Project detail

Non-linear Methods of Speech Enhancement

Duration: 01.01.2007 — 31.12.2009

Funding resources

Czech Science Foundation - Standardní projekty

- whole funder (2006-03-28 - 2009-12-31)

On the project

V moderních telekomunikačních systémech nasazených v pevných a mobilních sítích je nutné stále řešit základní problém odstranění parazitního rušení z pozadí řečového signálu a zlepšit poměr odstupu signálu od šumu při zachování nebo zlepšení srozumitelnosti výsledné řeči. Cílem projektu je pokračovat ve vývoji moderních metod pro extrakci řeči ze šumu a soustředit se na nelineární metody, které budou účinněji potlačovat zvláště nestacionární rušivé signály. Metody budou založeny zvláště na respektování maskovacích vlastností lidského sluchu, na moderních nelineárních metodách zpracování signálu včetně matematické statistiky (modifikované AR modely apod.), waveletové transformace, nerovnoměrně členěných bank filtrů a transformace typu time-scale.

Description in English
In modern telecommunication systems deployed in fixed and mobile networks it is all the time necessary to solve the basic problem of removing parasitic interference from speech signal background and to improve the signal-to-noise ratio while maintaining or enhancing the intelligibility of initial speech. The aim of the project is to follow up on the results obtained in the development of methods for extracting speech from noise, and to concentrate on the design of a group of non-linear methods that will suppress more effectively non-stationary interference signals. The methods will be based on respecting the masking properties of the human ear, on modern non-linear methods of digital signal processing, inclusive of mathematical statistics (the AR or ARMA models, etc.), the wavelet transform, non-uniformly divided filter banks, transformation of the type of scale, etc.

Keywords
Číslicové zpracování signálů, nelineární metody, zvýrazňování řeči

Key words in English
Digital signal processing, non-linear methods, speech enhancement

Mark

GA102/07/1303

Default language

Czech

People responsible

Rajmic Pavel, prof. Mgr., Ph.D. - fellow researcher
Stejskal Vojtěch, Ing., Ph.D. - fellow researcher
Sysel Petr, doc. Ing., Ph.D. - fellow researcher
Vích Robert, Ing., DrSc. - fellow researcher
Smékal Zdeněk, prof. Ing., CSc. - principal person responsible

Units

Department of Telecommunications
- beneficiary (2006-03-28 - not assigned)

Results

PŘINOSIL, J.; MÍČA, I. Supporting System for Digital Signal Processing. In Research in Telecommunication Technology RTT - 2007. Liptovský Ján, Slovenská republika: 2007. p. 341-344. ISBN: 978-80-8070-735-4.
Detail

RAJMIC, P. The Algorithm of Segmented Wavelet Transform. Poster na mezinárodní konferenci Trends in Harmonic Analysis. Trends in Harmonic Analysis. Strobl, Austria: 2007. p. 1 ( p.)
Detail

SMÉKAL, Z.; SYSEL, P. Speech Enhancement bz Spectral Subtraction Method using Enhanced Noise Estimate. In Proceedings of the 30th International Conference on Telecommunications and Signal Processing, TSP2007. 1.9.2007. Brno: VUT Brno, 2007. p. 89-92. ISBN: 978-80-214-3445-5.
Detail

SMÉKAL, Z. Semi-symbolic CAD Discrete System Analysis and Signal Flow Graphs. In Proceedings of the 30th International Conference on Telecommunicationsa and Signal Processing TSP 2007. 1.9.2007. ÚTKO: VUT Brno, 2007. p. 93-96. ISBN: 978-80-214-3445-5.
Detail

VRÁNA, J. Analysis of Transfer Functions in Quadrature Mirror Filter Bank with Sigma-Delta Modulators. In Telecommunications and Signal Processing TSP - 2007. Brno: SEI-UTKO, 2007. p. 59-62. ISBN: 978-80-214-3445-5.
Detail

SMÉKAL, Z.; SYSEL, P. Estimation of Threshold of Power Spectral Density of Noise in Single-Channel Methods of Speech Enhancement. In Proceedings of the 8th International Conference on Research in Telecommunication Technology RTT 2007. první. Liptovský Ján, Slovenská republika: Žilinská univerzita, 2007. p. 421-424. ISBN: 978-80-8070-735-4.
Detail

SMÉKAL, Z.; VÍCH, R. SNR-based Assessment of Quality of Speech Enhancement using Single-Channel Methods. In Studientexte zur Sprachkommujnikation, Elektronische Signalverarbeitung. Band 46. Germany: TUD Press, 2007. p. 149-154. ISBN: 978-3-940046-40-6.
Detail

VRÁNA, J. Quadrature Mirror Filter Bank with Noise Shaping. In TRANSCOM 2007, 7-th European Conference of Young Research and Science Workers. Žilina: EIDS-Žilina University Publisher, 2007. p. 211-214. ISBN: 978-80-8070-694-4.
Detail

MALÝ, J.; RAJMIC, P. Watermarking Schemes for DWT-based Image Codec. In Proceedings of the International Conference on Computer Science and Information Technologies. Lviv, Ukraine: Lviv Polytechnic National University, 2007. p. 202-205. ISBN: 978-966-2191-06-6.
Detail

VLACH, J.; RAJMIC, P.; PŘINOSIL, J.; VYORAL, J.; MÍČA, I. New algorithm of discrete wavelet transform optimized to real-time digital signal processing. In Proceedings of the 12th IFIP International Conference, Prague. 2007. p. 514-520. ISBN: 978-0-387-74158-1.
Detail

VLACH, J. Discrete wavelet transform frame-by-frame in signal processing. In Research in Telecommunication Technology RTT - 2007. Žilina: 2007. p. 1-5. ISBN: 978-80-8070-735-4.
Detail

RAJMIC, P.; MALÝ, J. Boundary Effects in the Wavelet Transform of Finite Discrete Signals. In Proceedings of the 30th International Conference Telecommunications and Signal Processing TSP-2007. Brno: VUT v Brně, 2007. p. 81 ( p.)ISBN: 978-80-214-3445-5.
Detail

RAJMIC, P.; VLACH, J. Real-time Audio Processing via Segmented Wavelet Transform. In Proceedings of the 10th International Conference on Digital Audio Effects (DAFx-07). Bordeaux: University of Bordeaux 1, 2007. p. 55-58. ISBN: 978-88-901479-1-3.
Detail

MALÝ, J.; RAJMIC, P. DWT-SPIHT Image Codec Implementation. In Technical Computing Prague 2007 15th Annual Conference Proceedings. Praha: Humusoft s.r.o., 2007. p. 90 ( p.)ISBN: 978-80-7080-658-6.
Detail

RAJMIC, P. Method of Forward Segmented Wavelet Transform. In Summer School DATASTAT 06, Proceedings. Brno: Masaryk University, 2007. p. 235 ( p.)ISBN: 978-80-210-4493-7.
Detail

VONDRA, M. Speech Enhancement using Cepstral Vocoder. In 17th Czech-German Workshop Speech Processing. 1. Praha: UFE AV ČR, 2007. s. 110-114. ISBN: 978-80-86269-00-9.
Detail

MAČÁK, J.; SCHIMMEL, J. Nonlinear Circuit Simulation Using Time-Variant filter. In Proceedings of the 12th International Conference on Digital Audio Effects DAFx09. Como, Italy: Politecnico di Milano, 2009. p. 1-5.
Detail

HANÁK, P.; SYSEL, P.; VRBA, K. Verification of Modular Design of VoIP PBX with Analogue Telephone Lines. In Proceedings of the 32nd International Conference on Telecommunications and Signal Processing - TSP' 2009. Budapest: Asszisztencia Szervezo Kft., 2009. p. 1-4. ISBN: 978-963-06-7716-5.
Detail

RAJMIC, P.; VLACH, J.; VYORAL, J.; PRŮŠA, Z. Real-Time Wavelet Transform with Overlap of Signal Segments. In ICSES 2008 International Conference on Signals and Electronic Systems Proceedings. Krakov: AGH University of Science and Technology, 2008. p. 39-42. ISBN: 978-83-88309-47-2.
Detail

PŘINOSIL, J.; SMÉKAL, Z. Robust Real-Time Face Tracking System. In 32nd International Conference on Telecommunications and Signal Processing - TSP'2009. Dunakiliti: ASSZISZTENCIA Congress Bureau, 2009. p. 1-4. ISBN: 978-963-06-7716-5.
Detail

ATASSI, H. Metody detekce základního tónu řeči. Elektrorevue - Internetový časopis (http://www.elektrorevue.cz), 2008, roč. 2008, č. 4, s. 1-17. ISSN: 1213-1539.
Detail

MALÝ, J.; RAJMIC, P. Fast lifting wavelet transform and its implementation in Java. Mobile and Wireless Communication Networks, 2007, vol. 2007, no. 9, p. 488-496. ISSN: 1571-5736.
Detail

SMÉKAL, Z.; SYSEL, P. Enhanced estimation of power spectral density of noise, Springer Verlag. Personal Wireless Communications, 2007, vol. 245, no. 2, p. 521-532. ISSN: 1861-2288.
Detail

SYSEL, P. Jednokanálová metoda zvýraznění řeči s využitím waveletové transformace v kmitočtové oblasti. Vědecké spisy Vysokého učení technického v Brně Edice PhD Thesis, 2007, roč. 2007, č. 425, s. 1-26. ISSN: 1213-4198.
Detail

VLACH, J.; PŘINOSIL, J. Lokalizace obličeje v obraze s komplexním pozadím. Elektrorevue - Internetový časopis (http://www.elektrorevue.cz), 2007, roč. 2007, č. 04, s. 1 ( s.)ISSN: 1213-1539.
Detail

ATASSI, H.; ESPOSITO, A. A speaker Independent Approach for Emotion Recognition. Tools with artificial inteligence, 2008, vol. 2008, no. 1, p. 147-152. ISSN: 1082-3409.
Detail

RAJMIC, P. Algorithms for Segmentwise Computation of Forward and Inverse Discrete-time Wavelet Transform. Journal of Concrete and Applicable Mathematics, 2010, vol. 8, no. 3, p. 1-15. ISSN: 1548-5390.
Detail

SMÉKAL, Z.; SYSEL, P. Single-Channel Noise Suppression by Wavelets in Spectral Domain. Lecture Notes in Computer Science, 2007, vol. 4775, p. 150-164. ISSN: 0302-9743.
Detail

SMÉKAL, Z.; ČERMÁK, J. Undertermined Blind Source Separation Using Linear Separation System. Lecture Notes in Computer Science, 2009, vol. 2009, no. 5398, p. 300-305. ISSN: 0302-9743.
Detail

SMÉKAL, Z.; ATASSI, H.; STEJSKAL, V.; MEKYSKA, J. Soubor programů pro práci se skrytými Markovovými modely (HTK). Elektrorevue - Internetový časopis (http://www.elektrorevue.cz), 2009, roč. 2009, č. 11, s. 11-1 (11-42 s.)ISSN: 1213-1539.
Detail

PŘINOSIL, J. Přesné sledování pohybu tváří v reálném čase. Elektrorevue - Internetový časopis (http://www.elektrorevue.cz), 2009, roč. 2009, č. 56, s. 1-5. ISSN: 1213-1539.
Detail

SYSEL, P.; MIŠUREC, J. Estimation of Power Spectral Density using Wavelet Thresholding. WSEAS Applied Informatics & Communications, 2008, vol. 2008, no. 1, p. 207-211. ISSN: 1790-5117.
Detail

SMÉKAL, Z.; SYSEL, P. Enhanced Estimation of Power Spectral Density of Noise using the Wavelet Transform. In Personal Wireless Communications. SSCS. USA: Springer, 2007. p. 521-532. ISBN: 978-0-387-74159-8.
Detail

MALÝ, J.; RAJMIC, P. Performance of discrete wavelet transform lifting scheme approach. In Proceedings of the 12th IFIP International Conference. 2007. p. 231 ( p.)ISBN: 978-0-387-74158-1.
Detail

Responsibility: Smékal Zdeněk, prof. Ing., CSc.