Microcomputer-based implementation of fast fourier transform for speech pattern analysis using ADSP-2101

Date

10-1998

Degree

Bachelor of Science in Applied Physics

College

College of Arts and Sciences (CAS)

Adviser/Committee Chair

Nelio C. Altoveros

Abstract

Development from voice recognition, discrete speech recognition, to continuous speech recognition has demanded much processing power on various computer systems, specially on Personal Computers (PC). Researches have aspired to develop computer systems that would respond to speech as it was spoken. Various companies have developed such continuous speech recognition systems for many powerful computer platforms, but not on the PCs. The ADSP-2101, a powerful processor specifically designed for analog and digital processing, was interfaced to the Personal Computer. Because of the nature of its instruction set, high speed performance, and reprogrammable property. the ADSP-2101 offers great possibilities for continuous speech recognition through faster FFT computations. The study showed that an FFT plot of the waveform from a certain speech diminished the task on phonetic pattern matching.

Language

English

Location

UPLB Main Library Special Collections Section (USCS)

Call Number

Thesis

Document Type

Thesis

This document is currently not available here.

Share

COinS