To simplify the mathematics involved and to facilitate derivations of intermodulation and crosstalk formulas, the Shimbo approach uses a double Fourier transformation and represents the characteristics of a memoryless nonlinear device by a Bessel function series. Such an approach seems to be a very brave approximation as the approximated characteristics vary wildly with respect to the number of Bessel function termsLand with the amount of actual characteristic data used(A_{F}). However, it is shown here that the approach works well for the case involving two carriers with equal constant amplitudes, provided that one choosesLand AFproperly.

There are a few errors, a couple of poor suggestions, and several assumptions made without being stated in the above paper by Shimbo and Pontano for crosstalk computation. The method does not seem to work when the two carriers (for a two carrier case) are not at equal power levels. Also, better approximations of TWTA characteristics do not seem to produce more reliable results.