sti.py : Speech Transmission Index (STI) from speech waveforms
Overview
--------
sti.py provides a library of functions used to calculate the speech
transmission index (STI) from real speech waveforms. The speech trans-
mission index provides an index measure of speech intelligibility, useful
for evaluating transmission channel or codec suitability to carrying speech
audio. The method used by sti.py was described first by Payton and Braida
in 1999, and is an adaptation of the original STI method developed by
Houtgast and Steeneken in 1985. Refer to those publications for further
technical information on the index itself or the calculation approach.
SOFTWARE DEPENDENCIES
---------------------
matplotlib
Numpy
Scipy, release version 0.8 or greater
INSTALLATION
------------
No installation is required OR AVAILABLE at this time. The only method of
running the program currently is by placing the 'sti.py' file into a python
project's source directory or including it in your module search path.
Usage
-----
sti.py provides library functions for calculating the speech transmission
index. At this time no command line interface or GUI utility is available
for accessing the functions. See the 'test.py' script contained in the
source code directory for a brief example of calculating the STI from
original and distorted raw audio samples.
Authors
-------
Jonathan Polom <jmpolom@wayne.edu>
Date
----
22 April 2011
Version
-------
0.5
LICENSE
-------
See 'COPYING' for the details of the GNU General Public License
Copyright (C) 2011 Jonathan Polom <jmpolom@wayne.edu>