Prosody in Speech Recognition and Understanding

October 22-24, 2001
Molly Pitcher Inn, Red Bank, NJ, USA

Implications of Prosody Modeling for Prosody Recognition

Chilin Shih (1), Greg Kochanski (1), Eric Fosler-Lussier (1), Melody Chan (2), Jia-Hong Yuan (3)

(1) Bell Laboratories, Lucent Technologies, Murray Hill, NJ, USA
(2) Yale University, USA
(3) Cornell University, USA

This paper introduces Stem-ML, which is a model of the prosody generation process with an associated description language, and suggests how it may help prosody recognition. We applied Stem-ML modeling to three topics: the modeling of prosodic strengths, intonation types, and noun phrase patterns. Stem-ML parameters derived from F0 contours may have a more consistent relationship with prosodic events than raw F0 values. This may improve identification of accent classes, accent strengths, and intonation types.


Full Paper (PDF)   Full Paper (Zipped Postscript)

Bibliographic reference.  Shih, Chilin / Kochanski, Greg / Fosler-Lussier, Eric / Chan, Melody / Yuan, Jia-Hong (2001): "Implications of prosody modeling for prosody recognition", In Prosody-2001, paper 25.