Microwave hearing is a special case within the verified acute effects of radiofrequency fields. It occurs in fields with a range from 200 MHz to 6.5 GHz with short strong pulses of 30 microseconds or less and long intervals between these pulses (as typically used in radar signals). The perceptions are described as humming, clicking or crackling sounds. The effects have been attributed to thermoelastic waves, which are generated in the brain by absorption of the pulse energy, and stimulate the inner ear mechanically in the hearable frequency range.
The threshold for perception depends on the energy flux density per pulse and not on the time-averaged power flux density of the overall signal. At 2.45 GHz with pulse durations of less than 30 µs, which corresponds to a specific absorption (SA) of 4-16 mJ/kg, the threshold for perception lies at an energy density of the exposure of 100-400 mJ/m². The SA is used for the assessment of exposure to short-pulsed signals, like in radar, instead of the SAR. There are, however, indications that SAR peak values of 130-520 W/kg occur in the brain during such a 30 µs pulse with the aforementioned SA values (cf. ICNIRP, p.506 and p.513). The recommendations of the ICNIRP include special limits for pulsed exposure in the frequency range from 300 MHz to 10 GHz, with a specific absorption of 2 mJ/kg for the general public and 10 mJ/kg at workplaces (cf. chapter Basic restrictions). In the Reference levels, corresponding limits are set for the power density in the range of pulses. According to this, microwave hearing is only possible in close vicinity to powerful radar units (i.e. at the workplace) when ICNIRP limit values are complied. The Swiss Federal Office for the Environment BAFU (p. 91) specifies a lower perception threshold than the ICNIRP and remarks that microwave hearing can be expected even if the ICNIRP limit value is complied.
Broadcasting, TV and mobile communication signals, however, cannot cause microwave hearing as the pulse patterns of these signals cannot be compared with radar signals and the energy flux densities per pulse are too low (or because pulses are not even existent, in the case of analog radio signals). Therefore, there is no evidence for corresponding perceptions in the case of everyday exposure, for example, to mobile communication fields (cf. BAFU, p. 91).