Using HTK & Julius

HTK (Hidden Markov Toolkit) is a portable toolkit for building and manipulating the statistical models used to represent sound in Speech Recognition (these are called hidden Markov models).   Julius is large vocabulary continuous speech recognition engine.

These two tutorials run through the steps to create an Acoustic Model for the Julius Speech Recognition Engine using the HTK toolkit, and show how to submit your audio files to VoxForge.


How-to This How-to uses a script is to automate most of the steps in the creation of a Speaker Dependent Acoustic Model.  This How-to only gives a quick overview of the required steps. 


Tutorial In this tutorial, all the steps are performed manually, with some explanations, following the approach used in the HTK Book tutorial.