M. Lang
     Department of Electrical and Computer Engineering - MS366
              Rice University, Houston, TX-77251

                             J. Bamberger
                           

                               ABSTRACT

  We examine the problem of approximating a complex frequency response 
by a real-valued FIR filter according to the L_2 norm subject to 
additional inequality constraints for the complex error function. 

  Starting with the Kuhn-Tucker optimality conditions  which specialize
to a system of nonlinear equations we deduce an iterative algorithm. 
These equations are solved by Newton's method in every iteration step.
The algorithm allows arbitrary tradeoffs between an L_2  and an L_oo
design. The L_2  and the L_oo solution result as special cases.