TAPI in telephone quality speech database

It is always been the speech recognition team’s vision to be able to apply the developed speech technology in real world so that more people can benefit form it. One of the targets is towards telephony. People will be able to talk comfortably to computer through telephone to obtain certain informa...

Full description

Saved in:
Bibliographic Details
Main Author: Bongsu, Mohd. Shukri
Format: Thesis
Language:English
Published: 2007
Subjects:
Online Access:http://eprints.utm.my/id/eprint/6119/1/MohdShukriBongsuMFKE2007.pdf
http://eprints.utm.my/id/eprint/6119/
http://dms.library.utm.my:8080/vital/access/manager/Repository/vital:62017
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:It is always been the speech recognition team’s vision to be able to apply the developed speech technology in real world so that more people can benefit form it. One of the targets is towards telephony. People will be able to talk comfortably to computer through telephone to obtain certain information. Studies and effort have been carried out to improve the accuracy and efficiency of telephone speech recognition. This project aims to build a computer telephony using Telephony Application Programming Interface (TAPI) to collect telephone quality speech database, which are very useful in testing and improving certain speech recognition system’s ability in recognizing telephone speech. First, discussion on TAPI itself will be presented. Then the whole system will be designed using TAPI. The implementation of the design will be done by using Visual Basic 6.0 as programming language. When the computer telephony is completed, speech samples will be collected to keeps as database. The speech samples will be collected through various type of Public Services Telephone Network. Parts of this database are the will be used for experiments to compare the performance of a certain speech recognition system that are trained in two ways: using soundcard quality speech (clean speech) as training samples and other one using telephone quality speech samples.