The tone AC signal has to be converted to DC power supply. A higher frequency means smaller components needed to do the conversion. Size roughly inverse proportional to frequency. So they probably chose close to the highest frequency possible on the audio output.

If the samples are put out at CD rates (44.1KHz) then 22.05KHz is the highest AC frequency available.

It's probably no coincidence it's right out of human hearing range. Most headphone drivers response curve are likely designed to be less efficient at that freq, and thus need more power to create the same loudness. the iPhone's audio driver probably takes this into consideration and drives those frequencies with more power to compensate.

Just a guess, I'm sure they arrived at it purely through trial and error, and were happy to find that it's also a frequency that will be minimally annoying if a headphone happens to be plugged in when the tone is played.

Minimally annoying my ass. If you are still young enough to hear it it's downright painful!

How young is young enough to hear 22 kHz? I am 25 and can barely hear 14kHz. http://www.noiseaddicts.com/2009/03/can-you-hear-this-hearin... In the comments section there people claim to be able to hear 20 and 21, but it seems like not many can hear 22.

In practice I'm pretty sure it's more about exposure to loud noises than actual age, and these increase as you age. I know I was able to hear 20kHz at 16, but probably down to 17kHz or so now at 25.

In other words, if you've been to a couple of rock concerts with no earplugs then all bets are off.

Also worth noting that most speakers play 0.5dB below the actual signal at 20000Hz, where anything above that usually isn't specified. So you need specialist audio gear or some measuring equipment to actually try to listen to really high frequencies.

Don't forget with tests like the one that you've linked it depends on your speaker's response to get an accurate reading.

I don't know the theory behind this stuff, but here it goes:

Headphone jacks tend to have a capacitor in-line with the output terminals to get rid of any DC bias -- this can cause problems like battery drain, lots of noise when you plug headphones in, and damage to the driver coil of headphones -- not good!

I like to think of a capacitor like a shock absorber, with the length of the shock absorber being like the voltage it's currently storing. If you push on one end of a shock absorber, the larger the force (current), the larger the rate of change of voltage (length), so this analogy might be useful.

Imagine holding on to one end of the shock absorber while someone else tries to move it back an forth at a given frequency. If they try to move it back and forth slowly, the maximum force (current) is going to be low, and so the total energy will be low and you won't get shaken much -- the shock absorber will just change length (voltage).

If they try to move it back and forth quickly, the force (current) will be much higher, and so you can extract a lot more energy from it.

I believe it's the same reason portable devices have poor bass response.

Wouldn't it be a simple relationship: higher frequency -> higher energy?