Minimum bandwidth interpolation of real discrete data

A DFT‐spectrum‐preserving transformation from one finite set of real numbers to a larger set of real numbers is developed. This transformation is completely in the time domain. Its convolution kernel, psinc(x;P), is analogous to the sinc(x) function familiar in Fourier transform work with continuous waveforms. The application to generation of periodic signals with specified bandlimited spectra is illustrated, as is interpolation using a mixture of time‐domain and radix‐2 fast Fourier transform (FFT) processing.