Phase-based speech enhancement

Other Title(s)

التحسين على الصوت باستخدام الطور

Dissertant

Shihadah, Bilal Abd al-Rauf Muhammad

Thesis advisor

al-Hanjuri, Muhammad Ahmad Muhammad

University

Islamic University

Faculty

Faculty of Engineering

Department

Department of Computer Engineering

University Country

Palestine (Gaza Strip)

Degree

Master

English Abstract

Speech is a fundamental means of human communication.

Speech enhancement has been a challenge for many researchers for almost three decades.

The problem involves improving the performance of speech communication systems and improve the quality and intelligibility of degraded speech in noisy environments.

The majority of speech enhancement methods perform noise removal in spectral domain and construct the enhanced speech signal from the estimated magnitude of clean speech discarding the phase side of the noisy speech.

In this thesis,the researcher aims to show the effect of modifying the complex spectrum of enhancement process in the quality of the signal in term of perceptual evaluation of speech quality (PESQ).

The main job of this study is to show the importance of the complex value of wave atoms transform for enhancing the degraded signalswith minimum speech distortion.

The researcher use NOIZEUS dataset and PESQ measurement in comparison between techniques.This study use wave atoms transform as main enhancement technique combining it with other transformation techniques like Karhunen–Loeve transform and use different statistical probability model like maximum a posteriori, prior probability and posteriori probability with speech presence probability.

The result of the thesis shows that best enhancement achieved when use multi transformation technique with wave atoms.

This thesis reach 17.5% PESQ enhancement with combining wave atoms with KLT and prior probability uncertainty.

The proposed algorithms are implemented using MATLAB environment.

The results indicate that the proposed system improves the classical methods performance in terms of speech quality.

Keywords: PESQ, KLT, Wave Atoms, MAP, Prior probability, Posteriori probability, speech presence probability

Main Subjects

Electronic engineering

No. of Pages

112

Table of Contents

• APA
• MLA
• AMA

Language

English

Data Type

Arab Theses

Record ID

BIM-531455