Naslov (srp)

Primena talasića na otklanjanje šuma i parametrizaciju audio signala

Autor

Damnjanović, Đorđe, 1986-

Doprinosi

Ćirić, Dejan G., 1967-
Perić, Zoran
Šumarac-Pavlović, Dragana
Nikolić, Jelena
Milošević, Marina

Opis (srp)

Methods for signals analysis have been developing more andmore in last few decades. By discovering new ones or by upgradingthe existing algorithms, results of signal processing have beensignificantly improved, especially if the emphasis is on the process ofremoving noise from the signal or signal parameterization. A numberof analog and digital filters are present in many scientific fieldsnowadays, and a technique that is more and more in use is the onebased on wavelets. Wavelet decomposition to its detail andapproximation coefficients is the method that provides significantresults in removing noise from the signal as well as in signalparameterization.This thesis presents the results of applying wavelets for thereduction of noise and its negative effects in synthetized andmeasured room impulse responses. Analysis is done with signals infull frequency range (broadband signals) and signals filtered in octaveand third-octave bands. Different sets of wavelet parameters areanalyzed with the aim of finding the optimal set of parameters thatwill give the best results. White and pink noises are used in the wholeresearch observing their influences on room impulse responses andenergy decay curves generated from these responses. Two usedmeasures that quantify the achieved results are the dynamic rangeimprovement of the decay curves and noise floor reduction. Otherstandard methods for noise removal and reducing noise effects inroom impulse responses are used for the purpose of comparison withresults achieved with wavelets.Besides the method of noise removal, wavelets have alsofound their place in the knee detection in energy decay curvesobtained from room impulse responses. The obtained results are alsocompared with the standard methods from literature.An important task when it comes to newly manufacturedindustrial products, such as direct current motors, is to findprocedures and algorithms that will separate non-faulty and faultymotors. Method that can be used for this purpose is related toapplication of wavelets to audio signals containing sound generatedby analyzed motors. The detail coefficients obtained by the waveletdecomposition process representing wavelet-based audio features mayshow the mentioned differences between motors at certain levels ofdecomposition. The achieved results are confirmed by applying theintroduced numerical measure called feature difference of the detailcoefficients at certain levels.

Opis (srp)

Biografija autora: str. 165,Bibliografija: str. 154-164. Datum odbrane: 11.03.2022. Audio signal processing

Jezik

srpski

Datum

2021

Licenca

Ovo delo je licencirano pod uslovima licence
Creative Commons CC BY-NC-ND 3.0 AT - Creative Commons Autorstvo - Nekomercijalno - Bez prerada 3.0 Austria License.

http://creativecommons.org/licenses/by-nc-nd/3.0/at/legalcode