Using HTK & Julius

This tutorial runs through the steps to *adapt*  a pre-existing acoustic model, such as the VoxForge acoustic model, to your voice using the HTK toolkit.  The VoxForge acoustic model is speaker independent.  Adapting it with your voice will increase its recognition accuracy for your voice, which can then be used with the Julius Speech Recognition Engine.

Note: This Tutorial was written when the most current version of HTK was release 3.3 (the most current release is HTK r3.4.1).  It uses an even older version of HTK (r3.2.1)  because there were problems with adaptation using HTKr3.3 (see Ticket #41 for details).

Julius works with models trained with any HTK release (3.3, 3.2.1 or the most current version of HTK)

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.