Deblurring Acoustic Signals for Optimum Perception

Authors

  • Viktoria Taroudaki Department of Mathematics, Eastern Washington University, USA
  • Costas Smaragdakis Department of Mathematics and Applied Mathematics University of Crete, Greece and Institute of Applied and Computational Mathematics, Foundation for Research and Technology-Hellas, Greece
  • Michael I. Taroudakis Department of Mathematics and Applied Mathematics University of Crete, Greece and Institute of Applied and Computational Mathematics, Foundation for Research and Technology-Hellas, Greece

DOI:

https://doi.org/10.14738/assrj.911.13417

Keywords:

Signal processing, Blurring of acoustic signals, Machine learning

Abstract

This work presents a preliminary study on the effect that a blurring mechanism has on the quality of acoustic signals. The mechanism is mathematically modeled by means of a matrix that distorts the clear (blur and noise-free) digitized signal. Then, a combined deblurring-denoising scheme, adopted from image processing, is introduced and is shown to be very efficient for the simultaneous deblurring and denoising of the acoustic signal in cases where the blurring mechanism expressed through the corresponding matrix is known. Approximate matrices close to the actual one may be proven adequate for specific applications but it is important that the type of the matrix is approximated as much as possible. Comparison of raw and clear signals is made using a statistical characterization scheme, which is also appropriate for cases where the signal features have to be exploited for applications involving machine-learning. The deblurring of acoustic signals improve their quality and enhance their perceptibility, in favor of their exploitation for a wide range of applications, including human communication.

Downloads

Published

2022-11-25

How to Cite

Taroudaki, V., Smaragdakis, C., & Taroudakis, M. I. (2022). Deblurring Acoustic Signals for Optimum Perception. Advances in Social Sciences Research Journal, 9(11), 221–242. https://doi.org/10.14738/assrj.911.13417