How DO you describe that audible inaudible difference properly?
If it's truly inaudible, then by definition it cannot be sensed by the human ear and is therefore irrelevant. (I am not talking about subsonic frequencies that can be felt as vibrations but not heard, but rather frequencies above the range of hearing, theoretically > 20KHz but for most of us anything over 15KHz to 18KHz.)
However, ultrasonic frequencies are not necessarily inaudible except when they exist as pure sine waves. In the presence of other frequencies, anything other than a pure sine wave, ultrasonic components interact with each other and with audible frequencies, generating sum and difference frequencies that are indeed audible.
Remember that much of the texture of sound that we hear in an acoustical space is the result of room dynamics, resonances, and acoustical summations and cancellations. The interplay between all of the components result in sum and difference frequencies that account for a great deal of the richness of complex sounds.
Cymbals are a good example because even though their fundamental frequencies are fairly low (a large crash cymbal might be down around 1.5-2KHz) they generate a very complex waveform with frequency components that extend way out past audibility. Those high frequencies combine to form
new frequencies that the cymbal itself did not generate, but that only came into existence as acoustical phenomena. It's the main reason cymbals sound so different live than when recorded. As those frequencies decay at different rates, the sum and difference frequencies ebb and flow, creating the rich dynamic texture that is the holy grail of every recordist to capture.
So there is a definite advantage to preserving "inaudible" frequencies during the analog phase of recording, especially for instruments that have a rich high-frequency component, such as cymbals and acoustic guitars. Of course, those frequencies will be completely eliminated during the A/D conversion process if your sample rate is 44.1KHz,
as they must be. Consequently, you are only going to record audible frequencies below 20KHz no matter what pains you took to capture higher frequencies within the analog realm. But it shouldn't matter -- the acoustical magic has already happened prior to conversion, and those synthesized frequencies formed by the interaction of ultrasonics are already in there. As Nyquist theorized and Shannon proved mathematically, a 40KHz sample rate can accurately record and recreate any audible frequency.
The question, then, is what advantage there might be in preserving those ultrasonic frequencies throughout the entire process, including within the digital domain, so that they have a second shot at acoustical interaction at playback time. The answer is a definite "maybe". You'll hear the effect on some instruments, but not on others. Whatever portion of your music was derived from synthesizers and samplers will not benefit. Unless you are recording primarily live acoustical instruments, you'll just be wasting disk space.
Geez, I hope this is making any sense at all. I am on medication this week, as I am recuperating from a heart attack that whacked me this past Monday night. I'm supposed to kick back and do nothing, but blathering on about digital audio is as close to doing nothing as I can manage.