DocumentCode
3391261
Title
Analysis and enhancement of locally harmonic signals using adaptive multi-kernel methods
Author
Dubost, Stéphanie ; Cappe, Olivier
Author_Institution
Ecole Nat. Superieure des Telecommun., Paris, France
fYear
1999
fDate
1999
Firstpage
163
Lastpage
166
Abstract
This paper is concerned with the analysis/synthesis and enhancement of signals that can be efficiently modeled as “quasi-harmonic”. We first provide a formal definition of quasi-harmonic signals and show that known estimation methods for such signals can be interpreted in a non-parametric local approximation framework. We then investigate the possibility of using “adaptive multi-kernel” estimation methods for which the duration of the analysis frame is tuned independently for each harmonic depending on the local behavior of the signal
Keywords
Gaussian noise; adaptive estimation; adaptive signal processing; harmonic analysis; polynomial approximation; signal synthesis; speech processing; white noise; Gaussian white noise; adaptive estimation methods; adaptive multi-kernel methods; formal definition; local behavior; local polynomial approximation; locally harmonic signals analysis; locally harmonic signals enhancement; locally harmonic signals synthesis; nonparametric local approximation; quasi-harmonic signals; voiced speech; Frequency; Harmonic analysis; Integrated circuit modeling; Interpolation; Noise reduction; Parameter estimation; Signal analysis; Signal processing; Signal synthesis; Speech synthesis;
fLanguage
English
Publisher
ieee
Conference_Titel
Applications of Signal Processing to Audio and Acoustics, 1999 IEEE Workshop on
Conference_Location
New Paltz, NY
Print_ISBN
0-7803-5612-8
Type
conf
DOI
10.1109/ASPAA.1999.810875
Filename
810875
Link To Document