Who can hear above 20Khz?

I think audiologists are concerned with everyday functioning. Speech, plus things like telephones and doorbells. Very high tones are not necessary.

FWIW, I'm 51 with an upper limit of 14 kHz.

 

I have read that harmonics relies on accurate timing of the waveform. That's probably a bad way to describe it but the point is that it's not just about a piece of equipment being able to record and reproduce extra high frequencies that results in good harmonic reproduction, but it's the timing accuracy of the recorded frequencies that's important.
With a full analogue waveform all nuances are captured at an accurate time line. With a complex waveform these nuances are important and noticable.
For example a simple 5 micro second 14 kHz signal could change in start time by a few nano seconds and if the recording equipment can't capture this accurately the timing of the waveform will be wrong. Now think of the incredibly complex waveform of music and you will see why timing is important.
Both digital and analogue would be able to capture this 14 kHz signal but the analogue capture will have a better time line accuracy because it will resolve the signal accurately at any time.
I think this is what makes analogue sound different to digital.
So although high frequency hearing ability is important much of the music information is in the timing of the lower frequencies. Even people with reduced hearing ability will still be able to benefit from music recorded at a high resolution.

 

Very interesting question. I can no longer hear sine waves at 16k but still seem to hear 20k material on recordings. Obviously we(I) react different to steady tones than we do to complex musical spectra.

 

As a kid, I could hear my parents' little tabletop stereo through two closed doors at night. I could make out songs but not what the DJ was saying. My dad was shocked when I told him, because he would typically turn the volume knob all the way down instead of turning the unit off, to avoid the possibility of a small thump from the speakers and slight change in ambient lighting.

I could also hear if someone had snuck downstairs and turned one of the televisions on with no sound. In that case, I could hear the high pitched squeal of the picture tube one and even two stories below.

Trust me, that ain't happening now!

 

Any theory why digital should lag behind analog a few nanoseconds?

 

Sorry if my explanation sounded like that but that's not what I am talking about.
What I am saying is that without the waveform being accurate over the timeline, harmonics can be affected, and harmonics are important to the
overall sound quality.
Often people say that CD is good enough because it records to 22 kHz and we can't hear more than that. But it can not resolve harmonics accurately because the sampling points are too far apart.
For example, if we were to consider 1 second of a 18 kHz signal recorded to CD versus 1 second of 18 kHz signal recorded to analogue they would be similar because the digital system would use filtering to rebuild the waveform reasonably well.
However, if we recorded just a micro second or even less of that 18 kHz signal with digital it would not have enough resolution to rebuild the waveform accurately. That's of course due to insufficient sampling points. The analogue recording would not have this deficiency. Music is complex and not just a simple sine wave like the above example and where high resolution is necessary.
So what I was saying is the ability of a recording system to record at any moment the correct value is important more so than just increased frequency response which we would not be able to hear anyway.
Higher sampling rates for digital will help but not for the purpose of recording higher frequencies, but for increased accuracy of the waveform on it's timeline.
Although it's subtle, high resolution recordings will make a difference even to people who can't hear the higher frequencies because they can still resolve the timing of the signal.
Hope this is a bit clearer explanation.

 

Is this an authoritative source? Not saying it isn't - just curious. I could use it for mixing.

 

Please provide your mathematical evidence disproving the Nyquist Theorem.

 

I think it also has to do with having more bit depth of the audio of Hi Res. Going from 16bits to 24 is a big jump. Smoothing out clipped signals.

The Real Differences Between 16-Bit and 24-Bit Audio - Tested.com

24-bit sound is a tricky thing to gauge. Does it provide for a greater resolution of sound? Definitively. It has room for 256 times the data, remember. Are you going to be able to hear that difference? Harder to judge. Human hearing supposedly tops out at 20kHz, but that doesn’t make higher sample rates useless. According to the Nyquist rate, to fully capture a wave, it should be sampled at twice its highest frequency. In other words, a higher sample rate, and a greater bit depth, gives your sound more wiggle room, meaning sound peaks are less likely to be truncated and the subtleties of the music are less likely to be drowned out.

 

Think about what I wrote and you will be able to use logic to work it out.
Even Sony thinks it's worthwhile to use higher sampling rates which suggests the Nyquist Theorem is flawed.
Nagra has tape recorders capable of up to 384 Kb/s and that was years ago. The engineers aren't designing equipment with these specs for no reason.

 

The 24 bit depth just is an easy way to get slightly better sampling quality while remaining compatible to the 44.1 KB/s standard.
96 KB/s 16bits would be much better than 48 KB/s 24bits.

 

If human hearing is limited to around 20kHz, having that extra frequency range does very little to improve the sound. I think the biggest benefit comes from much more detail by way of bit depth. And that extends all the way through the frenquency range. Thats my personal experience and was also outlined in the article I posted. You kind of stated it yourself there, "better sampling quality" as opposed to quantity.

Thats not to say having more sample rate doesnt benefit the music either. Because allowing more frequency affords less compression and clipped signals at the top end.

 

In order to complete just one single cycle of 14kHz would require 71.4us of time. So, what fractional portion of this single cycle are you referring to when you record only 5us of it? 5us of 14kHz would not be enough information for you to even determine the frequency of this tick. 5us is only about 7% of a single cycle of this 14kHz.

But, that's not really the point. Even lowly redbook (44.1/16) has timing resolution down in the nanosecond range. Medical studies have shown that humans are able to detect timing differences at least down to 10us, with some studies saying 5us is the human limit. CD timing resolution is way way better than that. So it would seem your incredibly complex waveform of music is no match for "standard" digital.

If we are talking about audio recorders, then the analog recording would MOST definitely share in the same deficiency that CD is hobbled by. Just recording a mere microsecond, of any audio signal, would be completely ignored by an analog recorder. More than likely, the amplifier paths within the recorder don't have a bandwidth into the MHz range. The head definitely won't be able to respond to this high of a frequency. And the tape won't be moving fast enough to allow even one magnetic particle to pass by the head in 1us. The CD may have insufficient sampling points, but the analog recorder has insufficient bandwidth. And, by some miracle of the universe, those two (sampling rate and bandwidth) are tightly related.

 

I remember as a child wondering what the big deal about dog whistles was, as I could hear the tone clearly, and it was sufficiently unpleasant to send me scurrying away rather than calling me in. Now, at 58 and after a lifetime of working with heavy equipment, I am fortunate to still have adequate hearing up to about 14khz.

I may be misremembering my music theory, but it was once explained to me that both a flute and piccolo can play a high C, at roughly 1050hz. What enabled the human ear to differentiate which instrument was being played was the ratio of the harmonics; that is, each would have a different relative strength at 2100, 4200, 8400 and 16800 hz (again, approximations). So even though there is precious little musical information above 8000 hz, those superharmonics produce the sonic fingerprints of closely related instruments that enable us to differentiate them, and the more multiples of those frequencies are contained within the recording and even marginally audible to us, the better we can identify all the component parts of a musical palette. It made sense to me many years ago.

 

Image is from The Music Espionage - Everything Music Tech That page looks pretty legit to me.

My DuckDuckGo keyword search pulled up a wide array of results from their Images link.
frequency spectrum musical instruments at DuckDuckGo
(fwiw, unlike Google, DDG still allows direct linking to images, instead of only linking to site sources. DDG doesn't spy on your Internet browsing, either...)

 

Well -- I am 61, and very dependent on my hearing for verifying audio processing software. I am very good at detecting/perceiving distortions somewhere below 10kHz, and my hearing is about gone at 14kHz. It starts disappearing at about 12kHz, however. I remember when the 15750 whine from TV sets (in USA), and the old traffic sensors at stop lights would be very intense. Also store security systems were very loud for me -- and I remember that I was about 25 when I last heard a store security system semi-ultrasonic signal.
Now, tinnitus messes with my ability to hear hiss, so I am partially dependent on spectograms and other means of detecting hiss.

However, my ability to detect distortions within my range of hearing is supurb -- and has been very helpful in some of my design/testing. Of course, I don't solely depend on my hearing, but it can give very good helpful hints.
I do worry about distortions which would be obvious to a young person, but I might totally miss. I had a bout of very poor hearing about 6mos ago, and produced some insanely distorting software (lots of distortion above 10kHz, and temporarily, I had lost about 10-15dB at 10kHz.)

Wish I had 25yr old ears again.

john

 

Regarding the sweep above -- I heard some intermodulation in it -- so if someone thinks that they might be hearing just barely to about 10kHz, some of that sound isn't really 10kHz.

John

 

I have tried many of these different online tests, but my soundcard causes aliasing on most of them(where you can hear a tone going up or up and down). The only one I have tried that seems ok is at the site linked below.

The standard high frequency sweep is a 44.1khz file and this has aliasing for me. The 48khz file does not seem to have this, so this is the only one I would use as a rough benchmark.

I know that in my 20s I used to be able to hear the high pitched tone that runs thru the original Love - Forever Changes CD, but I can't hear that now and it is at roughly 15.6khz.

Using the 48khz file at the site below, I hear the descending tone come in at around 15khz, so that would seem about right. I can hear something start to come in at 17khz, but the descending tone does not start until around 15khz. I'm 47.

Extended High Frequency Online Hearing Test | 8-22 kHz

It also has an aliasing test file.

 

How big are your paychecks? Not saying you are wrong, just wondering how big your paychecks are.

If they are big enough, I just might be able to find a doctor that will say whatever I want....

 

My Dog and my wife can hear 20k.

I’m at just around 18k. Just had it tested.

 

Yes, noise artifacts are possible, however I referenced the You Tube tones on my speakers with a sound level meter which easily reproduce to 20 kHz, on You Tube nothing, stone cold at 18kHz, but strong at 17kHz. You Tube (or the internet service I am connected to) brick walls above 17 kHz. You do not need calibrated equipment to measure no signal at all (or nearly nothing if anyone wants to argue tech on that, nothing useful in the real world to put it that way) True that the variables are many, such as the limitations of ear buds, or cheap speakers, or limitations of frequency response online. A sound level meter will always verify if you have a tone, regardless if it's heard or not.

More about this...

Audiologist tests are not for music, but should be. The test covers intelligibility of speech only, and yes, a plot is generated, and yes this test is calibrated. While useful, this test does not cover higher frequencies found in musical harmonics.

The online test can be useful even though not a calibrated test. But be careful, and know your equipment. We can test for what the audiologist exam does not cover, all frequencies above 8kHz. And we can ascertain hearing imbalance, or nulls in one ear, as subjectively we CAN tell is a frequency sounds louder in one ear than the other.

Our hearing isn't flat. Even this normal non-flat curve deviates person to person. Our ears hear differently, and I believe our brain does too. Localized nulls in hearing also develop with age, and hearing nulls from damage, ie: a firecracker at close range damages only those frequencies it produces, apprx 4kHz to 8kHz range. For instance, my left ear has a sharp null at 4kHz, much lowered sensitivity very narrow bandwidth. At 3kHz and 5kHz, both ears hear equally well. At low level below 60dB (phons) the imbalance is noticeable. At higher levels the null does not exist.. very strange indeed, (but some low level nuance lost at that precise 4kHz center) This info was found by a self non-calibrated test through Sennheiser HD420-SL headphones and confirmed through several speaker systems. I can not say how many dB the null is, except that it is noticed at lower levels... and I know it's there. I do believe many (or most) ears have these nulls at various frequencies.

Strangely for me this isn't a detractor listening to music, the vocals still sound correct and centered, instruments in their locations. Only pure tones reveal the problem. The first time I did a frequency sweep at the young age of 20 (as a new audiophile) I detected no nulls, just a nice increasing pitch all the way to 18 kHz, when it abruptly faded.

 

Thats interesting you write that. When I did a test (for fun) I noticed there is a range where I stopped hearing the test but Im able to hear the test after the "null". While reading your post I wonder if tinnitus phases out that portion of the test.

 

Interesting point about tinnitus, I think possible some kind of masking effect. I'm guessing the masking would be mostly at the tinnitus frequency? I'm hoping for a cure very soon for that.. just on the horizon.

 

Maybe someone can engineer something that restores the loss frequencies like implementing the RIAA curve
or through bioengineering repair damaged cells