The first time my hearing was tested, I was in the third grade (60 years ago). At that time I could hear up to 22000 hz. In my twenties I could still hear very high frequencies. I walked in to a Sears store back in the late 70s and I could physically hear all the "ultrasonic" burglar alarms in the store. I could point out ever one on the ceiling without looking up. Today at 67 years old, I can barely hear above 13000 hz. Rock and roll, loud cars/truck, shotguns, rifles and pistol have taken their toll. If I could give one piece of advice, PROTECT YOUR HEARING NOW. You will thank me later in life.

Amir from audio science review made Paul's explanation incorrect.

I don’t understand how this is still controversial. Surely it must be the easiest thing for a speaker / amp / headphone manufacturer or even a hobbyist to check once and for all whether we ‘hear’ very high frequencies in the super tweeter range? Just add a cutoff at, say, 20khz and find out whether it still sounds as good. The fact that so many comments below are about being able to hear very high tones strongly suggests to me that we all have this capability to some extent. It makes totally survival sense - survival of the fittest - to have very high hearing range since being aware of the very sharp click sounds would have been warning of the nearby presence of predators or prey. Why bring phase shift into it? Isn’t this just more commercial mumbo-jumbo?

Hello Paul, there is an additional research on human perception of frequency. There is a Japanese paper released in 2000 about "Inaudible High-Frequency Sounds Effect Brain Activity". Interesting research. Also there is, with the development of DSD audio microphones reaching 100 khz, namely the Sanken CU-100. There are musical instruments that reach beyond 20 khz. Best, Deon.

20khZ And 20hz, those frequencies that are not heard give u a sense of space, room and ambience. People who tune know this.

The human brain is quite amazing in terms of audio processing capabilities.

Thanks Paul. Sometimes I think you should use a white board (or something like that). Then you could draw the 3db points and that would clearly show that the signal slowly gets worse, like i.e. at the 1db points.

Of course, the real answer of why so much bandwidth is quite simple: Gain margins, and to a lesser extent, ensuring that any filtering needed can be done outside the audio band where you wont hear it. For the amplifier bandwidth, this is needed to ensure there is enough loop gain for the distortion mechanisms inherent in linear amplifiers to be dealt with. Since compensation is needed to ensure the amplifier is stable when there is sizable phase shifts present, the bandwidth and loop gain must be made quite high, especially since linear amplifiers inevitably have the open loop gain falling at a constant rate after only a few Hz. To ensure gain margins, the amplifiers normally have at least a MHz or so of useful open loop bandwidth. This is of no relevance to your auditory experience, but it is to the THD+N measurements and flatness of response across the audio band. In class-D and PCM audio, the situation is very different. There you need to stay away from the Nyquist frequency, so they often roll off quite abruptly above the audio band.

ultrasonic and infrasonic vibrations can not be heared directly, but they can create vibrations in our body that we feel. some of those vibrations can also create resonant frequencies in the audible spectrum. the linearity and phase are speaker related effects that are also things we can perceive. additionaly, a more complete spectrum allows the driver to accelerate and decelerate more fluently/ naturaly, improving the sound of audible frequencies. (within the speakers limit of course. very loud infrasonic bass might approach dc, damaging the system. and very loud ultrasonic frequencies may excede the drivers inertia, creating induction heating. but that both needs extreme amplitudes)

We could tell the difference in phase shift resulting in what? I think I get the explanation but not sure what I would be listening for?? Someone approaching from the side, I get. Rolled off highs or lows, I get. But just to say “ we would notice the phase shift”...how? In what way would we notice? I apologize if it was explained and I’m just not getting it, just trying to learn. I hear other reviewers, etc , mentioning things happening in the time domain...is something sped up? Slowed down? A difference from left to right? Anyway, I do appreciate these videos very much, thx Paul

I first encountered one of those pesky rat repellers that puts out 60,000 Hz at 120 dB SPL back in the mid 1990s and it really annoyed me. I'm turning 50 next year and yet I'm still annoyed by those rat repellers every time I get to within 10 feet of it.

Paul is confused. The phase shift effect he describes relates to steep band limiting filters back in the audio chain., not speaker/headphone bandwidth.

I'm from Bucharest also so....thank you!

Howdy. The sine wave frequency response is a thing that belongs to the Hi-Fi domain. In the High End domain other parameters become important. One is burst response. That is how well the amp. can deliver the dynamics of, say, a kettle drum and uphold the power without buckling. Another is Slew Rate that tells how well the upper bands can ride upon high power low bands. To have a good Slew Rate means that the voice and the instruments sound natural on high volume. Slew Rate also delivers the sensation of presence.and directionality. A High End amp. should be able to reach, say, 65 kHz or higher. Regards.

The snake 0:07 bit Paul. yes careful his snake oils bite. better get some antiseptic for that.

I'd bet with those headphones claiming frequency response to 65kHz you could indeed hear them playing at that frequency, but only due to the inevitable resonances and IMD being produced. If you're speakers or headphones play pas the audio band, then its a clear sign of nonlinearity.

ppl 'test their hearing' in uncontrolled environments and are under the impression that they can hear sounds that are in fact inaudible: most often that's because transducers produce harmonic distortions and that's what they actually hear. some play 10Hz on their subs and are convinced they can hear it. in fact the subs also produce 5Hz, 20Hz, 40Hz, etc when they are fed 10Hz. same goes for playing a 30kHz tone... fun fact: a spectrum analyzer shows clearly how much harmonic distortion a transducer produces for a given fq and amplitude. some subs produce the first or second harmonic distortion at a higher amplitude than the actual signal..

Interesting, thank you Paul. Never heard it explained so clearly.

What about harmonics and the ability to reproduce complex sound waves as a combination of multiple order harmonics, having individual frequencies beyond the human hearing capability? I would have thought this to be an additional reason. Comments?

What a coincidence that Paul mentions 25 Hz below. Because for the past two days I’ve been having a discussion on Facebook of people saying that anything under 25 Hz cannot be heard and just makes mechanical noise from the speaker. That is untrue. Although yes a speaker has inherent noise and especially when the speaker unloads but in the case of my system, I can play complete down to 10Hz and I can hear or at least sense the Bass pulses in my ear drums. The amount of output that my system does in the infrasonics far outweighs the amount of “mechanical noise” that they claim. Many people have not had any experience with these frequencies because they are usually tuned higher than 35Hz and then they claim to know it all which is very annoying.

Always enjoy these vids although this high end equipment is just WAY out of my price range. I'd love it if you could do a video on consumer grade equipment with modest price points, say a basic 500 or 1000 dollar system. A similar breakdown on single components in that range would be nice too, i.e. how much to budget per component. For a lot of us even a grand is a pretty big commitment.

Square wave is a compilation of all odd frequencies so 20k response can only reproduce a sign wave. Prolly only get a square wave up to 4k

An absolutely beautiful explanation.

Phase shifts occur even way upper and below the limits. Pushing limits do not cure what is damaged in other parts of frequencies Speakers or headphones can deliver 40kHz but important phase variations occur at much lower harmonics . In such speaker or phones even 12-16 kHz kHz may be masked and not noticeable by ear It is due to reflections of high tone sound and ball shape of dome . It may be easy calculated by difference of distance. Or it may be by deliberate tuning. The same 20 Hz may be covered by phase distortion in system acoustic resonance, which usually is i set not at lowest, or just resonate in headphones. My fellow has speakers they say capable of 40 kHz but in practice sound has highs deficiency so much that using them he states his ears already do not hear above 8kHz . I can hear 13 kHz but I also feel like 8kHz in his system which is shame because it is audiophile and good brand and as I said states 40 kHz. And still it may be true.

Always wondered about this.

Fortunately phase shifting happens to both channels simultaneously, so humans ears cannot detect difference between two channels.

Makes sense. I always thought of it the same way as resolution. When you compact 20hz to 20Khz you get 44.1 resolution. When you move up to 48 or 88 or 96 or even 192 and higher, it unpacks to a range of hearing well beyond humans. It is because of this, that nay-sayers believe having a higher resolution is meaningless. The way I see it is, 44.1 unpacks like a step-ladder. You still get the frequencies you'd expect, but if you unpack 192 it's like filling in each step with more information so now your step ladder is more like a smooth incline. As I often like to say, "Are you suggesting that a recording of an acoustic guitar at 44.1 sounds just as good as being in the same room live when that person is playing the guitar? There is a reason why recording studios record at higher rates than 44.1 but MASTER down to 44.1. You WILL lose some information.

"so as the amplitude rolls of, the phase starts to shift" - Yes, that is what filters do. Does it also happen e.g. in a tweeter, when the frequency reaches where the tweeter is no longer linear?

Paul (my middle name) : Can you explain overhead in reference to amplifier power, power supply rails etc. Thank you. John Paul in Michigan.

A ruler flat amplitude and phase response of your electronics. And then come the speakers and the room. Oh boy!

This is to me one of your best essays > To my decidedly non musician 72 year old not very good ears ( assume Utube spk. frequency response tests are somewhat accurate ) which quit @ 9 Khz.; explaining maybe there's useable stuff at twice that.

"we are very sensitive to thoes things" absolute nonsense.

My go to headphones are Sennheiser HD25's. They are rated at 16 - 22,000 Hz with .3% distortion. The HD25's are flat, super accurate and high fidelity. All I need. The 60kHz rating is more sell than science.

I worked for GTE Sylvania as a design engineer in an open room (~50 TVs), I can still hear the horizontal scan frequency (15,750 Hz), my wife can not, but when I was at college I tested another student who could hear out to near 100,000 Hz. He complained about the overhead ultrasonic sensors used to detect cars at stoplights especially in a convertible. Below 20 Hz you can hear fine but it sounds more like a series of clicks, at about 20 Hz it starts to sound like a tone.

Oh my goodness! What a simple way to explain! Thank you Paul!!

I’m still struggling to understand what phase shift is exactly..

Where is my ordered power plant 🤷‍♂️

Still not clear why, 18khz is the range of perfect young ears, usually it's much worse especially for older people. So instead of 18 it could easily be 12khz range. And we don't actually hear the phase, we hear the effect of phase on the frequency response, right? So it affects linearity of nearby frequencies the most, and since it's not within the range we hear, the effect on the frequencies we hear won't be, in most situations, distinguishable by human ear and brain because the higher frequencies have less power to make big influence. Please correct me if I'm wrong.

The phase shift explanation accounts for why sound gear designers extend the frequency output of their equipment beyond the hearing capabilities of the human ear. But the limits of our ears to pick up low and high frequencies remains.

Phase shift... brilliant, makes sense, thanks Paul ;-))

Huh

That's a shifty answer

Total garbage. The average 35 year old male has 11 dB hearing loss at 8KHz and probably no hearing above 15KHz. Speaker and headphone manufacturers who claim response to ultrasonics are playing to the technically challenged customer. Be very careful testing your hearing by playing high frequency tones through your headphones. The test should only be done at normal volume. It is tempting to play a 15KHz tone or above and increase the volume until you think you can hear it. This is very dangerous and may cause hearing damage.

It involves sales.

Mmm sensory reaction like our third ear on the back of our skull I kid you not check out Dave Rats video on the inner ear. That is detected my higher frequency’s fascinating stuff

Paul addicted

This man is genius

many of us still have hearing to 28khz and moreover can "sense" up to 32khz

whch is why digital media such as cd sound so lifeless