In short, ultrasonic frequencies will interact with each other in the room (reflections, reverb and so on) in a way that causes interference in the frequency range you can hear.

This is why it is important to record instruments with higher sample rates (192khz is standard) in order to fully preserve their sonic character for further processing.

This is one answer to the question "why do musicians record ultrasound," I hope it is correct. There are many answers given to this question and even a lot of sound engineers don't understand the concepts correctly, so hopefully I have got this right.

You also hear a lot of stuff about how filters need headroom to avoid creating aftefacts in the data; a LP filter creates more artefacts the sharper its cutoff curve, so therefore by having a lot of ultrasound one can have a very gradual cutoff (or a cheaper filter) and only discard data in the inaudible range. I'm less sure that this is true however.

Then there are a lot of people who think it's simply the usual "higher numbers are better" type of marketing games. They may well be right. All I know is, there is nowhere near enough double blind testing in the pro audio industry, and if they started doing in en-mass it would cause terrible trouble.

Hopefully, as usual, some expert will jump in here and lay down some solid facts. Be aware though that one cannot simply apply electromagnetic principles, it isn't always the same.

>In short, ultrasonic frequencies will interact with each other in the room (reflections, reverb and so on) in a way that causes interference in the frequency range you can hear.

If a vibrating membrane of skin in my head can react to it, then it is a wave traveling through air. Why wouldn't this hold for a microphone?

>If a vibrating membrane of skin in my head can react to it, then it is a wave traveling through air. Why wouldn't this hold for a microphone?

Microphones don't respond in the same way to the same frequencies that your eardrum does. They can generally only record a subset of the total audio spectrum. So they're unable to completely reproduce the sound that was created by the instrument. If the ultrasonic frequencies resonate with a frequency that you can hear, it can cause some odd things to happen if the mic isn't able to record the frequency directly.

( Background: I was a professional musician in a past life and I was quite involved in sound reinforcement. ) No two ears, nor no two microphones have the exact same response curve. I'm not yet convinced that what you or the original comment I reply'd to can be explained by differences in curves which, AFAIK, are relatively flat in the sonic regions we're discussing. The mic I referred to earlier looks to be flat to within 3db from 200hz to 20k. Are we discussing signals outside that range or within it?

The idea is that once you've recorded the sound, you want to be able to mix it with other sounds in an arbitrary virtual 3D space, so you need the ultrasound. I think.

I see, thank you. I'm not entirely convinced, but it sounds plausible enough. I'm no expert, though.

after reading the other responses, my guess is that there could be something to this if the ear's response is non-linear in some way that makes non-audible frequencies affect audible ones.

that might sound crazy, so let me give an example. cheap active speakers (the kind you plug into a computer) will often pick up radio transmissions from passing emergency vehicles. the speaker cable acts like an antenna and picks up the radio signal. now the radio signal is, of course, at radio frequencies, but there's some non-linearity in the amp that results in an audio signal (an understandable one) coming out of the speaker.

that's a fairly extreme example, but here's another more biologically believable one. perhaps there are frequencies which are high enough to be inaudible, but which influence the ear in how it responds to audible frequencies. for example, perhaps the ultra-high frequency signal stimulates the ear in some way that makes it more sensitive to audible higher frequencies. then you would hear a brighter tonal balance when listening live than when listening to a band-limited recording.

i am not saying this must happen; i am saying this may be the (physically reasonable) basis for some belief / rumour / knowledge in the audio business.

[also, note that the non-linearity must be in the ear; if it's in the surroundings then the mic will pick it up too. and it must be influenced by frequencies that are outside the recording bandwidth otherwise you'd hear it on playback.]

For those who don't know, if there weren't nonlinearities in the ear you wouldn't hear difference tones (https://en.wikipedia.org/wiki/Difference_tone). Those nonlinearities are interactions within the audible range though, not interactions between audible frequencies and ultrasound.

(Difference tones are a convenient way to tune a violin: when two strings are a fifth apart (as they should be), the difference tone is at the octave, and it's easier to get an octave right than a fifth.)