Signal reconstruction using the Wold decomposition
by
Iulia Borza
West University of Timisoara
Coauthors: Flavius Turcu, Mohamed Najim

We propose a new criterion for an optimal reconstruction of an audio signal based on Wold-like decomposition of stochastic processes. This decomposition in the 1D case implies two orthogonal parts, which can be modelled respectively by an autoregressive model and a harmonic model. The problem which we attempt to answer is the identification and the separation of the two parts, by a new criterion which combines the quality and the parsimony of the parametric representations.

Date received: June 14, 2008