Speech Recognition

Existing Solutions

• alphaWorks : Speech for Java
  
  Speech for Java is a Java programming interface for incorporating IBM's ViaVoice speech techonology into user interfaces.
   
• CloudGarden JSAPI
  
  CloudGarden's Java Speech API implementation for Windows.
   
• Java Speech API FAQ
  
  A collection of frequently asked questions (FAQ) that provides brief answers to many common questions about the Java Speech API (JSAPI).
   
• Microsoft Speech SDK v5.1
  
  Microsoft's SDK that includes text-to-speech (TTS) engines and speech recognition (SR) engines.
   
• ScanSoft
  
  "The leading supplier of speech and imaging solutions."
   
• Nuance
  
  Another company providing speech products and services.
   

Open-Source Solutions

• CMU Sphinx
  
  A real-time, large vocabulary, speaker independent speech recognition system.
   
• Open Mind Speech
  
  An open source project that aims to provide free speech recognition tools. applications.
   
• ISIP Automatic Speech Recognition
  
  Recognition software and various tools developed at Institute for Signal and Information Processing in Mississippi State.
   

Basic Algorithm

• Applying Hidden Markov Models
  
  Description about how to use Hidden Markov Models and phoneme as its sound unit to build a recognition engine.
   
• Basic algorithm not using Hidden Markov Models
  
  The simplest algorithm that contains only 2 phases in its design.
   

Pre-Processing

• Pre-Emphasis
  
  Provided by University of Cambridge's Speech Vision Robotics Group, this page has a brief description of what the pre-emphasis filter is and how to implement it.
   
• Isolated Word Endpoint Detection
  
  This page has an endpoint detector demo. as well as a brief overview of what the algorithm introduced by L. R. Rabiner and M. R. Sambur.
   

Feature Extraction and Digital Signal Processing

• Linear Predictive Coding (LPC)
  
  One of the speech analysis techniques available today.
   
• Front-end feature extraction algorithm
  
  Detail information on how to program methods to extract Mel-Frequency Cepstral Coefficients.
   
• Yehar's digital sound processing tutorial for the braindead!
  
  Information on digital signal processing including topics on filters.
   
• The Scientist and Engineer's Guide to Digital Signal Processing
  
  An online book on concepts in Digital Signal Processing.
   
• Digital Signal Processing Tutorial
  
  Introductory tutorial on Digital Signal Processing.
   

Vector Quantization

• Vector Quantization
  
  Explains what Vector Quantization is and also gives sample code and animated example.
   

Hidden Markov Models

• A Tutorial on Hidden Markov Models and Selected Applications in Speech Recognition
  
  Rabiner's tutorial on Hidden Markov Models.
   
• GNU Hidden Markov Model Library
  
  Hidden Markov Models Library.
   
• Hidden Markov Models
  
  Another page that shows how to program Hidden Markov Models.
   
• HMM
  
  Defines Hidden Markov Models and how it is characterized.
   
• Hidden Markov Models - Introduction
  
  A Slideshow presentation on Hidden Markov Models.
   
• HMM Methods in Speech Recognition
  
  Another page that shows how to Hidden Markov Models works.
   
• HMM Tut.
  
  A simple Hidden Markov Models tutorial.
   
• Hidden Markov Model Development Kit
  
  HMMSDK provides the core library HMMLib as well as other tools.
   

Audio File and Voice Capture

• Java Sound API
  
  Java Sound API can be used to capture voice from microphone.
   
• libsndfile
  
  libsndfile is an open source C library for reading and writing audio files in many different formats.
   
• spwave
  
  A great open source software that plays audio files in many different format and makes use of libsndfile library. There is Windows, Mac OS and Linux versions available.