories
based on the speaker's language, and further divided into
sub-directories based on the speaker's unique ID number.
These directory names are based on ID number divided by
ten. So, utterances from speakers 0-9 are in the
subdirectory /0, while utterances from speakers 10-19 are
in the subdirectory /1, and so on.

The /labels directory has a file structures exactly
parallel to the structure of the /speech directory.

File Formats
------------
The .wav file format used in this corpus is the RIFF
standard file format. This file format is 16-bit
linearly encoded.

The .ptlola file is a file containing time aligned
phonetic transcriptions. This file can be viewed with 
any standard text editor.