Filter banks allowing perfect reconstruction
Signal Processing
Multidimensional Digital Signal Processing
Multidimensional Digital Signal Processing
Hi-index | 0.00 |
The Euclidean algorithm, applied to zero-phase filters in the variable, w = (z + z-1)/2, is shown to provide a ladder structure realization of perfect reconstruction digital filter banks (DFB's) and wavelets. DFB matrices in the variable w are introduced and seen to have triangular factors corresponding to the ladder sections. Two-dimensional DFB's with diamond support are obtained from the McClellan transformation, and a related transform of a DFB in w to a two-dimensional DFB matrix is obtained. A design method based upon augmenting a given DFB with ladder sections is described. The application to the decomposition of biorthogonal wavelets is illustrated.