Shouldn't we upgrade the 20-20 audible range ?!

[kevinh]

71 yrs old, listen to music can't detect dounds above 12-14k. Funny thing if i hear recordings i have listened to for 50 + years they sound the same to my brain. So i don't care.

 

[oleg87]

Hi-res recording means nada if there is no equipment to play it. Or mics to actually record it. Or mix-room engineers who actually care about it.
That does not sound like much of a vision to me. And the skills/equipment are still missing in 202x.

Hi-res during the 80's was a pretty empty bottle. Still is, largely. Not a big wonder that it does not sell very well (actually, I'd say it sells amazingly well for an "empty bottle").

You don't need a mic to test your theories. Synthesizers, distortion applied in the mix, speeding up regular music until some significant amount of it is pushed to ultrasonic frequencies, etc.

 

[Mnyb]

My humble system goes from 15 hz to 36kHz -6dB / ( 20-24kz -3db) . Kef LS60 and dual KC92 subs .
speaker accepts 24/192 but are probably filtered due the typical tweeter resonance all hard domes have somewhere.

so my 24/96 records are fine and honestly that’s where I think the buck stops 24/96 or 24/48 records, but make em 24/96 just in case
and I buy no more than 24/96 if available so I skip the 24/192 versions to save some money and space .
24/96 has the slush margin to contain every relevant recording ever made .

and it’s not a first order problem, there are 1000’s of problems to fix before this moves up to being important it’s really all about priorities .
for example I possibly need to rebuild my whole room re the very real and blatantly audible issues in the typical audible range … first things first

 

[lashto]

You really seem to lack some knowledge here. There have been commercial hi rez releases that 'go well outside ' 20-20k, released on DVDA, SACD, and BluRay since around the year 2000*. You can verify the existence of 'signal' beyond 20kHz with any waveform viewer that has a spectral analysis function.

*not all such 'hi rez' releases actually have any content above 20kHz ....a rather amusingly frequent number of them e..g on HDtracks, do not ....which is one of the signs of marketing hype in that realm, aimed directly at true believers like lashto. But there certainlydo exist releases which were recorded and released at 'hi rez' rates.

That's pretty much the issue. There is quite a lot of 'hires' lately but those are just claims.
You are welcome to post some samples that you verified to have testable ultra/infra-sonic content.

"Super tweeters' have been around for decades too. Subs that 'go' below 20Hz are hardly uncommon.

yes super tweeters exist. They just cost thousands. And you have to rely on manufacturer claims cause noone ever tested those specs. Again, welcome to post an example that anyone can buy and use.

The super-sub situation looks a bit better. At least there is some reliable 3rd party data. They just cost thousands too ... plus the size.

Yet will you go on spouting about how important it is for it to happen?

I want it to happen. Is it important? Maybe .. maybe not ..
Is there any data that clearly says it's not? Not AFAIK. But there is a lot of data that says it is important.
Other than that, it is just an individual decision

 

[Mnyb]

You really seem to lack some knowledge here. There have been commercial hi rez releases that 'go well outside ' 20-20k, released on DVDA, SACD, and BluRay since around the year 2000*. You can verify the existence of 'signal' beyond 20kHz with any waveform viewer that has a spectral analysis function.



*not all such 'hi rez' releases actually have any content above 20kHz ....a rather amusingly frequent number of them e..g on HDtracks, do not ....which is one of the signs of marketing hype in that realm, aimed directly at true believers like lashto. But there certainlydo exist releases which were recorded and released at 'hi rez' rates.




"Super tweeters' have been around for decades too. Subs that 'go' below 20Hz are hardly uncommon.





Yet will you go on spouting about how important it is for it to happen?

I have hundreds of DVDA recordings, the ones from AIX are true HiRez not just old upsampled 70’s rock .
the mics does not have infinite slope after 20k it’s more like a slow rolloff . So it’s recorded and there .

the problem is monitoring the stuff as no one actually hears this with music ? How to adjust levels and mix things you can’t hear .
what AIX and 2L can do and probably does . Is keeping the production chain as linear as possible and even this content gets captured and do exist on their records.

 

[ahofer]

Is there any data that clearly says it's not? Not AFAIK. But there is a lot of data that says it is important.

First sentence is asking to prove a negative, Russell’s Teapot. However, audiology has very strong failure to disconfirm the null hypothesis at the sound pressure levels present in music.

If the evidence presented herein is what you are suggesting in the second sentence, I beg to differ.

 

[ZolaIII]

@lashto
1) there are true high depth/highs recordings but very little of them.
2) I participated in a ABX survey cuple years back when my hearing whose 1 KHz higher.
3) used good headphones as a trained listener and cheated doing spectrograps right after the testing and comprehensive analysis why I thought what I thought.

Conclusion whose that depth of 24 bit Vs 16 bit without dithering had a subtle difference and not always for sounding better (depending on original material). Listening critically and very loud analyticly fresh in the morning after a good sleep and not longer than 30 min for a session. High frequency range didn't had any impact nor it could as there were no additional reverberation (open headphones).
If you are interested in infra sound effects on materials and their properties start with Nikola Teslas work in that area. Of course it's not about hearing or using microwave would be painful experience. Theoretically it's possible that infra special refractions could impact and some higher end waves in hear able range but again it wouldn't have meaningful impact on how you hear them. Just remember how second harmonic of 10 KHz is 20 KHz and that's about it.

 

[Timcognito]

interesting idea. Did you also spend some time thinking about that test?
I actually did, here are some 'findings':

• there are no reliable commercial recordings that go outside 20-20. Only way to test is to buy a mic and record your own music. Mic alone costs several thousands.
• I have no audio equipment that reliably plays outside 20-20. And there isn't that much in the consumer area, particularly in the speaker area (although some do claim). That would be another multi-thousand investment in pro level übersubs, super-tweeters etc.. which I'll have to somehow test myself because that kind of testing is done by ~noone.
• I am just a sample of one .. one of those old guys who can barely hear 15-16 kHz. Not exactly the most helpful/relevant test.

So .. not gonna happen soon. Neither for me nor for anyone else.

Interesting but not the most helpful but a relevant thread. Nothing is heard above 20 KHz except by a few 20 something gifted listeners, cats and expensive microphones.

Thanks for a great discussion on musical insturments

 

[Cbdb2]

Ultrasonics are noise that we are incapable of hearing. Just like ultraviolet light is outside the range our eyes can see. Do you want a TV to produce infrared and ultraviolet ranges to be considered good? How about x-rays? Can we leave that out?

Reproducing and amplifying ultrasonic frequencies will accomplish nothing other than bothering the dogs and other small animals within hearing range.



A dog whistle is not a musical instrument.

The odd instruments have harmonics over 20khz. Mostly cymbals and percussion.

Percussion Frequencies Part 2 - Cymbals - Musical U

Learn to recognise audio frequencies by ear using these hi-hat, ride and crash cymbal sounds as your training material. Great for EQ and production ear training.

www.musical-u.com

 

[MRC01]

Upgrading the CD standard from 44.1 kHz to the DVD standard of 48 kHz would provide measurable benefits that could be audible to some people, with some kinds of sounds. This can be seen in just about every DAC Amir has reviewed here. At 44.1 kHz the transition band is so narrow that low pass filters either (A) fully attenuate by Nyquist, but also attenuate in the passband, or (B) have flat passband response but don't fully attenuate until 24 kHz, which can pass ultrasonic noise. Any noise would be at very low levels, but it could trigger intermodulation in the midrange, and even if that is not a real concern, shouldn't we use a standard that doesn't require engineering kludges?

It doesn't take much more to get it right. At 48 kHz sampling the transition band is just wide enough that many of the good DACs reviewed here have a proper filter - no passband attenuation and full attenuation by Nyquist.

With bit depth, going from 16 to 24 bit would be another good step, since the difference can be audible with some kinds of sounds.

That said, one can say that CD quality 44-16 is "mostly" transparent for "most" music for "most" people. But 48-24 should be fully transparent both musically and perceptually with no engineering compromises. And it would mean the same single standard for both music and movies (DVD/BluRay).

 

[ahofer]

My Harbeths have a super tweeter. I was just looking at the specs for the Ascend speakers. They claim the RAAL ribbon models go up to 27Khz.

 

[Mnyb]

Upgrading the CD standard from 44.1 kHz to the DVD standard of 48 kHz would provide measurable benefits that could be audible to some people, with some kinds of sounds. This can be seen in just about every DAC Amir has reviewed here. At 44.1 kHz the transition band is so narrow that low pass filters either (A) fully attenuate by Nyquist, but also attenuate in the passband, or (B) have flat passband response but don't fully attenuate until 24 kHz, which can pass ultrasonic noise. Any noise would be at very low levels, but it could trigger intermodulation in the midrange, and even if that is not a real concern, shouldn't we use a standard that doesn't require engineering kludges?

It doesn't take much more to get it right. At 48 kHz sampling the transition band is just wide enough that many of the good DACs reviewed here have a proper filter - no passband attenuation and full attenuation by Nyquist.

With bit depth, going from 16 to 24 bit would be another good step, since the difference can be audible with some kinds of sounds.

That said, one can say that CD quality 44-16 is "mostly" transparent for "most" music for "most" people. But 48-24 should be fullim transparent both musically and perceptually with no engineering compromises. And it would mean the same single standard for both music and movies (DVD/BluRay).

I agree in principle , but I don’t have to much fate in recording engineers or maker of playback systems so I think 24/96 with allow margins for sloppy filters in DAW plugins other diverse issues in the studio and bad playback equipment such as creative filter options in DAC .
even 48 cuts very close more margin could help some corner cases .

My personal experiments with downsampling real Hirez record to CD quality and AB test them has convinced me that your in fact are correct.
but I like more engineering margins

 

[Cbdb2]

We surely do . IMO, audible or not is not an end-all argument. But that would be a somewhat offtopic/parallel discussion

If you can't hear it why record it?

Here's my 'bottomline': those infra/ultra sounds are present in a live performance and I want them in my home too. A 'simple' accuracy wish.

What ultra sounds are present? Do they actualy reach the audiance? Do you also want all the reverb and noise present at a live performance in your home?

Whether that will make a difference or not, I don't know. Almost nooone knows, because the 20-20 limitation usually starts at the microphone level ... most of the time, that stuff is not even recorded. But apparetly everyone-but-me knows it is not important. How exactly?!

Did anyone actually try?

Yes. Every recording engineer who's ever recorded cymbals knows. There are a few mics that reach over 20khz in most recording studios.

 

[Mnyb]

My Harbeths have a super tweeter. I was just looking at the specs for the Ascend speakers. They claim the RAAL ribbon models go up to 27Khz.

Super tweeters may introduce real audible problems instead . It’s proper engineering to have the main tweeters cover the actual tweeter range and more if you fancy .

making crossovers that high and integrate dual tweeters migth be acoustically impossible to do properly, the distance between them needs to very very close , possibly closer than their actual size , placing a super tweeter on top of the cabinet is just an acoustic mess , interference and comb filtering everywhere ?

 

[kemmler3D]

Is there any data that clearly says it's not?

Yes, probably.

1. Find at least one recording with good ultrasonic content. Rare, but they exist.
2. Look up the masking thresholds for 20+ Khz content. (the data that may say it's not.)
3. Compare the levels of normal vs. ultrasonic content in the recording.

This will tell you whether (if you had the right equipment) the ultrasonics are definitely, definitely not, or maybe audible.

I think best case, people who can hear ultrasonics might detect a very slight difference between the attacks of certain percussion instruments with / without ultrasonics. In any other case I think the normal audio would mask it... and I'm not sure even then.

You have to remember that with everything except synthesizers, most of the energy is in lower frequencies, so even if there are ultrasonics present, it is harder to hear them than if they were alone. And they actually never appear alone in music.

 

[Taiga]

This is a section of the galaxy cluster Abell 426 in Perseus that I imaged from my backyard observatory. It is about 250 million light-years away. The reddish galaxy at the top of the picture is NGC 1275. While this object has an oval shape, we are actually looking at a spiral galaxy nearly edge-on as it emits gaseous filaments in multiple directions. NGC 1275 is also known as Perseus A and is a strong source of radio waves as well as X-rays from a black hole at its nucleus.

A 2003 study suggests that NGC 1275 may be generating sound waves at a period of 9.6 million years. [https://academic.oup.com/mnras/article/344/3/L43/1015198] Wikipedia states that “no human will actually hear the note, because its time period between oscillations is 9.6 million years, which is 57 octaves below the keys in the middle of a piano.” [https://en.wikipedia.org/wiki/Perseus_Cluster] Rumor has it that the famed singer Pavarotti suffered two hernias during his career trying to reach a bass note this low.

 

[Killingbeans]

There was the exact same reaction when Apple introduced the retina screens: laughing, go away, nooone needs it and so on and on ... 'surprinsingly', the tune changed after they sold millions of those devices.

The difference being that the first retina displays weren't pushing into the ridiculous. Every Tom, Dick and Harry could tell that the experience was "better" in the sense that the individual pixels were no longer visible. The only reason people said it wasn't needed, was because it didn't change anything related to functionality.

Try presenting an audio experience with ruler flat near-DC-to-30kHz all the way from the microphones to the eardrums, and I'll almost guarantee that any feedback you'll get from listeners, praising its superiority, is purely formed by expectation bias.

And just for the sake of argument, let's assume that it's all unnecessary .. it still does not explain the big discrepancy between audio and ~all other industries/domains.
Everyone else is still pushing for more, even if it's not needed and/or there are no clear studies. I am also not aware of any benefits of 1000 PPI screens or 20000 DPI mice but still, everyone is pushing those limits.
In audio, everyone seems to be pushing for ... ~nothing (i.e. nothing outside this 50 years old tech limitation of 20-20)

The industry doesn't even stand still. It seems to be actively resisting true innovation.

Products, that are aimed at fixing actual issues, are more often than not being branded as "sterile/soulless" or accused of removing all of the fun from the hobby.

There are far, far greater problems to focus on than ultrasonics, and yet people seem to love the idea of its indispensability.

Modern DAC chips support files with a 768kHz sample rate. I mean, if that's not unnecessary, I don't know what is.

 

[danadam]

... but don't fully attenuate until 24 kHz, which can pass ultrasonic noise. Any noise would be at very low levels, but it could trigger intermodulation in the midrange,

At 48 kHz sampling the transition band is just wide enough that many of the good DACs reviewed here have a proper filter - no passband attenuation and full attenuation by Nyquist.

Why would not fully attenuated 22-24 kHz band from 44.1 kHz recording be a problem but the same band from 48 kHz recording be ok? Sure, one is the mirrored image and the other is "legitimate" content, but the downstream hardware doesn't know that, right?

 

[oleg87]

There was the exact same reaction when Apple introduced the retina screens: laughing, go away, nooone needs it and so on and on ... 'surprinsingly', the tune changed after they sold millions of those devices.

Now you're just making things up. No remotely tech-savvy person was saying this. And obviously pre-iPhone 4 screens were nowhere close to the resolving limits of the eye in typical usage and no sane person would have argued they were.
High DPI screens were a massive improvement and actually made mobile browsers usable, especially when we were talking about iPhone 4-sized screens. And the reason I know off the top of my head that the iPhone 4 was the first model with a retina screen is specifically because that's what pushed me to buy one.

 

[MRC01]

Why would not fully attenuated 22-24 kHz band from 44.1 kHz recording be a problem but the same band from 48 kHz recording be ok? Sure, one is the mirrored image and the other is "legitimate" content, but the downstream hardware doesn't know that, right?

Suppose the passband always ends at 20 kHz. At 44.1 kHz sampling the proper stopband (Nyquist) is 22,050 Hz, so the transition band is only 2,050 Hz wide. At 48 kHz sampling the stopband is 24 kHz so the transition band 4,000 Hz wide - about twice as wide.

At 44.1 kHz, most DAC filters can't attenuate from 0 to -infinity in such a narrow transition band. So they must widen the transition band. If they widen the left/lower side, it creeps into the passband and we get passband attenuation. If they widen the right/upper side, the filter can leak ultrasonic frequencies. That is any frequency > 22,050 Hz (Nyquist) is by definition noise, so the filter will allow some ultrasonic noise to leak through.

This takes us to the engineering kludge I referred to earlier: most DAC filters at 44.1 kHz sampling fully attenuate by 24,100 Hz. This is no coincidence, because aliases mirror around Nyquist, and 22,050 is exactly half-way between 20,000 and 24,100. So all the ultrasonic noise the filter allows to leak through is between 20,000 and 22,050 Hz, above the passband and inaudible to most people. However, those ultrasonic noise frequencies are only a few kHz apart, which puts any intermodulation tones smack-dab in the midrange where our hearing is most sensitive.

Sampling at 48 kHz, no kludge is needed. The transition band is wide enough that a proper filter can be implemented.

Pragmatically, the kludge should not be a problem because the levels of this noise should be so low. That's why this kludge is so common. But why not do it right? Why use an audio standard that requires this engineering kludge of improper filtering? It only takes an extra 10%, sampling at 48 kHz, to do it right. Surely over the 40 years since the CD standard was introduced, technology has advanced to the point where using 10% more data is not an issue. OK call it 1.5 * 1.1 = 1.65x more data if we bump the bit depth to 24.