Design of prototype filter of DFT filter banks with low implementation complexity

In this paper, we consider the design of FIR prototype filter of DFT filter banks, where the coefficients are expressed as sums of signed powers-of-two terms. The infinite precision solution can be obtained for given amplitude distortion or aliasing power. However, the finite precision solution may not be satisfied. This problem is formulated as a constrained optimization problem and can be transformed into an unconstrained integer programming problem. An efficient algorithm based on a discrete filled function is developed for solving the finite precision solution.