PhD Materials

Comparison

Speech analysis/copy synthesis using four different methods; Sinusoidal Model (SM), Harmonic+Noise Model (HNM), adaptive Quasi-Harmonic Model (aQHM) and STRAIGHT. Here

Software in MATLAB

Remark: The code provided here represents working code and I do not claim that it is free of bugs. I decided to put it on this page, in order to make results reproducible for others. Feel free to use it and do not hesitate to contact me, if you have any problem, or, hopefully not, you found a bug!

DESA

Here you can download a version of Discrete Energy Separation Algorithm. Download

GMM-EM

A simple and fast implementation of Expectation-Maximazation for Gaussian Mixture Models. Download

LF model

Implementation of LF model for glottal source modeling. Download

Finding NE

An algorithm based on support enumeration for finding Nash Equilimbia in two-players non-degenerate games. Download

Basic QHM code

Reproducible Research for ICASSP 2010 paper: "On the Robustness of the Quasi-Harmonic Model of Speech" Download

Letter Code

Source code for the generation of figures of IEEE Signal Processing Letter "Iterative Estimation of Sinusoidal Signal Parameters". Download

Speech Examples

ICASSP 2010

Speech analysis/copy synthesis using Sinusoidal Model (SM) and adaptive Quasi-Harmonic Model (aQHM). Here

Web site and all contents (c) Copyright Yannis Pantazis 2008, All rights reserved. Last Update September 2010.