Testing audiophile claims and myths

[analogsurviver]

with so many people saying they can hear something at 17 or 18kHz on that sweep, you shouldn't reject the possibility that they really do.
oh wait! ^_^

just in case I also checked at various resolution settings for the vid, same thing.

It is perfectly possible for the people using lower quality devices to hear "something" up there ; decent devices should be quiet, but not-so-decent might well emit some kind of noise/distortion. Whatever it is, it is not intended signal, though.

 

[gregorio]

[1] I'm not an expert in any aspect of audio or electronics
[2] I have enough relevant science and engineering background to engage in discussions about measurements, test methodology, how science works, limits of theory, etc.

1. Yes, that's been clearly established.
2. You've actually established the exact opposite. You demonstrated that you didn't know that digital audio data is in fact just a measurement but worse still, you demonstrated that you didn't even know that a measurement could be represented as binary (digital information). How that's possible in this day and age, when it's taught at school, let alone in university, is mind boggling for someone claiming to have a formal, relevant science/engineering education! Clearly you do not and cannot engage in a discussion about measurements and the "limits of theory" as it applies to digital audio if you don't understand even the basics of how and what digital is. You continually deflect from this fact by casting aspersions on others and quoting cognitive errors, when you yourself have virtually no understanding of the subject. Dunning-Kruger taken to the extreme!

For the record, I am not a research scientist, I'm an audio engineer. I do not have a degree in the subject because when I started there were no degree courses in the subject, education in the subject was effectively by means of the apprenticeship model, so I'm not self-taught or rather, only partially self-taught. Ironically, that is why I was originally employed by the university in the first place, to design a degree course! My university (conservertoire) education was in the field of music performance, so as I did not have a degree in the specific subject area I had to undergo endless rounds of internal and external examinations, which eventually assessed my knowledge and certified me to teach students up to and including PhD level. In practise, I've been certified and employed as an expert in the field by universities, research scientists, industry bodies and a government, just to name a few and you've got what, a measly degree in an unrelated subject? However, I'm not a great fan of appeals to authority (and even less so with appeals to effectively ignorance!) and none of this is relevant most of the time here anyway, because we're often not talking about who is or is not currently a research scientist, we're talking about research that was done and dusted (and proven!) 200 years ago, 150 years ago in the case of electrical properties and 70 years ago in the case of digital audio!

[1] Most current studio recordings don't contain much musical information above 20 kHz.
[2] The reason for that is that most studio microphones have little response above 20 kHz.
[2a] (And most recording engineers don't expend any effort trying to record them or preserve them during the mix.)
[3] And the reason for that is that information above 20 Khz isn't considered to be useful because it is assumed that nobody can hear above 20 kHz.
[3a] Therefore, since the information is largely absent, and no attempt is made to preserve it when it is present, nobody is going to hear it.
[4] HOWEVER, this doesn't prove either way whether people would or would not hear content above 20 kHz IF IT WAS THERE.
[5] (Again, I'm not SPECIFICALLY claiming that it's audible or not... just that we haven't actually tested the possibility... we simply "go with the assumption".)

1. By and large true.
2. Also true.
2a. Not entirely true but again, by and large true.
3. Sort of true.
4. Correct, it doesn't by itself prove that.
5. And there we have it yet again! Some true or partially true statements followed by a completely false statement and a totally fallacious conclusion! You have in fact got it entirely backwards, we "go with the assumption" precisely because we HAVE "tested the possibility"! The reason we assume >20kHz content isn't useful is because ALL the evidence and practicalities of music recording creation demonstrates that it is NOT useful, while there's NO evidence to suggest that it might be, which incidentally is why most studio mics are designed to have little response above 20kHz!

[1] Generally microphones are chosen to "sound good" - which includes being able to pick up whatever frequencies contribute to that result.
[2] I have little doubt that, if tests were to find that "cymbals sound better when recorded with microphones whose response extends to 30 kHz".... We would soon see studio microphones, designed specifically for cymbals, whose response extends to 30 kHz... and sold specifically for recording cymbals.
[3] I've personally made very few recordings... And I've certainly never recorded a drum kit using a microphone whose response extends to 30 kHz "just to see if anyone thinks it sounds better"... Have you actually tried it?

1. Interesting, you do in fact have at least a very basic idea of why mics are chosen. Why then do you keep inventing scenarios and quoting mics which are contrary to this requirement? Hmmm!
2. Agreed, that's a logical idea, nothing wrong with it at all. However, in reality it's nonsense because we have done exhaustive tests and cymbals do not sound better when recorded with mics whose response extends to 30kHz and that's precisely why, after 20+ years of testing, there are no mics specifically designed for recording cymbals!!!!
3. You're joking right? If you suddenly had access to a new technology, what would be the very first thing you'd do with it? Wouldn't you test those new technology aspects exhaustively and start with the most likely and common source to expose that technology? So, of course I've actually tried it, countless times over many years and so has pretty much every experienced commercial sound engineer.

[1] Is it REALLY so much more unreasonable to buy a file whose frequency response extends to 45 kHz, "just because you prefer to have something a lot better than the minimum you actually need",
[1a] ... than it is to buy a sports car that can go 120 mpH and accelerate like the wind, or a precision stainless steel ruler that's accurate to 1/64", even though we all know that 1/8" is plenty accurate for that magazine rack you're building? [1b] Have you considered the possibility that some audiophiles simply enjoy knowing that their equipment, and their music, is significantly better than they'll ever need, instead of "just good enough"?

1. Yes, it is unreasonable because it's just a marketing lie that you're getting "something a lot better than the minimum". What you're actually getting in the best case scenario is something that's exactly the same and commonly, something that's actually worse!
1a. Is it unreasonable to buy and pay extra for a sports car that is advertised as a sports car with 120MPH top speed but at best only performs the same as a ford fiesta and at least some of the time can't even keep up with a ford fiesta?
1b. And have you, as a member of the trade, ever considered the possibility that convincing audiophiles they're getting something better than they'll ever need and omitting the fact that they've already got something better than they'll ever need (by falsely characterising it as not good enough or "just good enough") is a profitable marketing tactic?

A lot of the recent discussions have included the claim that "ultrasonic frequencies are so heavily attenuated by air that they never reach the listener in significant amounts".
[1] However, with all the anecdotal claims, I haven't seen much actual DATA, so I thought I'd provide some.
[2] The attenuation, at 30 kHz, under those conditions, is 0.647 dB / meter ..

1. Ideal, two falsehoods/fallacious tactics in one sentence, I'm impressed! Firstly, how come you haven't seen much actual DATA when so much of it has been presented, did you "inadvertently" miss it again? (For example this post #10636). Secondly, the old audiophile marketing tactic of providing "some" actual data but "inadvertently" ignoring the other relevant data and thereby misrepresenting the facts/data! ...

2. Under more realistic conditions (50% humidity, as musical instruments do not cope well with low or high humidity), the high freq absorption is 0.937dB/m which at say 20m is a loss of 18.74dB. However, as stated in the previous point (and in previous posts!), that is not the only relevant data! There is also damping of sound level vs distance, IN ADDITION to purely high freq absorption. At 20m (relative to say 0.5 meters) we loose 32.04dB, so now our 30kHz signal is down by 50.78dB. Another piece of ACTUAL DATA you've ignored (despite the fact it's been posted several times) is that in the case of a cymbal, only about 6% of it's energy is above 20kHz. So, we've got a relatively small amount of signal >20kHz to start with and what's left is going to be attenuated by over 50dB. We should also consider the rising thermal noise of mics and mic-preamps with frequency, which definitely comes into play with signal loss over distance and If we're talking about a live situation, then we should also include the significant damping cause by the audience itself. After all this attenuation, there's simply nothing left to record above the noise floor, let alone hear! So yet again, round and round in circles we go, caused by you yet again (inadvertently?) misrepresenting the facts. Are you ever going to stop insulting this forum?

This suggests several things.......
1) There in fact is significant attenuation of even audible high frequencies at the distances you might typically encounter in a living room or concert hall
2) Attenuation of even relatively high ultrasonic frequencies is significant - BUT FAR FROM ABSOLUTE - at typical listening distances
3) Attenuation of even high ultrasonic frequencies at one or two meters (at the distance of a "close mic", for example, on a drum set) is relatively minor

1. True.
2. True but only if you (inadvertently?) omit the relevant, uncontested facts and (inadvertently?) ignore some of the laws of physics!
3. True but I ask you again (because you refuse to answer) do you or an audience ever sit with your ears one inch away from the snare drum during a performance or a couple of feet away from the cymbals? Have you ever tried it? Have you any idea what it sounds like? Do you even have any idea how preposterous it would be?

here is where I stand right now: + we shouldn't reject audibility above 20khz as we have some kids who notice 22khz or even higher just fine and that alone disproves the notion that nobody is hearing ultrasounds.

I'm not disputing that given the right conditions 22kHz is audible, in fact quite the opposite, I've stated that given the right conditions, everything is audible, even much higher than 22kHz! The relevant question is: Do those conditions ever exist in the recording and reproduction chains of commercial audio (or live performances) at any reasonable listening level? Firstly of course, no one ever makes commercial audio recording of only a 22kHz tone. >20kHz content (when it does exist) is ONLY the result of higher order harmonics and is tiny in amplitude relative to the audible band. Even if it were possible to hear 22kHz at a level of say only 100dB, in commercial audio recordings that means the peak level (in the audible band) would have to be at least 130dB and typically, more than 140dB. If instead of 22kHz we were talking about only say a 15kHz signal, it's extremely unlikely that at >40dB below the (more audible band) content it could be "heard" in a commercial audio mix. So, what's the implication of the claim, that our ears perform better at 22kHz than at 15kHz? Admittedly there's essentially (AFAIK) no research in this area, a fact which is misrepresented/abused by those with an agenda. The reason there's no research is that it is too dangerous to carry out such research, peak levels of 130dB and above will cause pain and damage, >140dB will cause almost instantaneous damage. In other words, the reason the research has not been carried out is itself evidence/proof that it's inaudible.

What's disingenuous is that those supporting the claims actually seem to have some understanding of the facts but just abuse them depending on what they're arguing about at the time. For example @Phronesis stated "I think it's important to use a pure tone, otherwise harmonics can confuse us about what's going on." - As seems to be so often the case in these disputes, a true argument that's used to support one specific point is in fact self-contradictory and an argument against the claim itself. While Phronesis' statement is true, in practise it supports the argument AGAINST the claim because very high/ultrasonic frequencies in commercial audio recordings ONLY exist as harmonics and high order harmonics to boot! These tactics have gone way beyond tiresome!!

G

 

[KeithEmo]

Gregorio has said, and asked, a few things that have me a bit confused here.....

For example, we seem to agree that, at 30 kHz, there is significant attenuation over distance.
And Gregorio keeps pointing out that, at sixty feet or so, there wouldn't be much of it left at all.
However, in most multi-track recordings, there's a microphone a few feet above the cymbals.
And, at a distance of a few feet, the maximum attenuation, even at 30 kHz, is only a few dB.
(I've even seen a few microphones actually stuck inside the cymbals).
That suggests to me that the attenuation at further distances is somewhat irrelevant.
(Are they deliberately rolling things off at 20 kHz so it will sound like the microphone was 60 feet away?)

Gregorio and I seem to agree that most engineers and performers choose microphones by "sound".
However, we also seem to agree that most studio microphones have a limited high end response.
And, as Gregorio points out, there don't seem to be many if any "microphones for cymbals".

However, while Gregorio seems to imply that there are no studio microphones whose response extends above 20 kHz, or at least exceptionally few, he also suggests that, like any "interesting new technology", they have in fact "been tested over and over again over the years".(Am I to assume that tests were actually conducted to see whether listeners found microphones with extended high frequency response to NOT sound better? If so, I'd be curious to know if they used equipment for those tests designed to record and reproduce those frequencies, or if they ensured failure by testing them in conjunction with a signal chain not designed to reproduce what they recorded.)

(It seems to me that those recordings, made with those experimental microphones, with the response that extends to 30 kHz, that we know contain the ultrasonic harmonics that are produced by cymbals, and are only attenuated by a few dB at such close distances, would be ideal test files for testing whether people can tell the difference between CDs and high-res recordings. And, in fact, those are the microphones I would be using for all those tests. I guess luckily they do exist after all. So, are we to assume that every studio has a few, languishing in the back of the closet, because nobody thinks they sound especially good? )

I do agree that the concept of "better" is somewhat vague... and excessively open to interpretation. For example, we could argue whether a meter with 0.05% accuracy is "really better" than one with 0.5% accuracy, if all you really need is 1% accuracy, or whether a car that can go 120 mpH is "better" than one that can only go 80 mpH, so perhaps we should settle for saying "with technical performance that exceeds the minimum necessary requirements". (However, just as that Jaguar will actually do 120+ if you floor it, a 96k file is most certainly able to really reproduce frequencies up to 45 kHz if they are present.)

My problem here is that I keep seeing what I might call "an appeal to status quo".

Basically:
- Nobody does it because everybody knows it doesn't matter.
- Everybody knows it doesn't matter because nobody does it.

Well, maybe it's time to break the loop, and do some actual testing.

How about we start by recording a few drum solos, with lots of cymbals, using one of those long-forgotten microphones with the response to 30 kHz, stuck five feet right over the drum set? We'll record it at 24/96k so we know those ultrasonic overtones are really preserved (actually they're technically not harmonics with a cymbal). And we'll make sure to use a microphone preamp with a response into the ultrasonic, and confirm on an editor that those frequencies are really there. Then we'll play those recordings back, on speakers or headphones, which we've tested to ensure that they actually have a response that extends to 30 kHz or so..... and, once and for all, confirm how many people hear a difference and how many don't (and if that number is zero or not).

And, yes, selling people equipment that technically performs better, for a higher price, is often profitable.
And, for this reason, many companies are going to try to sell you their highest performing product.
Ferrari and Jaguar specialize in selling very expensive products that probably perform far above any need.
But Volkswagen had a very good run selling "a really cheap car that got the minimum job done very well".
However, neither has any bearing on what you or I, or anyone else, needs or wants.

I'm also a little confused about exactly what you mean by "damping of the sound". We have the attenuation experienced by the sound over the trip from the instrument to the microphone, which is all that will affect the sound arriving directly at the microphone from the source - over the shortest direct path. Then we have things like surface absorption, which will affect later reflections, and so the long-term averaged response. So, as those other losses increase, the average long-term response will get duller, but the original "bite", experienced over the first few tens of milliseconds, will be controlled only by the attenuation experienced from direct losses.) Also, as Gregorio pointed out, the attenuation increases significantly with distance..... so, for a microphone positioned five feet over the drum kit, we can probably ignore everything everything except the sound reaching the microphone directly through those five feet of air (with their 4 - 5 dB of attenuation at 30 kHz). We don't need to even consider all that stuff that's 50 dB down after reflecting off the audience.

Yes, cymbals sound quite different when you're standing three feet away than when you're standing twenty feet back in the audience. But, most of the time, they do in fact end up being recorded from three or four feet away - and not 20 feet. And, if 6% of the energy of the cymbal is at ultrasonic frequencies.... Then, one meter above the cymbal, where that overhead microphone is mounted, it's only down about 3 dB. That's still quite significant.

And, if it turns out that even a few people can notice when those frequencies are present...
But "they're never included in most commercial recordings anyway"...
Then I guess most commercial recordings fall short of "the limits of human hearing"...
Which sounds to me like a problem worth correcting...
(Especially since we both seem to agree that it wouldn't be that difficult to do so.)

I will always maintain that the recording engineer counts as "one of the artists"...
Therefore, if the recording engineer PREFERS to record cymbals with a microphone limited to 20 Khz....
(Or if he or she prefers to use a microphone that's limited to 10 kHz for that matter.)
That is certainly his or her business...
But then it's a personal preference... (or "artistic judgment")...
It's neither "a technical limitation" nor "a limit of human perception"....

1. By and large true.
2. Also true.
2a. Not entirely true but again, by and large true.
3. Sort of true.
4. Correct, it doesn't by itself prove that.
5. And there we have it yet again! Some true or partially true statements followed by a completely false statement and a totally fallacious conclusion! You have in fact got it entirely backwards, we "go with the assumption" precisely because we HAVE "tested the possibility"! The reason we assume >20kHz content isn't useful is because ALL the evidence and practicalities of music recording creation demonstrates that it is NOT useful, while there's NO evidence to suggest that it might be, which incidentally is why most studio mics are designed to have little response above 20kHz!



1. Interesting, you do in fact have at least a very basic idea of why mics are chosen. Why then do you keep inventing scenarios and quoting mics which are contrary to this requirement? Hmmm!
2. Agreed, that's a logical idea, nothing wrong with it at all. However, in reality it's nonsense because we have done exhaustive tests and cymbals do not sound better when recorded with mics whose response extends to 30kHz and that's precisely why, after 20+ years of testing, there are no mics specifically designed for recording cymbals!!!!
3. You're joking right? If you suddenly had access to a new technology, what would be the very first thing you'd do with it? Wouldn't you test those new technology aspects exhaustively and start with the most likely and common source to expose that technology? So, of course I've actually tried it, countless times over many years and so has pretty much every experienced commercial sound engineer.



1. Yes, it is unreasonable because it's just a marketing lie that you're getting "something a lot better than the minimum". What you're actually getting in the best case scenario is something that's exactly the same and commonly, something that's actually worse!
1a. Is it unreasonable to buy and pay extra for a sports car that is advertised as a sports car with 120MPH top speed but at best only performs the same as a ford fiesta and at least some of the time can't even keep up with a ford fiesta?
1b. And have you, as a member of the trade, ever considered the possibility that convincing audiophiles they're getting something better than they'll ever need and omitting the fact that they've already got something better than they'll ever need (by falsely characterising it as not good enough or "just good enough") is a profitable marketing tactic?



1. Ideal, two falsehoods/fallacious tactics in one sentence, I'm impressed! Firstly, how come you haven't seen much actual DATA when so much of it has been presented, did you "inadvertently" miss it again? (For example this post #10636). Secondly, the old audiophile marketing tactic of providing "some" actual data but "inadvertently" ignoring the other relevant data and thereby misrepresenting the facts/data! ...

2. Under more realistic conditions (50% humidity, as musical instruments do not cope well with low or high humidity), the high freq absorption is 0.937dB/m which at say 20m is a loss of 18.74dB. However, as stated in the previous point (and in previous posts!), that is not the only relevant data! There is also damping of sound level vs distance, IN ADDITION to purely high freq absorption. At 20m (relative to say 0.5 meters) we loose 32.04dB, so now our 30kHz signal is down by 50.78dB. Another piece of ACTUAL DATA you've ignored (despite the fact it's been posted several times) is that in the case of a cymbal, only about 6% of it's energy is above 20kHz. So, we've got a relatively small amount of signal >20kHz to start with and what's left is going to be attenuated by over 50dB. We should also consider the rising thermal noise of mics and mic-preamps with frequency, which definitely comes into play with signal loss over distance and If we're talking about a live situation, then we should also include the significant damping cause by the audience itself. After all this attenuation, there's simply nothing left to record above the noise floor, let alone hear! So yet again, round and round in circles we go, caused by you yet again (inadvertently?) misrepresenting the facts. Are you ever going to stop insulting this forum?



1. True.
2. True but only if you (inadvertently?) omit the relevant, uncontested facts and (inadvertently?) ignore some of the laws of physics!
3. True but I ask you again (because you refuse to answer) do you or an audience ever sit with your ears one inch away from the snare drum during a performance or a couple of feet away from the cymbals? Have you ever tried it? Have you any idea what it sounds like? Do you even have any idea how preposterous it would be?



I'm not disputing that given the right conditions 22kHz is audible, in fact quite the opposite, I've stated that given the right conditions, everything is audible, even much higher than 22kHz! The relevant question is: Do those conditions ever exist in the recording and reproduction chains of commercial audio (or live performances) at any reasonable listening level? Firstly of course, no one ever makes commercial audio recording of only a 22kHz tone. >20kHz content (when it does exist) is ONLY the result of higher order harmonics and is tiny in amplitude relative to the audible band. Even if it were possible to hear 22kHz at a level of say only 100dB, in commercial audio recordings that means the peak level (in the audible band) would have to be at least 130dB and typically, more than 140dB. If instead of 22kHz we were talking about only say a 15kHz signal, it's extremely unlikely that at >40dB below the (more audible band) content it could be "heard" in a commercial audio mix. So, what's the implication of the claim, that our ears perform better at 22kHz than at 15kHz? Admittedly there's essentially (AFAIK) no research in this area, a fact which is misrepresented/abused by those with an agenda. The reason there's no research is that it is too dangerous to carry out such research, peak levels of 130dB and above will cause pain and damage, >140dB will cause almost instantaneous damage. In other words, the reason the research has not been carried out is itself evidence/proof that it's inaudible.

What's disingenuous is that those supporting the claims actually seem to have some understanding of the facts but just abuse them depending on what they're arguing about at the time. For example @Phronesis stated "I think it's important to use a pure tone, otherwise harmonics can confuse us about what's going on." - As seems to be so often the case in these disputes, a true argument that's used to support one specific point is in fact self-contradictory and an argument against the claim itself. While Phronesis' statement is true, in practise it supports the argument AGAINST the claim because very high/ultrasonic frequencies in commercial audio recordings ONLY exist as harmonics and high order harmonics to boot! These tactics have gone way beyond tiresome!!

G

 

[analogsurviver]

Microphones with response over 20 kHz:

1.) Panasonic WM-61A ( spec'd to 20K, regularly used for break in alarms - breaking glass is in range 30-50 kHz )
2.) Bruel & Kjaer 4006 / DPA 4006 ( and 4003 )
3.) Bruel & Kjaer 4007 / DPA 4007 ( 40k )
4.) Earthworks QTC 30/40/50 ( model number denoting kHz with linear response )
5.) Sanken C-100K ( 100k)
6.) Audio Precision (100k) https://www.ap.com/news/audio-precision-expands-microphone-product-family/
7.) DPA 4060/4061 etc (spec'd to 20k, usable response > 40k )
8.) Microtech-Gefell ( their capsules also used by Josephson in the USA ; quite a few options )

There are MANY other studio mics with usable response over 20k - but very few with specified/certifdied response.

Regarding distances; properly positioned Jecklin Disk can be less than 10 metres away from the sources; here one documented example of my recording, for which I can provide say 30 seconds of either original DSD128 or converted to 192/24 PCM :

 

[bigshot]

Frequencies are a continuum. There is a sweet spot for human ears, and as the frequencies get further and further away from that sweet spot, they become less and less important... As they reach the outer extremes of human hearing they lose the ability to communicate pitch, then they reach the point where they aren't audible at all. You can continue to track them beyond that, but there is absolutely no point.

Music is designed to sit within the comfort zone of hearing. Musical instruments produce a pleasing combination of fundamentals and harmonics in the core frequencies of human hearing. Arrangers use them to produce an overall sound that fills the range of comfortable listening. Musicians don't compose a Concerto for Mosquito Squeal or Air on a Pitchless Bass Rumble. It's fine if your system is capable of reproducing that, but it isn't as important as presenting the core frequencies properly. And because you feel the need to include the frequencies at the bleeding edge of hearing in your system, it doesn't mean that you need to include inaudible frequencies too... just to be safe.

Audio fidelity is relative, but it also has finite limitations. Frequencies that humans can hear are called sound. Ones that can't be heard are called ultrasonics. Ultrasonics are irrelevant to sound reproduction by definition. The bleeding edges of sound are less and less important the closer they get to that line. Thankfully, CDs and inexpensive consumer electronics are able to perfectly reproduce sound. However transducers are not. If you are going to claim that the extreme octave of human hearing is important to you, you should measure the response of your transducers and see what you are really hearing. It is extremely difficult to achieve a balanced response from 20 to 20 using consumer transducers in a typical home. But if you've tried, you realize that the closer you get to the extremes, the less it matters.

If you sing out do re me fa so la ti... You've sung an octave. How much of that octave can you hear between 10kHz and 20kHz? Odds are if you are a normal adult, about half. You can feel from 20Hz to 40Hz. But you can't detect either of those octaves as musical notes... only as mosquito squeals and rumbles. Most of the mosquito squeals that exist in the upper harmonics of musical instruments are masked by lower frequencies or they are at such a very low volume level they can't be heard. In a reasonably large living room, the upper end of the top octave wouldn't even reach from the speaker to the main listening position. It would be absorbed by air and upholstery and wall board before it even reaches you. Yet audiophiles attach massive importance to those extreme octaves for some reason.

We listen to sound every day that is completely missing those outer octaves and we don't complain. Our $10K+ systems probably don't even do a great job of presenting all of them in a balanced way. But it really doesn't matter that much. Balance in the core frequencies is what is important for the reproduction of music. If you had a system that only produced sound from 40Hz to 10kHz, and it did it in a perfectly clean and balanced and dynamic way, it would still sound darn good. On many recordings it would sound just as good as one that did 20 to 20, because not all recordings contain those extreme frequencies.

Sure, go ahead and include those frequencies if you want. But that doesn't mean that they are as important as the core frequencies. It also probably doesn't mean that your transducers are reproducing them as accurately as the core frequencies. It doesn't even mean that you can hear all of them. And it certainly doesn't mean that you also need ultrasonic frequencies too.

It is all relative. And to make practical decisions and good compromises, you need to keep a firm grasp on relative importance. That is in short supply in audiophile circles.

Discussion like this make me feel like I'm Gulliver watching the Brobnagagian Kings argue over which end of an egg to crack.

 

[sonitus mirus]

It's funny they have a lot of hearing tests with frequency sweeps from 20Hz to 20kHz. For example this "20Hz to 20kHz Earrape":


And some people comment they hear something at 18k, but if you analyze the audio it looks like this:

When I listened to this same sweep, I did notice an abrupt cutoff and just assumed it was due to my age, being a product of the 60's.

 

[bigshot]

Sometimes you'll hear a blind spot in your hearing in sweeps too. I think I have one in the treble range, but it is very very narrow.

 

[james444]

When I listened to this same sweep, I did notice an abrupt cutoff and just assumed it was due to my age, being a product of the 60's.

Similar age here, just that for me the 15k cutoff is real. Too many too loud rock concerts in my younger days...

That said, next time someone brags about their hearing on Head-Fi, I might make them listen to this sweep and tell me the highest frequency they can hear.

Just hoping that's not a violation of forum rules, what you say @castleofargh ?

 

[bigshot]

There isn't much audible in music above 15kHz anyway. It's actually a blessing. I remember when I was a kid going into our local Sears store and being bombarded by the high pitched squeals of the fluorescent light ballasts on the ceiling. My parents couldn't hear it but it made me sick and I wanted to get out of there.

 

[KeithEmo]

I agree that the results may be interesting....

I also wonder if there is a tendency with human hearing to ecperience an illusion of pattern completion.
(If your brain, upoon hearing a sweep, tends to imagine hearing it continue after it ends.)
Certain optical illusions work because of a tendency to "see" things where we "expect" them to be.

However, I don't find it especially odd that a situation like that might happen on Youtube.
Most people who post audio or video files have little technical knowledge about the details.
It's quite possible that the person who posted it simply didn't KNOW that Youtube limits it.
They may have created it, or simply copied it from somewhere else, and simply not confirmed the posting.
It's also possible that Youtube has changed their settings.... and that it was a complete sweep when first posted.

I als doubt that Youtube specifically chose to limit audio bandwidth.
I find it more likely that it is simply a default setting in the encoder they currently use.

Commercial video encoders often include a list of "profiles" recommended for common situations.
(For example, the standard "audio and video quality profile" for "consumer DVDs".)
Odds are they simply chose a profile and didn't really think about or even look at the details.

It's funny they have a lot of hearing tests with frequency sweeps from 20Hz to 20kHz. For example this "20Hz to 20kHz Earrape":


And some people comment they hear something at 18k, but if you analyze the audio it looks like this:

 

[KeithEmo]

I recall reading, some time ago, that, if you are exposed to a certain frequency or range of frequencies at an amplitude and duration sufficient to cause hearing damage, it tends to create a gap or loss in sensitivity at a point in the spectrum ABOVE the frequency that caused the damage. This was quoted for instances of "industrial hearing damage" due to extremely loud equipment. (It has something to do with the energy being directed to the wrong nerve fibers in your ear due to some sort of mechanical overload condition.) I believe they suggested that, if damage occurs, it tends to be at a point approximately 1/2 octave above the frequency of the signal that caused the damage. (The counterintuitive part is that the damage does NOT render you incapable of hearing the sounds that caused the damage.. .)

Yes, this seems somewhat counter-intuitive...
No, I don't recall where I read it...
And, no, I'm not interested in arguing about it...
(I'm sur eit can be looked up somewhere...)

Sometimes you'll hear a blind spot in your hearing in sweeps too. I think I have one in the treble range, but it is very very narrow.

 

[james444]

It's quite possible that the person who posted it simply didn't KNOW that Youtube limits it.
They may have created it, or simply copied it from somewhere else, and simply not confirmed the posting.
It's also possible that Youtube has changed their settings.... and that it was a complete sweep when first posted.

Agreed. So if neither the uploader nor the listener knows about the 15k cutoff, then it's a perfect double blind test, isn't it?

AFAIK, Youtube changed their audio settings in 2014 to the same encoding quality for all video formats. So all the more recent videos have the cutoff. Prior to that, they had different audio qualities depending on video quality, so there may be some legacy stuff on YT that goes higher than 15k.

 

[analogsurviver]

Agreed. So if neither the uploader nor the listener knows about the 15k cutoff, then it's a perfect double blind test, isn't it?

AFAIK, Youtube changed their audio settings in 2014 to the same encoding quality for all video formats. So all the more recent videos have the cutoff. Prior to that, they had different audio qualities depending on video quality, so there may be some legacy stuff on YT that goes higher than 15k.

I beg to differ. Here the audio qualities available on YT up to at least 2016 : https://yd.3dyd.com/help/youtube_formats/

Here a foobar2000 component that allows you to play YT sound in F2K ( with all the processing you might find required ) ; it also identifies the EXACT audio codec used for each upload:
https://fy.3dyd.com/download/

 

[james444]

I beg to differ. Here the audio qualities available on YT up to at least 2016 : https://yd.3dyd.com/help/youtube_formats/

Here a foobar2000 component that allows you to play YT sound in F2K ( with all the processing you might find required ) ; it also identifies the EXACT audio codec used for each upload:
https://fy.3dyd.com/download/

Well, I said AFAIK... and I may be wrong. Mainly gathered that info from here: https://www.h3xed.com/web-and-internet/youtube-audio-quality-bitrate-240p-360p-480p-720p-1080p
https://www.h3xed.com/web-and-internet/youtube-audio-quality-bitrate-240p-360p-480p-720p-1080p
Note that I was talking about the upload date. Of course there are still other audio qualities available on YT from older uploads. I don't think you'll find a recent video that doesn't have the 15k cutoff though.

 

[analogsurviver]

Well, I said AFAIK... and I may be wrong. Mainly gathered that info from here: https://www.h3xed.com/web-and-internet/youtube-audio-quality-bitrate-240p-360p-480p-720p-1080p
Note that I was talking about the upload date. Of course there are still other audio qualities available on YT from older uploads. I don't think you'll find a recent video that doesn't have the 15k cutoff though.

I will have to check.

Always the same BS with digital whatever - IP providers included. Once they figure out what brings the most $, they drop/limit/discontinue everything else.