Engine noise separation through MCMC sampling in a hierarchical Bayesian context

Gianluigi Brogna (1) , Jérôme Antoni (1, 2) , Quentin Leclere (1) , Olivier Sauvage (3)

1 LVA - Laboratoire Vibrations Acoustique
2 CeLyA - centre Lyonnais d'Acoustique
3 PSA Peugeot Citroën

A noise separation algorithm is implemented to address the case of sources highly overlapping over time and frequency and observed through correlated references. The method is applied to simulated engine signals with the aim of separating the contributions due to different physical origins. The results is compared to the one provided by the Wiener separation method as developed in [1] and the advantages of the Bayesian approach is pointed out.

Domaines: Acoustique [physics.class-ph]

Gianluigi Brogna, Jérôme Antoni, Quentin Leclere, Olivier Sauvage. Engine noise separation through MCMC sampling in a hierarchical Bayesian context. Proceedings of ISMA 2016, 2016, Leuven, Belgium. ⟨hal-01468593⟩