Linear phase FIR paraunitary filter banks and higher multiplicity wavelet given several analysis filters

The theory and design of paraunitary filter banks where several analysis filters are given is considered in this paper. Factorization based algorithm is derived for paraunitary completion problem with linear and nonlinear phase filters and arbitrary number of channels. A time domain algorithm is derived which formulates the problem in quadratic form for computational efficiency. Both algorithms are shown to completely characterize all embedding solutions. The existence of orthogonal wavelet with higher multiplicity and with arbitrary number of channels is considered and design example is presented