While I was working in Ottawa, extensive searches of research publications for clues about likely transduction mechanisms revealed little that was useful apart from interesting papers on hearing tests, psychoacoustics, biophysical electrophonics (which is the perception of sound from direct electrical stimulation of areas near the ears) and reports of radio signals detected by tooth fillings. Apart from work on electrostatic loudspeakers, only one paper, by Sommer and von Gierke (1964), dealt with the direct human perception of electric fields varying at audio frequencies. They reported that large fields are required: several thousand volts per metre.
A subsequent visit to the Physics Department of the University of Western Ontario was more productive. The Head, Professor Parker Alford, expressed great interest in my progress thus far and encouraged me to make use of an anechoic chamber in his department for tests of the human perception of electric and magnetic fields varying at audio frequencies.
The magnetic field tests were inconclusive. The electric field results for the most part verified Sommer and von Giercke's findings, except for three of the volunteers who were markedly more sensitive than most, the best one able to detect electric field variations of only 160 volts peak-to-peak at 4 kHz frequency. The common factor proved to be their hair. Two females with the fashionable Afro hair styles and a male with very long soft hair showed the lowest threshold of sensitivity. Obviously their hair was acting as a transducer.
As well, there was a serendipitous discovery. Naturally I acted as the first test subject, and underwent the same test again as a check just prior to dismantling the equipment. My threshold for detection had risen 3 to 4 decibels! Luckily the answer was found: I was not wearing my glasses. When they were replaced my test results reverted to the same as they were initially. Clearly the glass frames were responding to the imposition of the varying electric field. This finding indicated that mundane objects in the immediate vicinity of observers may assist their perception of electrophonic sounds from bolides.
Later, in Newcastle, Australia, graduate student Trish Ostwald and I gained access to an anechoic chamber for tests of the transduction efficiencies of a variety of objects and common materials, including various types of vegetation. It was a brute force approach. We kept raising the voltage until an acoustic response was detected from the samples under test. This verified that there is an extremely wide variability of response and fully explained why whole groups of witnesses at appropriate locations may hear electrophonic bolide sounds and not at less suitable places.
These experiments amply explain the capriciousness of electrophonic sounds. One or two people in a group may hear the sounds while others do not. Or one entire group may report the sounds while other people in less favourable surroundings hear nothing. The experiments described above which have settled this conundrum are not difficult to perform yet I have found no evidence of them being performed by any other researchers.
The next question is the obvious one: how likely are you to hear an electrophonic bolide? That depends on the above factors as well as the rate of entry into the atmosphere of sufficiently large bolides.