I have extended the spectral division method (SDM) in a recent paper presented at the ISCCSP 2010 for the reproduction of focused sources. The mathematical analysis revealed some quite interesting facts about the role of evanescent contributions in this context. If one aims at modelling the field of an acoustic point source in front of a loudspeaker array including its evanescent contributions one ends up with evanescent contributions with excessive levels before the focus point. I have created two videos for my talk at the ISCCSP 2010 in order to illustrate the problem.
The first video shows the reproduction of a focused source using the SDM when including the evanescent contributions in the desired field:
The background for this effect is the repidly decaing evanescent contributions of the secondary sources. These have to be driven at excessive levels in order to produce the desired evanescent contributions at the focus point.
The second video shows the same situation when discarding the evanescent parts in the model of the desired wave field: