Speaker recognition system

This report givens an overview of a Gaussian Mixture Model – Universal Background Model (GMM-UBM) system which focusing on speaker identification. In this report we will be focusing on the traditional FFT-based Mel-Frequency Cepstral Coefficients (MFCCs) method to extract feature from wav file and G...

Full description

Saved in:
Bibliographic Details
Main Author: Song, Liyan.
Other Authors: Chng Eng Siong
Format: Final Year Project
Language:English
Published: 2012
Subjects:
Online Access:http://hdl.handle.net/10356/48504
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:This report givens an overview of a Gaussian Mixture Model – Universal Background Model (GMM-UBM) system which focusing on speaker identification. In this report we will be focusing on the traditional FFT-based Mel-Frequency Cepstral Coefficients (MFCCs) method to extract feature from wav file and GMM-UBM to create speaker model. The detail information of MFCC and GMM-UBM will be explained in the report. The program is build based using GMM-UBM and MFCC, the likelihood ratio of the testing speech are the output of the program. The experiment is carry out to evaluate the effects on accuracy when different mixture and file of MFC are used.