This letter presents a computational technique which allows one to resolve spectral amplitudes at frequencies spaced more closely than the inverse of data record length. This then alleviates problems often encountered in attempts at spectral analysis of low-frequency data.