Directeur de Recherche CNRS
Tel: +33 3 83 59 20 36
Fax: +33 3 83 41 30 79
e-mail: Yves.Laprie [at] loria.fr
I work on acoustic-to-articulatory inversion of speech signals and speech analysis. I am especially interested in the analysis-by-synthesis approach because I am convinced that it also enables the speech production process to be better understood. Together with Slim Ouni and Blaise Potard we proposed inversion algorithms guaranteeing a very good closeness with original acoustic data together with realistic articulatory gestures.
Since articulatory models play a central role in this approach I developed software (Xarticul) to process old X-ray images and build models, which approximate the vocal tract for vowels and consonants better than existing models. These tools also enable articulatory copy synthesis.
I am also interested in speech analysis and I am the main author of the WinSnoori software http://www.winsnoori.fr which offers a variety of speech analysis algorithms (pitch detection, automatic formant tracking, copy synthesis for a parallel formant synthesizer). Tools derived from WinSnoori are now used in foreign language learning.
I am especially interested in supervising PhD students or postdocs in the areas of articulatory processing of speech.
Two postdoc positions for 2015
PhD subject proposed for 2015-2018 (deadline May 31, 2015): Articulatory synthesis of speech
Snorri and WinSnoori are speech research software aimed at assisting researchers in the fields of speech recognition, phonetics, perception and signal processing.
Xarticul, software to process Xray images and other medical images of the vocal tract.
Mathematical tools (Guide to Available Mathematical Software): link to http://gams.nist.gov/
Introduction à l'analyse de la parole: Pdf version
Introduction à l'acoustique de la parole: Pdf version