Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Two Ultrasonic Frequencies Create an Audible Sound (Read 96172 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Two Ultrasonic Frequencies Create an Audible Sound

I found this interesting post on gearsluts ? The general gist is that if we add 2 very high frequency sines together they can create a beating effect that is audible to our ears.
Here is the quote of the example:-
Quote
the "beat" created between these two ultrasonic tones is equal to 1Khz (57K - 56K), which is a frequency right in the middle of the human hearing range and reproducible on pretty much any speaker...There is a mathematical propensity for harmonics to converge about three and a half octaves above the fundamental


I'd love to hear what people think of this?

Do Frequencies above 20khz really matter?

Reply #1 –
Quote from: boomtopper on
I'd love to hear what people think of this?


Short answer: bullshit.
Long answer: I didn't read, I just found what I expected to find scanning that site. You could repeat this with 1Ghz and 1.000001GHz and claim, we needed gear reproducing audio up to this frequency. But as a matter of fact, we neither store nor reproduce audio as sine waves. This 1kHz tone is well recorded and reproduced with standard 20-20 gear as it is (*obviosly*) within that range. The only difference is, that the two source frequencies are neither on the recording nor reproduced. Nyquist-theory as a hint.

Do Frequencies above 20khz really matter?

Reply #2 –
Quote from: boomtopper on
I'd love to hear what people think of this?


the beat may indeed be 1kHz but i tried it just there and couldn't hear a damn thing. my ultrasound sensor did however register a signal so there was an ultrasonic sound.

so, based on my experience, i'd have to say its more baloney from the audiophools.

Do Frequencies above 20khz really matter?

Reply #3 –
It's difficult to comment without having more details about the test, but there's a huge chance that IMD (Intermodulation Distortion) in the power amp is responsible for this effect.
Could the speakers reproduce the 57 kHz sine ? If so, was the 1kHz beat still there if the two sines were played back on separate sets of amp/speaker ?

Do Frequencies above 20khz really matter?

Reply #4 –
Quote from: Kees de Visser on
It's difficult to comment without having more details about the test, but there's a huge chance that IMD (Intermodulation Distortion) in the power amp is responsible for this effect.
Could the speakers reproduce the 57 kHz sine ? If so, was the 1kHz beat still there if the two sines were played back on separate sets of amp/speaker ?


Nothing spooky in here, its just beat.

Do Frequencies above 20khz really matter?

Reply #5 –
Someone could clearly create a recording, wherein two >20k frequencies are recorded separately and only "meet" during playback.
In this situation reproduction of said frequencies > 20k would be needed to hear the beat.

In real life, though, beats in the audible range will be picked up during recording, and perfectly reproduced by your common, garden variety, CD player.
Creature of habit.

Do Frequencies above 20khz really matter?

Reply #6 –
Quote from: Soap on
Someone could clearly create a recording, wherein two >20k frequencies are recorded separately and only "meet" during mixing / mastering.

Okay, right. I agree.

Quote from: Soap on
In this situation reproduction of said frequencies > 20k would be needed to hear the beat.

If they do meet during mixing, that is not true. Only if you had them on separate channels and they would meet in your room, you'd need that gear.

Do Frequencies above 20khz really matter?

Reply #7 –
Quote from: soulsearchingsun on
If they do meet during mixing, that is not true. Only if you had them on separate channels and they would meet in your room, you'd need that gear.

Yea, I caught that brain fart, thus my editing of my post.    Looks like you quoted me before I edited!
Creature of habit.

Do Frequencies above 20khz really matter?

Reply #8 –
Quote from: boomtopper on
I'd love to hear what people think of this?


I have no time to read the gearslutz thread but see this this product, a very directional loudspeaker that uses a similar principle.
Ceterum censeo, there should be an "%is_stop_after_current%".

Do Frequencies above 20khz really matter?

Reply #9 –
No it's not audible, as confirmed by insane_alien's test, and for good reason. It's incorrect to think of this (eg 57kHz-56Kz = 1kHz) as being in the "audible range". A better way to think of it is that it's more like a very high frequency signal (56.5k) that you can't hear being modulated (volume increased up and down) at a lower frequency. Imagine taking something that is completely inaudible and then turning it off and on at 100 times per second. Do you think that this will miraculously make it audible at 100Hz - it wont.

Edit : That was "Kees de Visser's" point back in post #4 btw. This is theoretically inaudible unless a non-linearity causes (via inter-modulation distortion) the production of a real frequency component (as in additive not multiplicative) at the modulation frequency.

Do Frequencies above 20khz really matter?

Reply #10 –
So for this to happen the signal needs to be passed through a non-linear system such as Amplitude Modulation or Distortion. The reason it occurs is because the 2 signals create sidebands much like ring modulation does. Is this right?

Do Frequencies above 20khz really matter?

Reply #11 –
Right (and this can be verified with basic trig too: f(x) = cos 9x + cos 8x; f^2(x)= ...?)

But: just because such intermodulation products are reproduced does not guarantee any sort of fidelity. That is, without knowing the details on exactly what causes such intermodulations in a "reference" environment, you have no way of knowing whether a different system will reproduce the intermodulations at the same levels. In other words, an extended frequency response alone does not guarantee "fidelity".

Do Frequencies above 20khz really matter?

Reply #12 –
I must have been misguided then. I'm apologizing for my arrogant tone.
(edit: omg I'm so embarassed)

Do Frequencies above 20khz really matter?

Reply #13 –
Quote from: Axon on
Right (and this can be verified with basic trig too: f(x) = cos 9x + cos 8x; f^2(x)= ...?)

But: just because such intermodulation products are reproduced does not guarantee any sort of fidelity. That is, without knowing the details on exactly what causes such intermodulations in a "reference" environment, you have no way of knowing whether a different system will reproduce the intermodulations at the same levels. In other words, an extended frequency response alone does not guarantee "fidelity".


For this effect to be audible what sort of circumstances would make it audible?

Do Frequencies above 20khz really matter?

Reply #14 –
Researching this further. It looks like it's not possible if transmitting 51khz and 50khz thorough the air to cause the 1khz harmonic as there is nothing to produce the harmonic. Unless the signals are being transmitted loud enough to cause the non-linear effects in the air. This of course could be bad for the ears. Another factor that could create this harmonic could be the speakers as they are also a non-linear system.

Do Frequencies above 20khz really matter?

Reply #15 –
Also another thought that came to mind is the room. If the room was fairly reverby this could lead to the 1khz harmonic being smeared thereby making it difficult to hear.
And as Kees De Visser said the Power Amp could be causing the harmonic via intermodulation. So putting some of this together leads me to believe that there are alot of variables involved in this such as:-
Speakers
Systems being passed through
Amps
Room

Another thing to note is in the article is that is says "beat" now I at first took this term to mean in the sense of binarual beats. But seeing as it is intermodulation that is causing the harmonic it is not in this sense.

Do Frequencies above 20khz really matter?

Reply #16 –
For the domain of sound reproduction this is all irrelevant. You can surely create setups where two 50 kHz waves in a room create a 1 kHz modulation. But when reproducing sound this will usually be an unwanted artifact and a function of a specific listening room (and not related to the original soundstage). All sub 20 kHz intermodulations of higher frequencies, that actually were existent at the original event (at the microphone positions), are perfectly preserved with standard Redbook audio CDs. Any > 20kHz content, that could be recorded at the microphone positions is not supposed to cause any more <20 kHz intermodulations at playback time!

If you think this through, it would even make sense to low pass >20 kHz content intentionally in real world setups for higher fidelity. Subjects can't hear HF content directly, so why reproduce it in their listening rooms, where it could only cause audible intermodulations that were not part of the original soundstage? Since they can't hear it directly, they won't miss it and also get less listening room specific artifacts.

Do Frequencies above 20khz really matter?

Reply #17 –
I felt the need to register to chip in a bit here - the discussion I am having with @boomtopper is too big for Twitter!

I am not trying to post the 'final word' on this - the subject is just fascinating and it is a topic well worthy of debate (a traditional method for sharing and improving understanding of science!)

I love the subject of beating because it is the simplest form of dissonance, and I have done extensive academic research into dissonance theory (or as we call it, 'gride' theory) - see the link to my thesis at the end of this post.

The original question was essentially:

Quote
If we add 2 very high frequency sines together, do they can create a beating effect that is audible to our ears?


This is entirely possible, but I don't think it is proven. Just as when audible freqs are within critical bandwidths (1/6 of the freq), they bounce, for example when you play two piano keys that are next to each other. The beating is at a very low rate, like a tremelo. Beating is proven and common, listening for beating is helpful when tuning guitar string (for example).

Beating will occur regardless of whether we can hear the 'original' two frequencies at, say, 56 and 57KHz. The question above, and really the point of the debate, is whether we can hear only a side effect of interaction between freqs out of our hearing range.

I wanted to quote a few posts to this thread and dispel a few myths/assumptions...

Quote
Short answer: bullshit.
Long answer: I didn't read, I just found what I expected to find scanning that site. You could repeat this with 1Ghz and 1.000001GHz and claim, we needed gear reproducing audio up to this frequency. But as a matter of fact, we neither store nor reproduce audio as sine waves. This 1kHz tone is well recorded and reproduced with standard 20-20 gear as it is (*obviosly*) within that range. The only difference is, that the two source frequencies are neither on the recording nor reproduced. Nyquist-theory as a hint.


Probably worth reading things before you dispel them, as you say "the two source frequencies are neither on the recording nor reproduced" - it never says anything about this test being recorded, and there is no evidence the person who did this test does not have the equipment with the range and ability to play back frequencies up at 56 and 57k.

Let's assume they hypothetically do have the kit that could reproduce - some super high grade speakers with a tweeter sensitive enough to move back and forward up to 60 thousand times a second, analog gear with an operational limit up to or past 60k (I know op amp freq responses don't always go that high, but it IS possible, I don't see why certain equipment wouldn't have the capacity to do it - but this is a hypothetical question, so stick with me).

So we have a system that can produce these two 56 and 57 KHz signals. This system would produce the frequency at 1k, because the speakers are trying to reproduce two waveforms which phase and then sum an amplitude at a speed of 1000 times per second. So the speakers ARE producing an audible frequency of 1KHz.

This is under the condition that the 56 and 57k are coming out the same speaker What if the speakers were playing one frequency each? So the left one is playing the 56k, the right one the 57k. We wouldn't hear the sounds in isolation, but I suppose we don't know whether the 1k would be audible in this instance - the only place the beating could occur is on the basilar membrane in the cochlea. Naturally occurring beating, where the source of the beating is not a physical object trying to represent two close frequencies as described in the speaker example above, is just your basilar membrane reacting in the same way, unable to 'play' two frequencies in consonance because they are within a critical bandwidth of each other (1/6 of each others freq.)

So two sine waves on a piano, say a C and a C#, played at the same time (at the same./similar amplitude or masking can occur, where the louder note masks the other), hit two separate springs, which are both representing the C accurately on one string and the C# on the other string, have no place for beating to occur other than when it actually gets into your ear and your basilar membrane (For a longer explanation of the basilar membrane, and how it acts/reacts and looks like, see my academic work I link to at the end of this post).

(An open question here is actually whether normal air has the capacity to represent the frequencies, because then beating is occurring before it gets to your ear - yet another area in psychoacoustics which is not fully researched. Seriously, this field is so open that if you want to make a name for yourself in science, this is the area with enough niche to go and find something groundbreaking. It is also why so many people have these kind of debates, because actually we don't know for certain some of the points discussed in this thread, and neither side of the argument is proven or disproven other than in theory, and that we should all have an open mind, or perhaps sometimes some skepticism when talking about 'proven' facts.)

Of course, this is assuming a system with a super high grade frequency range. I don't know for certain that such a system is easily availabe, though if anyone would have equipment that high grade, Geoff Emerick probably does.

Let's not get into the digital/analog debate, we all know the pros and cons of both systems, so let's assume we all know that a digital system at a sample rate of 44.1KHz just wouldn't reproduce the 56 and 57k at all, that sample rate can only theoretically reproduce frequencies as high as 22.05KHz (and only if the converters are any good... we'll move onto this).

However, a 192KHz sampling rate could, so again assuming someone had the speakers that could play a freq as high as 57k, a digital system with decent clocking (at least apogee or prism audio quality) could indeed process all those frequencies we are taking about. We just wouldn't hear them... unless two frequencies within a critical bandwidth of each other are played, and then we have the same situation as with the above hypothetical analog system.

You are welcome to say that it all ends up on CD at 44.1khz and it would lose it. I think it would, unless the side effect of the 1KHz 'beating' does actually get recorded, but that is more of the debate which is open, because you'd have to test it, and the discussion is something of a headfuck as it already is! And of course, why would we even burn this test to CD?

I want to touch on Nyquist theory for people, because I get the impression people don't fully understand it.

I'm sure you all know the gist of it - the highest potentially audible frequency, supposedly 20KHz, can only be properly represented with a sampling rate of twice that. If you aren't aware of that much, do and read a little more about it somewhere else.

But if this is the case, why would anyone record with a sampling rate above 44.1KHz?

Hands up, who records at a rate higher than 44.1KHz?

In fact why isn't the sampling rate just 40KHz? Think of the sampling rate having a bit of a slope to it, like a low pass filter - 44.1KHz works because the 'slope' is higher, going up to 22.05KHz, rather than lower down in an audible range. This is with poor/common converters, you probably won't get this with Apogee (or above) quality. This is why, when we don't have that expensive quality, we use 48 or up to 96KHz, which people often describe as a more 'open' sound, because the theoretical LPF when using those sample rates is way higher and therefore does not affect the range we can hear, up to 20k.

If anyone says to you about how they use 96k sampling because the 'resolution' makes it sound better, they are actually wrong (stick with me), because 44.1 is technically as much resolution as you need to represent audible frequencies. 'Resolution' is an often cited reason for why higher sampling rates are better, and yes the resolution is higher, but it is not the reason people would naturally think at first, probably to do with presuming it is the same kind of thing as when an image has a lower resolution. It isn't at all. (There may indeed be a benefit with plug in processing at higher sample rates, but I am not qualified to answer that aspect of it).

--------------------

Okay, back to responding to quotes! I've covered everything but want to get specific.

Quote
the beat may indeed be 1kHz but i tried it just there and couldn't hear a damn thing. my ultrasound sensor did however register a signal so there was an ultrasonic sound.


You may have realised after reading my post that you may need:

- Speakers that represent beyond 57KHz
- A sampling rate of 192KHz
- High quality converters
- Your ears exactly in line with the tweeters, since freqs are more direction as the get higer (a bass sub can go anywhere in the room, it is non directional, but higher freqs increasingly get more directional. Try a sine at something like 16KHz from only one speaker, listen to it of axis, then listen to it exactly on axis to the tweeter, it will be louder

And IMO there is no need to spend that kind of money of kit for a little experiment like this!

Quote
Imagine taking something that is completely inaudible and then turning it off and on at 100 times per second. Do you think that this will miraculously make it audible at 100Hz - it wont.


Interesting example, because this is actually how beating works - the quick, repetitive phasing between the two signals IS turning it on and off because is nulls then sums at a new freq. BUT... to turn something on and off at 100hz just means something inadiuble is turned on and off, and remains inaudible. You are forgetting the summing part of the beat phenomenon.

Quote
It looks like it's not possible if transmitting 51khz and 50khz [he means 56 and 57k] thorough the air to cause the 1khz harmonic as there is nothing to produce the harmonic. Unless the signals are being transmitted loud enough to cause the non-linear effects in the air. This of course could be bad for the ears. Another factor that could create this harmonic could be the speakers as they are also a non-linear system.


You mean 56KHz and 57KHz. I think the part of my post here answers where the beating effect is produced.

I can why you might think you need the >20KHz  to be really loud to cause the beating at an audible volume, but they wouldn't have to be that loud - remember the equal loudness curve/fletcher munson curve? We are SUPER sensitive to sounds in the upper mid range, like 1KHz to 5KHz (roughly, depends on the person). We would be able to pick up a sound at 1KHz easier than ANY potentially audible frequency. Consider this - imagine you have a bass sine at like 80Hz and a sine at 1KHz. If you have a mixer in front of you now get them up on two tracks, then try and match their apparent volume. The 1KHz fader would have to be a lot lower, because you are more sensitive to it. I heard that if that 1k sine actually had the same power as the bass note, it could kill you! Anecdotal of course, but the point is that a bass note has to have a LOT of energy just for us to hear it like normal. An upper mid range sound could sound just as loud as that bass note with very, very little energy, such is our sensitivity to that range.

Now, I don't know for certain how much energy the 56 and 57k signals would need to produce enough energy for the 1K beating effect to be audible, but just remember that out of every feasible frequency, for example 0Hz to 1 million Hz(!!), the range of 1-5KHz(ish) needs the LEAST energy to it to be audible to us. Get me? Just for your consideration.

--------------

Now to answer some of @boomtopper 's points on twitter:

Quote
There was one guy on that thread saying he could hear the differance with a soldered joint?


Yes, this is a facinating story. Although it is not written on an academic, peer reviewed paper, we should also give it the credibility it deserves, for the following reasons:

This is an interview with Rupert Neve, of course as in Neve, the effing better preamps and EQs etc of the entire history of recording. Rupert Neve, while talking to Fletcher (BTW, Fletcher is a very trusted and respected individual in the pro echelons of modern recording culture - he runs mercenary audio - though his credibility is irrelevant to this story, but perhaps of interest to you), recalls a story about Geoff Emmerick, the most famous and emulated of engineers - you surely all know he engineered The Beatles, and pretty much invented many of the technique we use on a daily basis.

If anyone had equipment good enough to playback these inaudible but somehow perceivable >20KHz frequencies, and trained enough ears, then it would have been this guy in this studio. And in turn, Rupert effing Neve agreed there was something different about these channels. And lo and behold, they found there actually were transformers incorrectly fitted. Assuming the story is true, (and remember Rupert Neve was a first hand witness to this, and Fletcher is as trusted as it comes), then this would be evidence that somehow, some people have perceived a difference when there was a +3db peak at 54KHz. I am not saying 'look, this is guarenteed proof of the concept', but it is a convincing demonstration of an effect we definitely do not know enough about.

We should admit that science has not discovered everything to do with audio and the brains perception of it. We haven't by a long shot, and if you disagree, read a book on psychoacoustics, write down all the questions you think of along the way, then discover most of everyone's questions have not be researched let alone answered.

Quote
It [the 1KHz side effect of 56 and 57KHz] is not a beating effect.


Yes it is.

Quote
The bottom line is supersonic frequencies matter a lot less then you think


They don't particularly matter, according to our current understanding of science - I'm not trying to make out they matter even a percent of how much the audible range of 20-20k matters, but I want everyone to accept we really don't know that they don't matter.

Quote
the other way of creating this 1k sideband would be to play the 2 [56 and 57k] sines throught the air at very high volumes

This of course would be bad for the ears.


It wouldn't be bad for the ears, because the 56 and 57k sines are inaudible, so they wouldn't damage you. (as described ealiar, you wouldn't necessarily have to have those sines loud to get the 1k beating side effect. Maybe you would, but like I said before, probably not).


---------------

This discussion on the effects of higher inaudible frequencies is very valid and overdue, in my opinion. They may be inaudible, but they may be perceivable. At the other end of the spectrum, few would argue that sub 20Hz freqs are at least percievable. If a hypothetical 10Hz wave was coming out a speaker the size of a house, you wouldn't 'hear' a wave at 10Hz, but by god you would feel it. You might haemorrhage and die, such is the power! (infrasonics, the <20Hz range, is just as interesting and slightly more studied that ultrasonic sound).

Psychoacoutics is TOTALLY under-researched. Really, just because we can't hear a frequency above 20KHz (or lower, of course, because of ageing) doesn't actually mean that it's presence doesn't in some way affect our perception of it. In all my reading on the subject I have not come across conclusive proof (or any kind of proof) that we do not react to higher frequencies in some way. The topic of a side effect with beating that we have been discussing fascinates me, because it should remind us that their presence can have an audible effect. What other tricks are lurking up there? Could research on this subject actually make sense for showing that HD audio can actually have a measurable effect on quality/perception/air/space on a record? Could this be the unique selling point for HD audio, something everyone had previously decided wouldn't make a difference since we can already get up to 20KHz with a regular CD player?

I urge anyone at uni with a dissertation coming up to study this, it would be a great topic to cover. There are gaps in our knowledge and perception of audio, and it may or may not be of our benefit to fill these holes.

May the discussion continue.

For my academic paper on 'Analysis, Testing and Discussion of Extreme Dissonance' which covers similar topics as here (but <20KHz), click here and download the relevant PDF.

Do Frequencies above 20khz really matter?

Reply #18 –
Oh, please say whether you read all that or not. I understand it is a big read, but try not to argue against parts of it without seeing it all, the whole piece was just bashed out in one go so important points are a bit scattered.

Do Frequencies above 20khz really matter?

Reply #19 –
I read most of it and still ask myself why this should matter. There is no reason to question the possibility of audible <20kHz modulations between multiple >20kHz waves. But all this doesn't matter for the domain of sound reproduction due to the reasons given in my last post, which were also already mentioned in earlier posts.

This (once again) wave of Gearsluts hysteria isn't really worth so much attention. Weee! Somebody has shown that the product of two > 20kHz waves can have audible components. Fine, then you just record these components and ignore the HF parts, as we are doing it for ages. That's it.

Do Frequencies above 20khz really matter?

Reply #20 –
Indeed, the relevance of >20k in music production is totally questionable. But it does not mean to write it off. My entire point is that we should be open to the potential psychoacoustic effects of these ultrasonic frequencies have, not beating specifically, just that there could be things that one day will prove useful in music production. Or not.

Do Frequencies above 20khz really matter?

Reply #21 –
Quote from: rpp3po on
If you think this through, it would even make sense to low pass >20 kHz content intentionally in real world setups for higher fidelity. Subjects can't hear HF content directly, so why reproduce it in their listening rooms, where it could only cause audible intermodulations that were not part of the original soundstage? Since they can't hear it directly, they won't miss it and also get less listening room specific artifacts.


Filtering with both LPF and HPF is something I always do to clear headroom in a mix - filtering (apparantly) inaudible information clears amplitude space for the entire mix, especially at the lower end.

BUT...

Quote
it could only cause audible intermodulations that were not part of the original soundstage?


Maybe... just maybe... these unintended modulations could be a critical part of an ethreal quality for the recorded sound. If in an A/B test, the unfiltered version for whatever reason sounded better, you would keep it, yes?

Did you read my section on sampling rates? What sampling rate do you record at?

Do Frequencies above 20khz really matter?

Reply #22 –
Quote from: PlayMusic on
Indeed, the relevance of >20k in music production is totally questionable. But it does not mean to write it off. My entire point is that we should be open to the potential psychoacoustic effects of these ultrasonic frequencies have, not beating specifically, just that there could be things that one day will prove useful in music production. Or not.


I can't rule that out here. But it usually goes like this:

A. Subject listens to < 18 kHz signal and identifies it.
B. Subject listens to < 18 kHz plus ultrasonic signal and is able to discern it from A.

Ergo ultrasonics are audible.

No! Ultrasonics can modulate audible signals. The correct test would not compare A to B but B to properly low passed B, so that intermodulations within the audible band are present in both cases.

Of course you need a proper low pass. But in my experience the following is true:

For any signal B including ultrasonic content, there is a possible low pass filter F, that both eliminates all frequencies above a ultrasonic threshold and results in a signal B', that is indiscernible from B for humans.

Edit: typo

Do Frequencies above 20khz really matter?

Reply #23 –
Quote from: PlayMusic on
Maybe... just maybe... these unintended modulations could be a critical part of an ethreal quality for the recorded sound. If in an A/B test, the unfiltered version for whatever reason sounded better, you would keep it, yes?


Unintended audible modulations, that were present during the original recording and part of its "ethereal quality", can easily be captured with 44.1kHz equipment.  Anything unintended beyond that point is completely contingent, but in any case no improved capture of the "ethereal quality" of the original scene.

I do record in 96kHz but only to keep more headroom for filtering. For pure listening of final mixes and records I don't need anything above Redbook.

Do Frequencies above 20khz really matter?

Reply #24 –
This is what needs to be tested (the AB test you suggest).

Interesting:

Quote
Unintended audible modulations, that were present during the original recording, can easily be captured with 44.1kHz equipment. Anything unintended beyond that point is completely contingent, but in any case no improved capture of the "ethereal" quality of the original scene.


This is a good point - but two things arise from it.

- It suggests that regardless of whether or not the ultrasonic frequencies are recorded, their audible effects are somehow useful. We don't even know for sure that there are common effects, but if they are, then recording them is irrelevant here - the point is whether humans somehow perceive them.

- If we refer to a beating effect from any ultrasonic sounds, you assume in your example that the effects only can be recorded if the effects are present during recording. But you have forgotten the importance of where the beating effect can potentially take place (the instrument/speaker/air/ear), which was an integral part of my post, like when I mention the two separate piano strings. In a way, these effects don't exist unless there is something to perceive it! Brings a whole new meaning to the 'If a tree falls in the woods...'