Abstract
We present a sequence of experiments with one–class classification, aimed at examining the ability of such a classifier to detect spectral smoothness of units, as an alternative to heuristics–based measures used within unit selection speech synthesizers. A set of spectral feature distances was computed between neighbouring frames in natural speech recordings, i.e. those representing natural joins, from which the per–vowel classifier was trained. In total, three types of classifiers were examined for distances computed from several different signal parametrizations. For the evaluation, the trained classifiers were tested against smooth or discontinuous joins as they were perceived by human listeners in the ad–hoc listening test designed for this purpose.
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
References
Bellegarda, J.R.: A novel discontinuity metric for unit selection text-to-speech synthesis. In: Proceedings of the 5th Speech Synthesis Workshop (SSW5), pp. 133–138. Pittsburgh, PA, USA (2004)
Grubbs, F.E.: Procedures for detecting outlying observations in samples. Technometrics 11, 1–21 (1969)
Hanzlíček, Z., Matoušek, J., Tihelka, D.: Experiments on reducing footprint of unit selection TTS system. In: Habernal, I. (ed.) TSD 2013. LNCS, vol. 8082, pp. 249–256. Springer, Heidelberg (2013)
Karabetsos, S., Tsiakoulis, P., Chalamandaris, A., Raptis, S.: One-class classification for spectral join cost calculation in unit selection speech synthesis. IEEE Signal Process. Lett. 17(8), 746–749 (2010)
King, S.: Measuring a decade of progress in text-to-speech. Loquens 1(1), e006 (2014)
Klabbers, E., Veldhuis, R.N.J.: Reducing audible spectral discontinuities. IEEE Trans. Speech Audio Process. 9(1), 39–51 (2001)
Legát, M., Matoušek, J.: Analysis of data collected in listening tests for the purpose of evaluation of concatenation cost functions. In: Habernal, I., Matoušek, V. (eds.) TSD 2011. LNCS, vol. 6836, pp. 33–40. Springer, Heidelberg (2011)
Legát, M., Matoušek, J., Tihelka, D.: On the detection of pitch marks using a robust multi-phase algorithm. Speech Commun. 53(4), 552–566 (2011)
Legát, M., Tihelka, D., Matoušek, J.: Pitch marks at peaks or valleys? In: Matoušek, V., Mautner, P. (eds.) TSD 2007. LNCS (LNAI), vol. 4629, pp. 502–507. Springer, Heidelberg (2007)
Markou, M., Singh, S.: Novelty detection: a review-part 1: statistical approaches. Signal Process. 83(12), 2481–2497 (2003)
Matoušek, J., Romportl, J.: On building phonetically and prosodically rich speech corpus for text-to-speech synthesis. In: Proceeding of the 2nd IASTED International Conference on Computational Intelligence, pp. 442–447. ACTA Press, San Francisco (2006)
Matoušek, J., Tihelka, D.: Voting detector: A combination of anomaly detectors to reveal annotation errors in TTS corpora. Submitted to the Interspeech (2016)
Matoušek, J., Tihelka, D.: Anomaly-based annotation errors detection in TTS corpora. In: Proceedings of the 16th Annual Conference of the International Speech Communication Association (Interspeech 2015), pp. 314–318. Dresden, Germany (2015)
Matoušek, J., Tihelka, D., Romportl, J.: Building of a speech corpus optimised for unit selection TTS synthesis. In: Proceedings of 6th International Conference on Language Resources and Evaluation, LREC 2008. ELRA (2008)
Pantazis, Y., Stylianou, Y.: On the detection of discontinuities in concatenative speech synthesis. In: Stylianou, Y., Faundez-Zanuy, M., Esposito, A. (eds.) COST 277. LNCS, vol. 4391, pp. 89–100. Springer, Heidelberg (2007)
Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., Blondel, M., Prettenhofer, P., Weiss, R., Dubourg, V., Vanderplas, J., Passos, A., Cournapeau, D., Brucher, M., Perrot, M., Duchesnay, E.: Scikit-learn: machine learning in Python. J. Mach. Learn. Res. 12, 2825–2830 (2011)
Přibil, J., Přibilová, A.: Evaluation of influence of spectral and prosodic features on GMM classification of Czech and Slovak emotional speech. EURASIP J. Audio Speech Music Process. 33(3), 1–22 (2013)
Schölkopf, B., Platt, J.C., Shawe-Taylor, J.C., Smola, A.J., Williamson, R.C.: Estimating the support of a high-dimensional distribution. Neural Comput. 13(7), 1443–1471 (2001)
Stylianou, Y., Syrdal, A.K.: Perceptual and objective detection of discontinuities in concatenative speech synthesis. In: Proceedings of the IEEE Acoustics, Speech, and Signal Processing (ICASSP), pp. 837–840 (2001)
Syrdal, A.K., Conkie, A.D.: Data-driven perceptually based join costs. In: Proceedings of the 5th Speech Synthesis Workshop (SSW5), pp. 49–54. Pittsburgh, PA, USA (2004)
Tax, D.M.J.: One-class classification: concept learning in the absence of counter-examples. Ph.D. thesis, Technische Universiteit Delft (2001)
Tihelka, D., Grůber, M., Matoušek, J., Jůzová, M.: Examining the ability of one-class classifier to ensure the spectral smoothness of concatenated units. Submitted to the 13th IEEE International Conference on Signal Processing (ICSP) 2016. If not accepted, the paper will be placed to github, under ARTIC-TTS-experiments/2016_SPECOM/ repository where the experiment data are
Vepa, J.: Join cost for unit selection speech synthesis. Ph.D. thesis, The University of Edinburgh, College of Science and Engineering, School of Informatics (2004)
Vepa, J., King, S.: Kalman-filter based join cost for unit-selection speech synthesis. In: Proceedings of the EUROSPEECH 2003 - INTERSPEECH 2003. Proceedings of 8th European Conference on Speech Communication and Technology, pp. 293–296. ISCA (2003)
Vít, J., Matoušek, J.: Concatenation artifact detection trained from listeners evaluations. In: Habernal, I. (ed.) TSD 2013. LNCS, vol. 8082, pp. 169–176. Springer, Heidelberg (2013)
Acknowledgments
This work was supported by the Grant Agency of the Czech Republic, project No. GA16-04420S and by the grant of the University of West Bohemia, project No. SGS-2016-039. Computational resources were provided by the CESNET LM2015042 under the program “Projects of Large Research, Development, and Innovations Infrastructures”.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2016 Springer International Publishing Switzerland
About this paper
Cite this paper
Tihelka, D., Grůber, M., Jůzová, M. (2016). Experiments with One–Class Classifier as a Predictor of Spectral Discontinuities in Unit Concatenation. In: Ronzhin, A., Potapova, R., Németh, G. (eds) Speech and Computer. SPECOM 2016. Lecture Notes in Computer Science(), vol 9811. Springer, Cham. https://doi.org/10.1007/978-3-319-43958-7_35
Download citation
DOI: https://doi.org/10.1007/978-3-319-43958-7_35
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-43957-0
Online ISBN: 978-3-319-43958-7
eBook Packages: Computer ScienceComputer Science (R0)