Evaluation of Nonlinearities in a Diode Clipper Circuit based on Wave Digital Filters

This thesis examines the method of virtual analog modeling known as wave digital filters, and specifically its application in adequately capturing the behavior of nonlinear audio circuits. An overview of various methods of virtual analog modeling are
presented, as well as other common approaches to nonlinear system identification. In order to evaluate the wave digital filters, we construct a Python library with various circuit elements and circuit classes. We focus on evaluating the approximation of a
diode clipper, a common nonlinear circuit used in audio signal processing. In the evaluation, we compare the frequency response of our wave digital model and SPICE simulations of the same circuit. We make such comparisons at several sample rates
and parameter changes. We further perform transient analysis and examine the harmonics added by the circuit and measure the amount of total harmonic distortion introduced, at multiple sample rates and parameter changes. We lastly process the clipper with a swept sine wave across all frequencies, which is convolved with an inverse filter of the sweep, and is then used to calculate a transfer function of the system that accounts for the nonlinear behavior.