Here's an overview of our ECE 846 Speech Tools Software:

The tools are used for a series of practical labs in conjunction with a speech processing course. Routines are included for visual phoneme segmentation, speech production modeling, coding, cepstrum, zero-crossing, etc. The tools were coded by Duanpei Wu with modifications and additions by Eric Patterson and Sami Niska. The software runs under Linux or Solaris and possibly other Unix platforms. It is freely available with the Linux binary and all source code.

The downloadable package is available under the "Tools" page on this website. If you find the software useful, interesting, or have other comments, please send us an email. We have a series of labs in conjunction with the software that may be posted on the web in a future update.

It is recommended that the software be placed in /usr/local/ece846 and that "tcsh" is run as your command shell, with the lines in the file "addto.tcshrc" added to your .tcshrc file. We may be able to help if you have problems installing, but no official support is available, unfortunately.

Here are some screen captures.

The above image shows the root menu for the software.

 

This is a snapshot of the speech production model tool.

 

The image above shows basic coding examples from the speech coding tool..

 

This image shows newer routines for zero crossing, cepstrum, autocorrelation, etc.