Abstract
Automated speech recognition systems work typically with pronunciation dictionary for generating expected phonetic content of particular words in recognized utterance. But the pronunciation can vary in many situations. Besides the cases with more possible pronunciation variants specified manually in the dictionary there are typically many other possible changes in the pronunciation depending on word context or speaking style, very typical for our case of Czech language. In this paper we have studied the accuracy of proper selection of automatically predicted pronunciation variants in Czech HMM ASR based systems. We have analyzed correctness of pronunciation variant selection in forced alignment of known utterances used as an ASR training data. Using the proper pronunciation variant, more exact transcriptions of utterances were created for further purposes, mainly for the more accurate training of acoustic HMM models. Finally, as the target and the most important application are LVCSR systems, the accuracy of LVCSR results using different levels of automated pronunciation generation were tested.
This is a preview of subscription content, log in via an institution to check access.
Preview
Unable to display preview. Download preview PDF.
Similar content being viewed by others
References
Psutka, J., Ircing, P., Psutka, J.V., Hajič, J., Byrne, W.J., Mírovský, J.: Automatic transcription of Czech, Russian, and Slovak spontaneous speech in the MALACH project. In: Proc. Interspeech 2005, Lisbon, Portugal, pp. 1349–1352 (2005)
Nouza, J., Ždánský, J., David, P., Červa, P., Kolorenč, J., Nejedlová, D.: Fully automated system for Czech spoken broadcast transcription with very large (300K+) lexicon. In: Proc. Interspeech 2005, Lisbon, Portugal, pp. 1681–1684 (2005)
Dupont, S., Ris, C., Couvreur, L., Boite, J.-M.: A study if implicit and explicit modeling of coarticulation and pronunciation variation. In: Proc. Interspeech 2005, Lisbon, Portugal, pp. 1353–1356 (2005)
Wolff, M.: On representation and training of pronunciation dictionaries. In: 8th Czech-German Workshop ’Speech Processing’, Prague, Czech Republic (1998)
Wolff, M., Eichner, M., Hoffmann, R.: Evaluation of automatically trained pronunciation dictionaries. In: Proc. Czech-German WS on Speech Processing, Prague, Czech Republic (2002)
Pollák, P., Hanžl, V.: Tool for Czech pronunciation generation combining fixed rules with pronunciation lexicon and lexicon management tool. In: Proc. of LREC 2002, Third International Conference on Language Resources and Evaluation, Las Palmas, Spain (May 2002)
Pollák, P., Volí, J., Skarnitzl, R.: Influence of hmm’s parameters on the accuracy of phone segmentation - evaluation baseline. In: ESSP 2005, Electronic Speech Signal Processing, Prague (September 2005)
Wells, J.C., et al.: Czech SAMPA home page (2003), http://www.phon.ucl.ac.uk/home/sampa/czech-uni.htm
Palková, Z.: Czech phonetics and phonology. In: Czech language - Fonetika a fonologie češtiny, Charles University. Karolinum (1994)
Pollák, P., Volí, J., Skarnitzl, R.: Analysis of glottal stop presence in large speech corpus and influence of its modelling on segmentation accuracy. In: 16th Czech-German Workshop on Speech Processing, Prague (September 2006)
Psutka, J., Müller, L., Matoušek, J., Radová, V.: Mluvíme s počítačem česky (Talking to the Computer in Czech). Academia, Prague (2006)
Pollák, P., Černocký, J.: Czech SPEECON adult database (November 2003), http://www.speechdat.org/speecon
Young, S., et al.: The HTK Book, Version 3.3, Cambridge (2005)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2009 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Hanžl, V., Pollák, P. (2009). Accuracy Analysis of Generalized Pronunciation Variant Selection in ASR Systems. In: Esposito, A., Vích, R. (eds) Cross-Modal Analysis of Speech, Gestures, Gaze and Facial Expressions. Lecture Notes in Computer Science(), vol 5641. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-03320-9_37
Download citation
DOI: https://doi.org/10.1007/978-3-642-03320-9_37
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-03319-3
Online ISBN: 978-3-642-03320-9
eBook Packages: Computer ScienceComputer Science (R0)