CHATR is a workbench environment for developing speech synthesis and voice-creation systems. It offers a simple but powerful modular architecture to create and modify speech. Many parallel modules are included, offering different ways of performing the same task. We will show you how to switch between different phoneme sets and lexicons, alternative intonation systems, and multiple methods for waveform synthesis. Setting of parameters and control of flow is done through CHATR's (Lisp) command interface, either interactively or from files. CHATR also helps you display the results of its synthesis. It can also be just a simple multi-language, multi-voice, text-to-speech system.
CHATR has many settable options and commands. Parts of the system are always under development, so it will be constantly changing (and hopefully improving). Any written document will inevitably be out of date and may not correspond to the latest implementation. These pages, however, contain many details which will not yet have been superseded.
(C) Copyright 1997
ATR Interpreting Telecommunications Research Laboratories.
All rights reserved.
This page was created by Martyn Weeks.
CHATR is being developed by Nick Campbell and the members of ATR-ITL Dept 2
Return to the Top, Go to the Introduction, Table of Contents.