SUBSPACE APPROACH FOR THE DESIGN OF COSINE-MODULATED FILTER BANKS WITH LINEAR-PHASE PROTOTYPE FILTER

In this correspondence, a method for designing perfect reconstruction (PR) prototypes for paraunitary cosine-modulated filter banks is prese nted. The design procedure is based on a subspace approach that allows linear combinations of even-length linear-phase PR prototype filters in such a way that the resulting filter is also a linear-phase PR prot otype. Within a given subspace, the weights of the optimal Linear comb ination can easily be computed via an eigenanalysis. The filter design is carried out iteratively while the PR property is guaranteed throug hout the design process. No nonlinear optimization routine is needed. As a special case, the proposed approach allows the design of discrete -coefficient prototypes, which are of great interest for efficient har dware implementations.