Answers are grounded in this paper and cite their sources.
Near-to-far-field transformations require the tangential electric and magnetic fields over a surface, which we call the integration boundary. However, the staggered nature of the finite-difference time-domain grid is problematic in that the electric and magnetic fields are not collocated in either space or time. For harmonic transformations, i.e., ones which rely upon a Fourier transform of the time-domain near-fields, one can account for the temporal offset with a simple phase correction in the frequency domain. To account for spatial offsets, previously an arithmetic mean of the time-domain fields to either side of the integration boundary has been used. Here we show that superior results are obtained by instead using a geometric mean of the harmonic fields to either side of the integration boundary.
No takes yet. Share an insight, caveat, or question.
Robinson et al. (2007) studied this question.
Synapse has enriched 5 related papers. Consider them for comparative context: