10th International Conference on Heterogeneous Networking for Quality, Reliability, Security and Robustness

Research Article

Greek Folk Music Denoising Under a Symmetric α-stable Noise Assumption

Download367 downloads
  • @INPROCEEDINGS{10.4108/icst.qshine.2014.256282,
        author={Nikoletta Bassiou and Constantine Kotropoulos and Ioannis Pitas},
        title={Greek Folk Music Denoising Under a Symmetric α-stable Noise Assumption},
        proceedings={10th International Conference on Heterogeneous Networking for Quality, Reliability, Security and Robustness},
        publisher={IEEE},
        proceedings_a={QSHINE},
        year={2014},
        month={9},
        keywords={music denoising markov chain monte carlo overlap-and-add reconstruction greek folk music},
        doi={10.4108/icst.qshine.2014.256282}
    }
    
  • Nikoletta Bassiou
    Constantine Kotropoulos
    Ioannis Pitas
    Year: 2014
    Greek Folk Music Denoising Under a Symmetric α-stable Noise Assumption
    QSHINE
    IEEE
    DOI: 10.4108/icst.qshine.2014.256282
Nikoletta Bassiou1, Constantine Kotropoulos1,*, Ioannis Pitas1
  • 1: Aristotle University of Thessaloniki
*Contact email: costas@aiia.csd.auth.gr

Abstract

The noise in musical audio recordings is assumed to obey an alpha-stable distribution. A sparse linear regression framework with structured priors is elaborated. Markov Chain Monte Carlo is used to infer the clean music signal model and the alpha-stable noise distribution parameters. The musical audio recordings are processed both as a whole and in segments by using a sine-bell window for analysis and overlap-and-add reconstruction. Experiments on noisy Greek folk music excerpts demonstrate better denoising under the alpha-stable noise assumption than the Gaussian white noise one, and when processing is performed in segments rather than in full recordings.