Annotating Speech Data for Pronunciation Variation Modelling
2005 (English)In: Proceedings, FONETIK 2005, Göteborg, Sweden, 2005, 25-27 p.Conference paper (Other academic)
This paper describes methods for annotating recorded speech with information hypothesised to be important for the pronunciation of words in discourse context. Annotation is structured into six hierarchically ordered tiers, each tier corresponding to a segmentally defined linguis-tic unit. Automatic methods are used to seg-ment and annotate the respective annotation tiers. Decision tree models trained on annota-tion from elicited monologue showed a pho-neme error rate of 9.91%, corresponding to a 55.25% error reduction compared to using a canonical pronunciation representation from a lexicon for estimating the phonetic realisation.
Place, publisher, year, edition, pages
Göteborg, Sweden, 2005. 25-27 p.
Computer Science Language Technology (Computational Linguistics)
IdentifiersURN: urn:nbn:se:kth:diva-51892OAI: oai:DiVA.org:kth-51892DiVA: diva2:465186
FONETIK 2005, Department of Linguistics, Göteborg University
tmh_import_11_12_14. QC 201202012011-12-142011-12-142012-02-01Bibliographically approved