An efficient FFT interpolation scheme, using the concept of multirate digital filtering, is presented in this paper. The method relies on a block representation and employs a parallel implementation. Timing comparisons show that this method is faster than the conventional FFT technique demonstrated by Gold and Rader. In addition, it permits a shorter word length, smaller memory size, and is well suited to seismic data processing. Application of this method to common-depth-point trace summation ('stacking') is illustrated.