This page summarise some of the available resources for building spoken dialogue systems.

General tools

Automatic speech recognition

  • C++ decoder
  • decoding with 60k words vocabulary in real-time
  • generates confusion networks
  • free for commercial use
  • C++
  • WFST - based decoder,
  • Also includes AM training tools
  • WFST based decoder
  • can be user with OpenFST
  • free for commercial use
  • C++ decoder (Pocketsphinx)
  • The rest is in Java
  • Also includes acoustic and language model training tools 
RWTH ASR - http://www-i6.informatik.rwth-aachen.de/rwth-asr/
  • C++
  • WFST
  • generates confusion networks
  • non-free for commercial use
  • Python
  • OpenFST
  • probably not maintained any more
The acoustic models can be trained using:
Acoustic data:

Speech synthesize

  • C++
  • System for building voices
  • Includes unit selection synthesis and HTS 
Project Festvox - http://festvox.org/
  • Source of new voices for Festival and Flite
  • Java 
  • Complete toolkit
  • Unit selection and HTS