In this paper, a Period-Synchronous Wavelet Transform (PSWT), based on the concept of the Multiplexed Wavelet Transform (MWT), is proposed for periodic acoustic signal estimation. In PSWT decomposition, the analyzed signal is reconstructed in terms of a sum of scaling residue component and different scale wavelet partials. The scaling residue component consists mainly of harmonics of the periodic acoustic signal. After PSWT processing, the scaling residue component is passed through the Fourier Linear Combiner (FLC) algorithm to produce an adaptive estimate of each harmonic using the least-mean-square (LMS) algorithm.
展开▼