A variety of signal processing techniques can be used to minimize the effects of noise on linear, wideband, electromagnetic (EM) systems operating in the time-domain. All systems use repetitive waveforms with polarity reversal in alternate half-cycles. Exponential averaging or digital integration (stacking) is employed to increase signal-to-noise (S/N) ratios by limiting the noise acceptance to narrow frequency bands centered on odd harmonics of the repetition frequency, the width of the acceptance bands being inversely proportional to stacking time. For certain types of nonstationary noise (e.g., occasional transients) or coherent noise (e.g., powerlines) it is possible to increase S/N ratios above those obtained by simple stacking for an equal time by use of techniques such as pruning, tapered stacking or randomized stacking. With some system waveforms and when the noise spectrum is not 'white', use of pre-emphasis filtering in the transmitter and a corresponding de-emphasis filter in the receiver may significantly improve the input S/N ratio before stacking. Specific applications of the various techniques are discussed with reference to one particular time-domain EM system, the UTEM 3 system. By their use, improvements in S/N ratio of as much as 6 to 1 have been regularly achieved without any increase in transmitter power, depending on the nature of the local noise.