README file for the CD annexed to the master thesis titled:
  "Techniques d'apprentissage machine appliquées à la modélisation du
  timbre d'un individu pour la synthèse de voix réaliste" par
  Pierre-Alexandre Fournier.  December 13th 2005

------------------------------------------------------------------

Copyright 2005 Pierre-Alexandre Fournier, all rights reserved.


This CD contains various files related to the research described in my
thesis. Here is a short description:
 

Fournier_memoire20051208.pdf : The final version of my thesis (in
French).


images/ : Images used in my thesis and in the presentation.

images/sources.txt : Description of the source of the images, when I
am not the author of these images.


presentation/ : Files used during my thesis presentation, Thursday
November 24th 2005.

presentation/fournier_presentation.pdf : Presentation slideshow (in
French).

presentation/tts-presentation.sh : shell script using mplayer to play
the *.wav files introducing my presentation.

presentation/*.wav : Wave files used in the presentation. These files
were created with the demo of AT&T's text-to-speech software
(http://www.research.att.com/projects/tts/demo.html). These files were
created respecting the policy regarding the use of the online demo
located at this address ("A class project is probably OK.") :
http://www.research.att.com/projects/tts/faq.html#WebPolicy


voyelles/ : Folder containing synthesized voice samples using measured
(orig) and predicted (pred) spectrums from the 18 first speakers for
the 5 vowels studied in this research.
(aa = /a/, ee = /e/, ii = /i/, oo = /o/, uu = /u/).


NOTE: If you use the files on this CD, or if you cite this research,
please contact me at this address:

pierre-alexandre.fournier@polymtl.ca



