Theoretical and Experimental Comparison of Three Methods for Compensation of Electrodynamic Transducer Nonlinearity

Based on a simplified nonlinear lumped-element model of an electrodynamic loudspeaker, nonlinear compensators are derived, simulated, and implemented on a digital signal processor (DSP). The model comprises three major nonlinearities: voice-coil excursion dependent force factor, (suspension) stiffness, and self-inductance. A Volterra series expansion is used to estimate nonlinear parameters from distortion measurements. The first compensation method utilizes the second-order s-domain kernel of this expansion to synthesize a second-order compensator. The other two methods employ extended mirror filter and state-space techniques, respectively. All three approaches are compared with respect to needed dynamic elements, computational complexity, robustness, and effectiveness for the case of a low-frequency direct radiator in a closed cabinet.