International Symposium on Tonal Aspects of Languages
With Emphasis on Tone Languages

Beijing, China
March 28-31, 2004

Quantitative Tone and Intonation Modeling across Languages

Hansjörg Mixdorff

Faculty of Computer Science, TFH Berlin University of Applied Sciences, Germany

Although it has been shown for many languages that the Fujisaki model is in principle capable of decomposing F0 contours in these languages, problems remain as to the automatic extraction as well as the interpretation of parameters extracted with respect to their phonological relevance. The claim is made in the current article that the Fujisaki model parametrization used must be rooted in the linguistic system of the particular language, and that this consideration should always be given priority over simple fitting accuracy. To this effect, this article presents settings and results from a number of more recent studies by the author and his co-workers in various languages, such as German, Finnish, Thai and Vietnamese.

Full Paper

Bibliographic reference.  Mixdorff, Hansjörg (2004): "Quantitative tone and intonation modeling across languages", In TAL-2004, 137-142.