In this paper, we present a method to correct the amplitud distorsion produced by a non linear system. First of all it is shown how this correction can be achieved making use of a least square error criterium by means of a polynomical function. An adaptive method is also studied and a discussion of the convergence condition is made. Finally, as an example, the method is aplied to design the expansor which makes use of a mu law.

Peer Reviewed