Auditory-Visual Speech Processing 2005

British Columbia, Canada
July 24-27, 2005

The History of Articulatory Synthesis at Haskins Laboratories

Philip Rubin (1,2), Gordon Ramsay (1), Mark Tiede (1,3)

(1) Haskins Laboratories, New Haven, CT, USA
(2) Yale University School of Medicine, Dept. of Surgery, New Haven, CT, USA
(3) Massachusetts Institute of Technology, Research Laboratory of Electronics, Cambridge, MA, USA

In the 1970s, Rubin, Baer and Mermelstein turned Mermelstein's vocal tract model into the first articulatory synthesizer (ASY) regularly used as a research tool for exploring the relationship between speech perception and production. This system was designed to allow for simple control of tract shape by direct manipulation of articulators. A significant extension is CASY, the configurable articulatory synthesizer, that lets users fit outlines of the vocal tract model to acquired sagittal images and represents fixed surfaces of the tract parametrically. Future developments include the integration of a new key articulator controlling parasagittal shape of the tongue dorsum motivated by volumetric MRI data and an improved voice source model. This paper will provide an overview of the history of articulatory synthesis at Haskins, including current work which aims to incorporate better models of biomechanics and fluid dynamics. The model will also be related to other approaches to articulatory synthesis.

Full Paper

Bibliographic reference.  Rubin, Philip / Ramsay, Gordon / Tiede, Mark (2005): "The history of articulatory synthesis at Haskins laboratories", In AVSP-2005, 117-118.