Though heterodyne detection provides a valuable technique for detecting small signals, the conventional system has several inherent disadvantages. In applications such as communications and radar, obtaining a reasonably high signal-to-noise ratio (SNR) requires (1) a good knowledge of the velocity of the transmitter or target, (2) a stable yet tunable local oscillator, and (3) a target or source that presents a minimum of frequency broadening. These conditions are frequently not adhered to by real systems, particularly in the ir and optical, giving rise to detection capabilities far below optimum. Calculations are presented for the use and operation of a three-frequency nonlinear heterodyne system that eliminates many of the stringent conditions required for conventional heterodyne detection, while maintaining its near-ideal SNR. The technique, which is similar in principle to heterodyne radiometry, makes use of a two-frequency transmitter and a nonlinear second detector and is particularly useful for signal acquisition; for signals of unknown Doppler shift, in fact, performance is generally superior to that of the conventional system. Although primary emphasis is on the ir and optical because of the large Doppler shifts encountered there, application of the principle in the microwave and radiowave is also discussed. For cw radar and analog communications, the SNR, power spectral density (PSD), and minimum detectable power (MDP) are obtained and compared with the standard configuration. Both sinewave and Gaussian input signals are treated. A variety of specific cases are discussed including the optimum performance case, the typical radar case, and the AM and FM communications case. Evaluation of the technique for pulsed radar and digital communications applications (both in the absence and in the presence of the lognormal atmospheric channel) is reserved for Part 2 of this paper [Appl. Opt. 14, 680 (1975)].