(#video) X-Ray Film Database for Speech Research "It's ten below outside"

The X-Ray Film Database is a collaborative project by Dr. Kevin Munhall of Queen's University and Drs. Eric Vatikiotis-Bateson and Yoh'ichi TohkuraATR Human Information Processing Research Laboratories, Kyoto, Japan. It was conceived to preserve cineradiographic vocal tract footage, to make the images available to the speech research community, and to develop techniques for the automated digital processing of the images. of


Forty years after they were first utilized in speech research, high speed x-ray films yield the best dynamic view of the entire vocal tract and provide important information about it's time-varying properties. In particular, they accurately depict the complex movements of the tongue.


Due to ethical concerns about the high radiation dosages necessary, x-ray imaging technology using normal subjects is no longer practiced and it has become imperative to preserve those films which were origially captured on the fragile medium of 35 mm film.


The Queen's University/ATR Labs X-Ray Film Database for Speech Research [1],[2] offers twenty-five high quality x-ray films, totalling 55 minutes of footage, compiled on CAV laserdisk. The films were contributed by C. Rochette (Université Laval), and K. Stevens and J. Perkell (M.I.T.). The subjects are 14 native speakers of Canadian English or French, reading phonetically contrastive sentences. The x-ray film database is available to researchers at no cost, with a limitation of one disk per institution. A DAT recording of the original audio tracks is also available. To cover the cost of materials and postage, there is a small fee for the DAT.





Cineradiography examples:


These sample movies were digitized from the Queen's University/ATR cineradiographic database. Copyright for the images remains with the original researchers and their institutions. Non-commercial use of the images for research purposes is unrestricted.


The movies are available in MPEG format, with accompanying AIFF audio file, or as a QuickTime movie, which includes sound. These have been subsampled and optimized to transfer quickly, at the expense of the original image quality. A full source video size (640x480) JPEG image is available for comparison purposes.


"It's ten below outside"

MPEG movie 81K
AIFF auido file 29K
JPEG image 31K
Quicktime movie 233K


This sample was taken from a movie recorded in 1974 at the Départment de Radiologie de l'Hôtel-Dieu de Québec, Québec, Canada. Supervising the project was Dr. Claude Rochette. For information regarding the film recording technique used for this and other films, please see Rochette (1973) Les Groupes de Consonnes en Français, Les Presses de l'Université Laval, Québec. The subject was a 26 year old male native speaker of Canadian English who attended primary school in British Columbia.


"Le boulanger but onze bières"


MPEG movie 105K
AIFF auido file 45K
JPEG image 32K
Quicktime movie 307K


This sample was taken from another Université Laval film recorded in 1974 by Dr. Rochette. The subject of this movie was a female native speaker of Canadian French.


"Why did Ken set the soggy net on top of his deck"


MPEG movie 195K
AIFF auido file 47K
JPEG image 34K
Quicktime movie 469K


This sample was taken from a movie recorded in 1962 at the cineradiographic facilty of the Wenner-Gren Research Laboratory at Nortull's Hospital, Stockholm, Sweden. The investigators were Drs. K. Stevens and S. Öhman. The film recording techniques are documented in detail and the film is analyzed in part in J.S. Perkell (1969) Physiology of Speech Production, MIT Press, Cambridge, MA. The subject was a 38 year old male native speaker of Canadian English who attended primary school in Ontario.