DIGITAL INTERPOLATION BASED ON FUNCTIONAL ITERATION.

Using functional iteration to generate digital interpolators with any desired accuracy from very approximate building blocks or 'seeds' are analyzed. The objective is to find a computationally simple seed and iteration formula that converge fast enough to result in computationally efficient filters. The iteration is carried out as many times as necessary to satisfy the given design specifications. An important aspect of this approach is the use of the iterated structure to implement the filter without expanding it into equivalent, but computationally undesirable, forms.