In 1951, Licklider first suggested autocorrelation as a method to model pitch perception in his Duplex and Triplex Theories. It was not until 1991 that a full-fledged autocorrelation model was developed by Meddis and Hewitt. Over the past decades several investigators have refined autocorrelation or autocorrelation-like models to account for pitch perception and processing. While there have been a few arguments against autocorrelation for explaining pitch phenomena, autocorrelationlike models have been very successful in accounting for a wide range of data including those associated with the pitch of iterated ripple noise. This presentation will review the development of autocorrelation models of pitch perception, especially in accounting for the pitch and pitch strength of iterated ripple noise. The presentation will also cover how refining well-established models or theories may offer a good opportunity for developing skills to build additional models and theories. [Work supported by a grant from NIDCD.]