Application of the narmax method to the modelling of the nonlinearity of dynamic loudspeakers

The application of the NARMAX method to the modelling of
the nonlinearity of dynamic loudspeakers is described. The principle of creating
a polynomial representation of a model, the problems stemming from a too large
number of model coefficients and the method of optimizing the model are
presented. The method was tested on data from actual loudspeaker measurements.
Different models are compared as regards their accuracy depending on the
modelling parameters. Finally, the model characteristics are compared with the
results of loudspeaker measurements performed by classical methods.