"An exemplary Java Speech Demo, Dynamic Time Warping"

Supervised by P.Green

Abstract

The objective of this dissertation is to investigate feasibilities of making demonstrative tools in Java for the taught speech courses in the department of Computer Science at University of Sheffield.

This dissertation is going to reveals graphical components of MATLAB Auditory Demo (MAD), an exemplary previous practice, and the other tools and literatures. The literatures include the necessary theories and phases of development of Java Auditory Demo (JAD) with the intension of providing only one or two demos.

One particular demo that included in the JAD is Dynamic Time Warping (DTW) introduced by Sakoe and Chiba (1978). A simple Automatic Speech Recognition (ASR), is the process by which a computer maps an acoustic speech signal to text, to draw the map and plots the cost path. However, another demo for Hidden Markov Models (HMMs) is excluded for further development.