LCE Kotisivu

S-114.720 Kommunikaatio ja kognitio -tutkijaseminaari

Multimodal speech perception and production

Syksy 2004/Fall 2004 (2ov, L)


Michael Frydrych

Sisältö - Content

Although visible speech is not as informative as auditory speech, because many phoneme articulations look similar, visible articulatory information is an important part of general speech perception. Visible speech is particularly effective when the auditory speech is degraded, because of the noise, bandwidth filtering or hearing impairment. Compelling evidence for the importance of visible speech is provided by examples where auditory speech signal is non-degraded. It has been shown that the speech perception improves significantly also with computer animated audio-visual speech synthesizers, so-called talking heads.

This course will give the students an overview on issues related to multimodal speech perception and synthesis. The focus will be on visible speech.

The seminars will be held in English.

Program of seminars:

  1. Introduction
  2. Multimodal Speech Systems
  3. Methods and Standards for Facial Animation and Modeling
  4. Multisensory Integration, Information for Visual Speech Perception
  5. Spatial and Temporal issues, Coarticulation
  6. Articulatory Models and Parametric Control
  7. Measurement of Facial Motion, Relation to Acoustics

Vaatimukset - Requirements

Active participation in the seminar. Each student acts once as talker and once as opponent, and writes a short report from the presented research article.

Kirjallisuus - Literature

Will be given at the first seminar.

Aikataulu - Schedule

The seminars will be held on tuesdays at 13.15-14.45 in B317, Tekniikantie 14, Innopoli 2, starting October 26th, 2004, with December 7th the last seminar. Please note that the number of participants is limited to 12.

Ilmoittautumiset ja lisätietoja - Registration and Further information

Michael Frydrych

Tästä sivusta vastaa
Sivua on viimeksi päivitetty 25.10.2004 / Last updated October 25, 2004