The following changes, additions and bug fixes were made to the OGI
Speech Tools since the BETA release in July 1993. The subdirectory
<speechtools/src/changes> contains the output after using the
unix {diff} command.


ADDITIONS
=========

VQ_LBG.c
~~~~~~~~
Vector Quantization using the algorithm of Linde Buzo and Gray (LBG)


phonebase
~~~~~~~~~
A database utility to efficiently automate speech related queries.
The utility creates a database based on a listing of speech files and
their corresponding location and label files (lola).  It allows the
user to specify a particular label or set of labels in a given
context, display all occurrences of the label, and relabel the
occurrences if desired.


BUG FIXES
==========

addhdr.c
~~~~~~~~

a. Changed to a variable sized buffer, when adding a header to a data
stream of unknown length. The previous version had a constant buffer
size, which causes problems when the incoming data stream is larger.
(addhdr.c - line 144-153)

b. The definition of the Byte ordering was also wrong and this has been
fixed. (addhdr.c - line 23,24)


auto_lyre.c
~~~~~~~~~~~
a. Spelling mistake fix in the Negative Range label in the TDAT data
   structure. (Gspectool.c - line 540)

b. Site dependent code (Lgraph.c - line 874)

c. Keyboard entry fix (ToolUtil.c). Changed overrideShellWidgetClass
to transientShellWidgetClass.


halfrate.c
~~~~~~~~~~
a. Output of halfrate is now no longer compressed before saving.
(line 94)


lyre.c
~~~~~~
See auto_lyre.c


libnspeech
~~~~~~~~~~
lpc.c
~~~~~
a. Fixed LPC-cepstrum calculation part.


rasta.c
~~~~~~~
a. Added a parameter tp set the pole of the rasta filter.


shortpack.c
~~~~~~~~~~~
a. Fixed file io handling when using stdin and stdout as the
   data streams.



CHANGES
=======

nopt.c
~~~~~~
a. Added two extra command line switches to control the number of
   hidden nodes and the random seed respectively.





