Measurements of double-click-evoked otoacoustics emissions (2CEOAEs) and double-chirp distortion products (2ChDPs) are reported for normal-hearing adults based upon theory presented in an earlier report [Keefe, J. Acoust. Soc. Am.103, 3489–3498 (1998)]. The nonlinear acoustic response of a probe assembly used in ear-canal measurements is tested in a calibration cavity to compare the double-evoked (2E) technique with existing OAE techniques. The 2E technique reduces the peak distortion by approximately 30 dB relative to existing click-evoked techniques. The 2E subtraction of click responses is partially analogous to current techniques in that the linear response is eliminated, but differs in that high-frequency measurements are improved by eliminating time gating of the cochlear response, and low-frequency measurements are improved by reducing probe distortion, especially when two acoustic sources are used. Because time gating is eliminated, it is straightforward to measure the onset of a click-evoked OAE. The nonlinear coherence function is used to measure the nonlinear distortion signal-to-noise ratio (DNR) for the 2ChDPs and 2CEOAEs. The DNR is typically 20–30 dB. An intermittent noise rejection technique is implemented in real time that compares a currently acquired ear-canal response with a stored response. Dissimilar responses indicate the presence of intermittent noise, and the noise-contaminated responses are thereby discarded before ensemble averaging.