23

It's commonly held that humans can hear sounds in the range 20 Hz - 20 KHz.

But most acoustic instruments will emit sounds outside that range, for instance breath sounds from flutes and recorders, bow noise on bowed string instruments, plucking noise on plucked string instruments, and so on.

One strand of reasoning says that if an instrument produces sounds with components above 20 KHz then they're inaudible and so don't really matter.

But is that true? Does the ultrasonic part of an acoustic instrument's spectrum affect its perceived sound?

I'm specifically asking about acoustic instruments and what it's like to be in the same room as that instrument as it's performed. I'm specifically not asking about audio recording/storage/replay, as that's a whole subject on its own.

This question was prompted after hearing a program Costing the Earth on BBC Radio 4 last night which reported that in the rainforest there's significant ultrasonic sound and that's why we feel better there than in the urban environment.

Subsidiary Question Seems no-one has the answer at their fingertips. I've been wondering what experiment would answer this question?

12
  • 5
    Do bats flying by outside keep you awake? I'm sure this could descend rapidly into a deep philosophical discussion ;)
    – Tetsujin
    Mar 30, 2023 at 12:48
  • 5
    @Tetsujin - not often, but bats inside...
    – Tim
    Mar 30, 2023 at 13:27
  • 2
    What about constructive interference? Could two ultrasonic frequencies combine to generate a sonic one? Maybe? Though, as noted in some other answers, this gets messy because you're asking about actual acoustic instruments. Their waveform is so complex, and there are so many individual sound sources going on (as you noted—breath, string sounds), that's it's very hard to parse which frequencies are "affecting" which others. Mar 30, 2023 at 13:46
  • 3
    For a single instrument, the constructive interference would already have happened and the audible frequency would be there. I'm a bit skeptical about the idea that the ultrasounds from several different acoustic instruments would interfere to produce audible sounds or that those sounds would be musically significant.
    – ojs
    Mar 30, 2023 at 14:19
  • 1
    @ojs well, the interference and distortion may occur inside the ear... also some young people do hear above 20 kHz. Mar 30, 2023 at 15:31

9 Answers 9

20

As far as I know up to this point there is exactly 0 scientific indication that any human will perceive music differently due to ultrasonic sound content, no matter what turntable producers want to make you believe.

High frequencies of certain parts of instruments might affect the sound due to non linearity. E.g. an Aetherophone uses a very high frequency oscillator mixed with a second high frequency oscillator, which produced the difference frequency which might then be audible.

But generally this will rather affect low frequencies than high ones, as high frequencies experience more dampening, and less effects will actually result in an audible distortion.

But the high frequencies actually emitted by the instrument – most likely no.

10
  • 2
    I agree that ultrasonic content in the air has no affect when it’s heard by a human. It’s possible that ultrasonic signal levels in audio electronics could contribute to intermodulation distortion and generate new non-ultrasonic artifacts. But I wouldn’t expect that to be a factor 99% of the time and only a very minor factor that other 1% of the time. Mar 30, 2023 at 15:22
  • 1
    The more technically minded audio marketing people usually claim that the benefit of higher sample rates is that filtering artifacts are shifted outside the hearing range.
    – ojs
    Mar 30, 2023 at 16:29
  • 3
    @ojs The benefit of slightly higher sampling rate than necessary is to give more head space for the necessary low pass filter, required to avoid quantization noise. Having a few KHz more space for your low pass filter to operate may in fact make an audible difference. But this question is not about digital audio.
    – Lazy
    Mar 30, 2023 at 17:02
  • @Lazy I agree that the question is not about digital audio but the subject often comes up with these unusually high sample rates, and usually it's not about few KHz but twice or four times the usual rate. Quantization noise however is related to bit depth, not sample rate. Steep lowpass filters have different problems.
    – ojs
    Mar 30, 2023 at 19:08
  • 1
    @ojs Bit depth is a result of quantization along the elongation dimension, sample rate a result of quantization along the time dimension. The first leads to noise in the sense of rounding error, the second leads to noise in the sense of nonuniqueness of representation for not rate/2 bandwidth limited signals.
    – Lazy
    Mar 30, 2023 at 19:44
9

As most of what we hear of other people's instruments is from a recording, most of this becomes moot anyway.

If you're in the room listening, you're still not hearing what the mic hears. Just try standing near a drum kit in a recording studio, then comparing it with what the control room hears for an extreme example.
Don't try this at home - by putting your ear 1" from the snare head, you'd get what the mic gets… except the mic won't even blink at those SPL levels [120 - 140 dBSPL potentially]. You, on the other hand, will suffer at least some temporary, it not permanent, hearing loss & will have experienced more pain than 'hearing'.

If you've got a violin on your chin, then half of what you hear is distortion in the nearest ear - you don't notice it because you're used to it. If you're trapped in a small room with one, it makes your teeth squeak.

So… where do we measure this possible high frequency 'interference' that we may or may not be able to perceive? Inside the instrument, 1/2" from it, far enough away that we're getting a good amount of room reflection in it, outdoors, in an anechoic chamber? Take your pick.

My best, flattest mic [so good they took one to Mars in 2020 to capture sound, I kid you not] is only good from 20 - 20kHz. You can get mics that go higher, but they tend to be for specialist use, not for music.

7
  • 1
    You mention a lot about mics, recording, which OP states isn't part of the equation here.
    – Tim
    Mar 31, 2023 at 11:57
  • My entire answer questions the circumstance under which this may be perceived, quite specifically. The OP's question is so general as to be almost unanswerable - as we see from the myriad attempts to answer small sub-sets of the overall theme. The entire second half of the question was added after many of these answers were written - moves the goalposts entirely. My personal take on all this is highly dubious, involving psychoacoustics.
    – Tetsujin
    Mar 31, 2023 at 12:06
  • Without detail of their experimental method, I would question anything that just bluntly claims, rainforest makes you feel good because of the ultrasound. Without significant double-blind tests, removing other senses, then it's no better than the simple statement 'people feel better in the woods than they do in an urban jungle surrounded by noisy people & lots of traffic.'… which I'd easily accept as true, though you'd probably still feel that way with your fingers in your ears; so long as it wasn't pi**ing down with sideways rain & 6°C.
    – Tetsujin
    Mar 31, 2023 at 12:21
  • 2
    @Clockwork - the numbers will likely be skewed from regular extrapolation by the fact that the ears can 'relax' to protect themselves, up to a point. I honestly don't know any exact figures for anything like that. I do know every time I hit the snare in a rehearsal room, everybody but me blinks. That continues until everybody else is playing, by which time, they all 'got used to it'.
    – Tetsujin
    Mar 31, 2023 at 15:48
  • 1
    @Clockwork - you're just replacing one bit of hand-waving with another. It makes no difference whatsoever what the sound is, your ears just naturally compress the sound & de-sensitise after a while. If I were a doctor, I'd know what it's called. I'm not… & I don't ;)
    – Tetsujin
    Apr 1, 2023 at 11:23
4

Simple answer, most likely not - to humans.

There must be many, many sounds out there that are sounds, but are inaudible to humans. Most, it figures, we are unaware of, for obvious reasons. Dog whistles, for example. We know they work, as they have an effect on dogs, but we're blissfully unaware of the racket they make to dogs!

As we know, most instruments will not only produce the fundamental sound of a note, but varying degrees of its harmonics. An interesting point would be how that sound (due to its harmonics) would be 'heard' by different individuals. Some time ago, I built a speaker cab with an 18" speaker and a horn tweeter. The band members could hear the white noise emanating from the tweeter - I was blissfully unaware! So we heard the sound even when notes were played, in different ways.

So, basically, no, and speakers, headphones, et al, which are said to reproduce sounds outside the human hearing range may well do so, but they're as much use as the anti-elephant paint on my chimney pot...

4

There's some evidence that ultrasonic frequencies are in fact audible! https://en.wikipedia.org/wiki/Ultrasonic_hearing

But I think most arguments for reproducing frequencies above 20Kz come from the fantasy world of Russ Andrews. https://www.russandrews.com/

There's also the issue of what happens when you sample with a Nyquist frequency TOO close to the top of audible range. There's no true 'brick wall' filtration.

3
  • I once listened to a simulation of the sound of a jet engine with and without its ultrasonic frequencies, and could clearly hear the difference. Ultrasonic frequencies made the sound more realistic. Mar 30, 2023 at 21:13
  • 2
    @PhilFreedenberg Was this a blind test?
    – DKNguyen
    Mar 30, 2023 at 22:23
  • 5
    @PhilFreedenberg - Russ Andrews sells snake oil to gullible 'audiophools'. I didn't find these jet engine examples, but a) do you know what they were recorded with & the audio spec right through, & b) were you listening on monitors capable of better than 20kHz at the top, from an appropriately high-rate source? A late c) have you ever had your hearing tested to know what range they currently have? btw, you can fake this stuff for the gullible in several ways - by squeezing cut off filters really, really close to nyquist or vice versa, so there's aliasing right at the very top of the range.
    – Tetsujin
    Mar 31, 2023 at 7:59
2

One can start with the fact that the nominal 20kHz limit is somewhat different between individuals and some (esp. younger ones) are sensitive to a kHz or two above 20.

Could this make difference exactly in regard to music is an open question.

I personally can hear up to e.g. 15kHz, but cannot really distinguish the pitch of a single tone between 7kHz and anything above. It may be generally impossible or a matter of training, I don't know. Even if it is a matter of training, probably few people will have it.

On the other hand, human ears (and the processing circuity between them) are known to be non-linear in a number of ways.

Loud sounds (and even not so loud ones) are known to alter the sensitivity of the ear outside of the sound's own spectrum, so e.g. an ultrasonic beat could leave auditory artifacts over the usual, audible sound.

Again from personal experience, the ultrasonic rat repellents (be damned whoever invented them) make me almost deaf even if set at frequencies that I am unable to hear.

In short, one can hear artifacts of an ultrasound even if unable to hear the ultrasound itself. The effect is small and I think it will be cruel to the listeners to use it even if possible.

7
  • 1
    With such a 'precision' device as a rat scarer, it's quite likely the cheap circuitry or even the speaker itself is subject to modulation dropping into human-audible territory.
    – Tetsujin
    Mar 31, 2023 at 10:42
  • Hell yeah, some of them do, but my comment is about the non-audible frequencies that still have audible (over some other sound) effect.
    – fraxinus
    Mar 31, 2023 at 10:47
  • You're not hearing the 'well out of range' sound, what you're hearing is 'interference', lower frequencies caused by cross-modulation or resonance. You'd hear this even on a limited frequency recording.
    – Tetsujin
    Mar 31, 2023 at 10:51
  • Whatever the interference is, it happens in my ears - up to not understanding a normal speech - without hearing the ultrasound itself.
    – fraxinus
    Mar 31, 2023 at 10:55
  • Have you tried recording it?
    – Tetsujin
    Mar 31, 2023 at 10:56
1

It is possible to hear ultrasonic sound under the right circumstances, essentially if you have interfering ultrasound waves that generate beats at a lower frequency. This can be used to transmit audible sound directionally, some key words are "parametric speakers", "audio laser" and "beam forming".

As explained on directional-loudspeakers:

The speaker array actually produces a modulated wave made of two separate ultrasound waves. One of them is a steady, reference tone of a constant 200,000 hertz (Hz) frequency (the carrier wave) and the other is the signal that fluctuates between 200,200 Hz and 220,000Hz (the modulating wave). Although they're combined, it's easiest to think of them as two separate waves traveling out in parallel straight lines through a column of air without overlapping. If they meet an obstruction (such as your curious head), they suddenly slow down and mix together so they interfere constructively (by adding together) and destructively (by subtracting from one another). By the principle of wave superposition, two ultrasound waves of those frequencies can subtract from one another to produce a third wave with a much lower frequency in the range 200–20,000 Hz—and that's in the frequency range that your ears hear.

Could it be that interferences of ultrasound waves therefore effect how a music instrument sounds? I think so.

Can it be experimentally tested? It is tricky, as in the explanation above: As soon as these interfering ultrasound waves hit an obstacle (such as an eardrum or a microphone) they turn into regular sound waves, so even if you record a full audio spectrum from 16Hz to 200 kHz, these would be recorded at the lower interference frequency. Therefore filtering out the high frequencies would not change that. What should be possible, at least in theory: Record only the ultrasonic part of the spectrum (or record the full spectrum and filter out the audible frequencies) and play it back. Whatever is audible then would be the contribution of the ultrasound frequencies to the sound of the instrument. Whether or not that is technically feasible, I don't know.

0

I find the question VERY theoretical. The given frequency range is just the extension of the sequence of harmonics, which will surely be influenced by any constructional modifications you may attempt to strenghten or weaken the supersonic parts. But with changes already in the easily hearable harmonics, how to decide (beyond recording/low-pass filtering which may have more impact than the effect you are searching for)?

2
  • Perhaps theoretical, but well-established theories based on a long history of experimental science.
    – Theodore
    Mar 31, 2023 at 19:13
  • When does supersonic regime come into effect? Honest question difficult to find answer. Apr 1, 2023 at 5:05
0

Not in any direct sense, but there's an indirect sense where it might matter: all of the sound (audible and inaudible) is physical movement, and all of it is energy given up to the atmosphere by the instrument. So if you were to take an instrument and modify it to make more or less ultrasound, you might find that doing so alters the timbre of the instrument in ways that you wouldn't expect from a simple analysis.

The tension on a drum's head could evolve differently, giving it a different "note" and a faster or slower decay. The distinctive sound of bells and gongs is caused by the amplitudes of a large number of different resonances evolving over time (transferring energy between each other) after the instrument is struck, and any ultrasonic resonances would have an effect on that. A bowed instrument might exhibit a different amount of friction between the string and the bow. And so on: musical instruments are complex and nonlinear.

I'm not qualified to say whether any of these effects are significant — I'm not an expert in the field and I would love to hear from someone who is — but I would be very surprised if none of them were.

-1

I would say yes and no. If there is an ultrasonic frequency or fundamental, it would have harmonics which you will hear because of the ultrasonic part. However that sound can be produced with out ultrasonic wavelengths. So it’s still sub 20khz that you hear. But higher then 20 kHz still effects and resonates with lower frequencies and produces new sound you will hear but in the sub 20 kHz region and so these harmonics do not require super 20 kHz fundamentally but are naturally produced by them where there is harmonics.

So I think in working reality the simplest answer is yes. They can definetly effect the sounds you hear though you can not hear them alone.

Further if the given the ultrasonic frequency, frequencies which divide the ultrasonic frequency, will have a harmonic and the amplitudes will be effected at a given frequency which divides the ultrasonic frequency. You may hear this lower frequency that divides the ultrasonic frequency get changed. All the common divisors will be effected. But because the high frequency happens more times per second, it will have smaller and smaller effect as the area of the amplitude of the wave gets smaller and smaller. There is less power per amplitude.

So you would think that these ultrasonic waves probably in natural occurrence become less and less important in the sound we hear. But with digital audio or anything specifically designed to utilize these ultrasonic regions to hit harmonics, they certainly can have hearable effects.

It is also interesting that because you can not hear these frequencies on their own they can be used to construct harmonics, that can’t be constructed and sound the same- as with waves using audible wave lengths.

Thus ultrasound is not totally insignificant, in sound and music.

9
  • 2
    I find it difficult to follow your argumentation. Much of what you're writing seems circular / begging the definition, or simply unsupported claims of dubious veracity. Apr 2, 2023 at 10:27
  • Ultrasonic frequencies DO factually effect audible sounds via harmonics. That is fact. However when the wave as a total is broken down to its constituent frequencies, the ultrasonic part can be removed and no difference would be heard. Apr 6, 2023 at 16:12
  • So in the non analog and digital world. Using various instruments or machines say a violin and a dog whistle. You should be able to get harmonics which you hear, IN THEORY. Apr 6, 2023 at 16:13
  • The issue is this is best to understand with science and Laplace transform however with out that music theory etc and natural music does not typically cover these areas. Science deals with waves and hence Laplace transform a lot. This question is precisely analogous to many physics and engineering concepts. Be it light waves, sound waves, it all is modeled by the same math and science. Which is quite common for physics etc. And this question quite easily decided with a YES. Apr 6, 2023 at 16:21
  • Basically the laplace transform takes any crazy repeating wave like a real sound wave, and tells you the frequency spectrums which are present. The frequencies can all be added up or these many sound waves of different frequencies mix together and produce the original wave. Apr 6, 2023 at 16:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.