Abstract
It is shown that the best ASR results are attained when a pre-processing is carried out synchronically with pitch. Specifically, an analysis step has to be equal to the current one-quasiperiod duration and current analysis intervals have to consist of an entire number of quasiperiods with total 45–60 ms duration. Quasi-periodicity and non-qusiperiodicity models and measures as well as their applications for the optimal segmentation of speech signals into one-quasiperiods are given and discussed. Then the ways to embed these pre-processing results into the recognition procedure are described.
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
T.K. Vintsiuk. Analysis, Recognition and Understanding of Speech Signals.-Kyjiv: Naukova Dumka, 1987, pp. 264, in Russian.
T.K. Vintsiuk, L.I. Khomenok. Influence of Discretisation Effects, Analysis Interval Length and Step to Speech Signals Recognition Accuracy.-Proceedings, 15 All-Union Seminar ldAutomatic Recognition of Sound Patterns”, Tallinn, 1989, pp 81–82, in Russian.
Taras K. Vintsiuk. Optimal Joint Procedure for Current Pitch Period Discrimination and Speech Signal Partition into Quasi-Periodic and Non-Periodic Segments.-In ldText, Speech, Dialogue”, Proc. of the First Workshop on Text, Speech, Dialog-TSD’98, Brno, 1998, pp 135–140.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 1999 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Vintsiuk, T.K., Sazhok, M.M. (1999). Speech Analysis and Recognition Synchronised by One-Quasiperiodical Segmentation. In: Matousek, V., Mautner, P., Ocelíková, J., Sojka, P. (eds) Text, Speech and Dialogue. TSD 1999. Lecture Notes in Computer Science(), vol 1692. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-48239-3_32
Download citation
DOI: https://doi.org/10.1007/3-540-48239-3_32
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-66494-9
Online ISBN: 978-3-540-48239-0
eBook Packages: Springer Book Archive