Fir half-band filters satisfying prescribed amplitude and phase specifications in weighted least mean square sen

Abstract EN

This paper introduces a new method for designating half-band filters that approximate given amplitude and phase characteristics in a weighted least mean square sense.

The method is based on magazine the integrated square error for passband and stop band amplitude deviations while imposing the phase response to the required specifications.

Th ses filters have the advantage of providing the designer with the most optimal design in the sense of smallest filter length for a given set of specifications.

This makes the filter's group delay under control.

Consequently, the design of half-band filters with fractional sampling periods is possible.

Compared with existing methods, this technique is sim le and effective since the coefficients of the filter are obtained by solving a set of linear equations.

Moreover, the design algorithm can adapted to the design of half-band filters with discrete coefficients that have economical hardware implementations.

Numerical examples are presented to demonstrate the usefulness and flexibility of the design method