Researchers usually utilize a few previously employed ground motions or simple sinusoidal input for earthquake simulations in centrifuge tests. Given the inherent randomness of earthquake motions as well as their importance in determining seismic response, it is beneficial to use a broad suite of motions. This paper describes the earthquake motion selection and filtering processes employed for a recent series of centrifuge tests. A simple procedure is proposed for calibrating recorded ground motions for use in the University of California at Davis Center for Geotechnical Modeling centrifuge. The procedure employs an iterative frequency-domain transfer function to reasonably match the motion achieved during the test with what is targeted by the researcher. Insights about the performance of the shaking table are also made.