14th International Congress of Phonetic Sciences (ICPhS-14)
San Francisco, CA, USA
In automated phonetogram registration, several acoustic voice parameters are computed as a function of fundamental frequency and vocal intensity. We show how such a phonetogram can be described by a set of Hidden Markov Models. This allows a comparison of phonetograms utilizing all underlying data. The method makes it possible to recognize new phonetograms in terms of the most likely phonetogram that could have generated the new data. Promising implications for increased clinical usefulness of phonetograms are discussed.
Bibliographic reference. Bloothooft, Gerrit / Dorssen, Margreeth van / Koppen, Kim / Pabon, Peter / Wijk, Mieke van (1999): "On the comparison of computerised phonetograms", In ICPhS-14, 1001-1004.