Few have the time or inclination to read the standards that make up the implementing measures of electronics legislation.

It appears from the EN 61000 series of standards that the maximum permitted audio apparatus frequency is 40kHz - without it actually saying so. It is because of the 4KV ESD test, and that the only way to protect an input transistor from destruction is a low-pass filter. In fact, the filter is actually written into the implementing measure as a hint.

The actual low-pass frequency is higher than 40kHz, but when factoring in the EN 61938 matching system, the permitted output impedance of the previous stage being 2k2, then 40kHz is the result.

These standards are called harmonised - they "sing" in harmony - and it's as if a standard design was agreed on, and then the legislation has been designed around it.

Even though I think it sensible to have some sort of input filter, to apply it to every item in a system of multiple separates leads to a problem - that of phase. Two items in the chain might not introduce sufficient phase shift to be noticeable, but three or more items in a system will bring the phase frequency low enough to have a noticeable affect on sound quality.

The amplifier author, Doug Self, argues that such input filters are a source of distortion and should not be used. The harmonised standards make it plain that whatever Mr Self's distortion is, the consumer shall put up with it.

The calculations which reveal the filter's turnover frequency will (hopefully) appear in the "Conclusion-P" power amplifier construction manual, that I'm in the midst of writing.