Date of Award

5-2013

Degree Type

Report

Degree Name

Master of Science (MS)

Department

Electrical and Computer Engineering

Committee Chair(s)

Jacob H. Gunther

Committee

Jacob H. Gunther

Committee

Todd K. Moon

Committee

Jeffrey B. Larsen

Abstract

A speech synthesis system with an original user interface is being developed. In contrast to most modern synthesizers, this system is not text to speech (TTS). This system allows the user to control vowels, vowel transitions, and consonant sounds through a simple 2-d vowel pad and consonant buttons. In this system, a synthesized glottal waveform is passed through vowel filters to create vowel sounds. Several filters were calculated from recordings of vowels using linear predictive coding (LPC). The rest of the vowels in the North American English vowel space were found using interpolation techniques with line spectral frequencies (LSF). The effectiveness and naturalness of the speech created from transitions between these filters was tested.

In addition to the vowel filters, filters for nasal and liquid consonants were found using LPC analysis. Transition filters between these consonants and vowels were determined using LSFs. These transitions were tested as well.

Share

COinS