Abstract
This paper describes the neural network algorithm flexibly incorporated into the text-to-speech (TTS) system Epos. A particular network forming the fundamental frequency F0 in Czech indicative sentences read in neutral voice is described. The results show that the network was able to learn the characteristics of the speaker. The output generated with its usage is much more lively than the previous one formed by rules which were collected over several years.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Hanika, J., Horák, P.: Epos – A New Approach to the Speech Synthesis. In: Proceedings of the First Workshop on Text, Speech and Dialogue – TSD 1998, Brno, Czech Republic, September 23-26, pp. 51–54 (1998)
Hanika, J., Horák, P.: Text to Speech Control Protocol. In: Proc. of the Int. Conf. Eurospeech 1999, Budapest, Hungary, September 5-9, vol. 5, pp. 2143–2146 (1999)
Hanika, J.: Epos on-line documentation, http://epos.ure.cas.cz/
Horák, P.: Modelling of the Suprasegmental Characteristics of Spoken Czech Using Linear Prediction. Ph.D. thesis, Czech Technical University, Prague (2002) (in Czech)
Adámek, J.: Neural Networks Controlling Prosody of Czech Language. Master thesis, Faculty of Mathematics and Physics, Charles University, Prague (2002)
Horák, P.: Automatic Speech Segmentation Based on DTW with the Application of the Czech TTS System. In: Keller, E., Bailly, G., Monaghan, A., Terken, J., Huckwale, M. (eds.) Improvements in Speech Synthesis, pp. 328–338. John Wiley & Sons, Ltd., Chichester (2002)
Traber, C.: SVOX: The Implementation of a Text-to-Speech System for German. Ph.D. thesis, Swiss Federal Institute of Technology, Zurich (1995)
Touretzky, D.S., Mozer, M.C., Hasselmo, M.E.: Advances in Neural Information Processing Systems, vol. (8), pp. 225–231. MIT Press, Cambridge (1996)
Krušina, P.: Bang on-line documentation, http://bang2.sf.net/
Hanika, J., Horák, P.: Depedences and Independences of Text-to-Speech. In: Wodarz, H.-W. (ed.) Forum Phoneticum 69, Frankfurt Am Main, pp. 27–40 (2000)
Strecha, G.: Multilingual Labelling of Natural Speech Signals Basing on Synthetic Reference Signals., Dresden, Dresden University of Technology, Master thesis (2000) (in German)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2004 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Horák, P., Adámek, J., Sobe, D. (2004). Using Neural Networks to Model Prosody in Czech TTS System Epos. In: Sojka, P., Kopeček, I., Pala, K. (eds) Text, Speech and Dialogue. TSD 2004. Lecture Notes in Computer Science(), vol 3206. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-30120-2_43
Download citation
DOI: https://doi.org/10.1007/978-3-540-30120-2_43
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-23049-6
Online ISBN: 978-3-540-30120-2
eBook Packages: Springer Book Archive