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

[pkane]

I prefer great engineering.

I do too, but I see no reason to exceed FR requirements by a huge margin when it brings no benefit. Good engineering isn’t about exceeding all possible specs, it’s about prioritizing specs and ensuring they are met at an acceptable cost.

Pushing the envelope to go beyond what’s necessary is admirable when it solves some problems that were previously unsolved or were hard to solve. Building circuits that handle 100kHz, MHz, or even GHz range is done routinely, this isn’t new. So what is the purpose of over-designing audio circuits where there is no actual benefit to the extra effort and cost?

 

[dlaloum]

I do too, but I see no reason to exceed FR requirements by a huge margin when it brings no benefit. Good engineering isn’t about exceeding all possible specs, it’s about prioritizing specs and ensuring they are met at an acceptable cost.

Pushing the envelope to go beyond what’s necessary is admirable when it solves some problems that were previously unsolved or were hard to solve. Building circuits that handle 100kHz, MHz, or even GHz range is done routinely, this isn’t new. So what is the purpose of over-designing audio circuits where there is no actual benefit to the extra effort and cost?

Sometimes, it is about overcoming issues that are not well defined/understood, and example is transient capabilities...

A circuit with bandwidth up in the 100KHz or higher range, will have no issues handling transients and strange waveforms in the audible range...

A prime example is the ability to pass a square wave.... which can be mathematically characterized as a series of related sine waves, ranging in frequency up to infinity (ie: a square wave is never perfect without infinite bandwidth) - the wider the amps bandwidth, the "better" the square wave it can reproduce. (and yeah you can easily reach the limits of typical measuring instruments this way!)

So bandwidth and square waves, can be used as a "fill in" to mark for things like slew rate, rise time, transient capability... these are aspects that definitely DO impact on audible performance.

But what level is required? what level of performance is audible? where are those audibility thresholds?

We really don't have a lot there in the general set of measurements published for components and speakers that tell us anything at all about such aspects. The closest we get to it, is those occasional amps that advertise their wide bandwidth capabilities...

 

[pkane]

Sometimes, it is about overcoming issues that are not well defined/understood, and example is transient capabilities...

A circuit with bandwidth up in the 100KHz or higher range, will have no issues handling transients and strange waveforms in the audible range...

A prime example is the ability to pass a square wave.... which can be mathematically characterized as a series of related sine waves, ranging in frequency up to infinity (ie: a square wave is never perfect without infinite bandwidth) - the wider the amps bandwidth, the "better" the square wave it can reproduce. (and yeah you can easily reach the limits of typical measuring instruments this way!)

So bandwidth and square waves, can be used as a "fill in" to mark for things like slew rate, rise time, transient capability... these are aspects that definitely DO impact on audible performance.

But what level is required? what level of performance is audible? where are those audibility thresholds?

We really don't have a lot there in the general set of measurements published for components and speakers that tell us anything at all about such aspects. The closest we get to it, is those occasional amps that advertise their wide bandwidth capabilities...

Reproducing high frequencies isn’t overcoming issues — as I said, non-audio circuits handle GHz range without an issue. Adding a carburetor to an EV, while may be challenging, isn’t solving any real problems, just complicating the design and increasing cost.

 

[dlaloum]

Reproducing high frequencies isn’t overcoming issues — as I said, non-audio circuits handle GHz range without an issue. Adding a carburetor to an EV, while may be challenging, isn’t solving any real problems, just complicating the design and increasing cost.

My point was that the high frequency bandwidth was (is) being used as an alternate for something that is not commonly measured or reported on - it is not itself that goal... the goal being transient response/slew rate, the amps response to the "Time" parameter...

But the bandwidth, provides an indication of capability in that area, as a high bandwidth design necessarily has high slew rate / transient response.

So sometimes an apparently "irrelevant" parameter (who needs to be able to reproduce frequencies in those ranges!?!) - is there as an indicator of something important - we do need components to be able to respond very quickly to transients and to maintain time relationships between components of the signal...
People tend to ignore slew rate / transient measures (hard to understand), and focus on bandwidth measures (easy to understand) - so marketing follows suit.

And yes it is a non issue from a design perspective - there have been designs capable of ultrasonics for decades (since the dawn of audio really!) - it becomes more of an issue from an engineering perspective (and design engineering) - and then there is the issue of additional undesirable ultrasonic "noise" potentially generated by current digital and PWM/ClassD designs - and whether they need to be filtered (potentially to protect tweeters?) etc...

Of course in a perfect world we would not need to provide these alternate measurements - and would provide more directly relevant specifications.... but first the market would need to be educated (and not for lack of trying, the last 50+ years!)... and the marketing departments brought into line.

 

[pablolie]

For recording/processing purposes 32 bits would be great! - but for reproduction/playback purposes, 16/44 hits the spot.

Seems for recording 24/192 is kind of a standard when it isn't even higher. Doesn't mean the recording is any good, just may mean you have a high rez baseline that may be crap.

Nor does it in any way mean you can't easily play back things at over 16/44 and beyond 20Hz to 20kHz these days.

The poor horse in this thread has been very dead for a while and flogged into minced little pieces....

 

[EJ3]

But the bandwidth, provides an indication of capability in that area, as a high bandwidth design necessarily has high slew rate / transient response.

What happened to SLEW RATE, DAMPENING FACTOR & TRANSIENT RESPONSE being part of the STANDARD measurement suite?
DAMPENING FACTOR is still seen from time to time, but the other 2 seem to have lost favor totally???

 

[dlaloum]

What happened to SLEW RATE, DAMPENING FACTOR & TRANSIENT RESPONSE being part of the STANDARD measurement suite?
DAMPENING FACTOR is still seen from time to time, but the other 2 seem to have lost favor totally???

Exactly!

But if the manufacturer specs include a wide bandwidth (say 100Khz) - then you know that the slew rate & transient response will be OK (within the audible range)

 

[Sal1950]

Of course in a perfect world we would not need to provide these alternate measurements - and would provide more directly relevant specifications.... but first the market would need to be educated (and not for lack of trying, the last 50+ years!)... and the marketing departments brought into line.

What happened to SLEW RATE, DAMPENING FACTOR & TRANSIENT RESPONSE being part of the STANDARD measurement suite?
DAMPENING FACTOR is still seen from time to time, but the other 2 seem to have lost favor totally???

The poor horse in this thread has been very dead for a while and flogged into minced little pieces....

A great line of discussion and all very relevant my friends.
And yet, I'm constantly involved in some quite heated debates with a big number of OUR members who
tell me over and over that great engineering in our electronic path doesn't matter. There is no need to have a
system that's "straight wire with gain" like from source to speaker. It can distort the source in what-ever way and
as long as they like the sound, that's OK. To hell with High Fidelity.

 

[solderdude]

some FOMO some DOHM .. the usual

(Desire Of Having More)

I am not worried about either... nor Euphonics.
Makes life a lot easier and cheaper.
In my younger years I always wanted 'better' until I went to work in a high-end audio store...
Am not fussed now, not even when I find better speakers/headphones (at considerable higher cost) and am happy with what I own.

 

[Brian Hall]

You're still missing the point of the practical implications of setting immovable goal posts.

I want immovable goal posts if they are already in the perfect position to accomplish the desired goal.

Moving them around just because it is technically possible is not beneficial to the end user of the goal posts, only to the person who is being paid to move them.

 

[Killingbeans]

Sometimes, it is about overcoming issues that are not well defined/understood, and example is transient capabilities...

A circuit with bandwidth up in the 100KHz or higher range, will have no issues handling transients and strange waveforms in the audible range...

A prime example is the ability to pass a square wave.... which can be mathematically characterized as a series of related sine waves, ranging in frequency up to infinity (ie: a square wave is never perfect without infinite bandwidth) - the wider the amps bandwidth, the "better" the square wave it can reproduce. (and yeah you can easily reach the limits of typical measuring instruments this way!)

So bandwidth and square waves, can be used as a "fill in" to mark for things like slew rate, rise time, transient capability... these are aspects that definitely DO impact on audible performance.

But what level is required? what level of performance is audible? where are those audibility thresholds?

We really don't have a lot there in the general set of measurements published for components and speakers that tell us anything at all about such aspects. The closest we get to it, is those occasional amps that advertise their wide bandwidth capabilities...

There's just one problem.

No matter how much you avoid having any low-pass filter below 100kHz anywhere in your playback chain, there's one you can't escape; Human hearing.

Super wideband audio playback might give really pretty spiky impulse resonses and lovely boxy square waves, but your hearing doesn't care.

Bandwidth is bandwidth. If you can't hear it, you can't hear it.

 

[RandomEar]

My point was that the high frequency bandwidth was (is) being used as an alternate for something that is not commonly measured or reported on - it is not itself that goal... the goal being transient response/slew rate, the amps response to the "Time" parameter...

But the bandwidth, provides an indication of capability in that area, as a high bandwidth design necessarily has high slew rate / transient response.

So sometimes an apparently "irrelevant" parameter (who needs to be able to reproduce frequencies in those ranges!?!) - is there as an indicator of something important - we do need components to be able to respond very quickly to transients and to maintain time relationships between components of the signal...
People tend to ignore slew rate / transient measures (hard to understand), and focus on bandwidth measures (easy to understand) - so marketing follows suit.

And yes it is a non issue from a design perspective - there have been designs capable of ultrasonics for decades (since the dawn of audio really!) - it becomes more of an issue from an engineering perspective (and design engineering) - and then there is the issue of additional undesirable ultrasonic "noise" potentially generated by current digital and PWM/ClassD designs - and whether they need to be filtered (potentially to protect tweeters?) etc...

Of course in a perfect world we would not need to provide these alternate measurements - and would provide more directly relevant specifications.... but first the market would need to be educated (and not for lack of trying, the last 50+ years!)... and the marketing departments brought into line.

It's the same thing, though. "Steep transients" are just sounds with a relevant amount of HF data. That's it. There's no magic to it. If a transient contains a relevant amount of 40 kHz data and assuming that it was recorded using a high bandwidth mic and ADC, you can lowpass that recording to ~22 kHz and play both versions on a high bandwidth audio system. Your ears won't be able to differentiate the two, if you can't hear past ~22 kHz (or 18, 16 or 15 kHz, in case you're older). The the hair cells in your ear don't react to sounds that are as "fast" as 40 kHz containing transients, because their inertia and stiffness impose a physical limit on their movement speed. At least that's my understanding of the matter.

 

[pkane]

A prime example is the ability to pass a square wave.... which can be mathematically characterized as a series of related sine waves, ranging in frequency up to infinity (ie: a square wave is never perfect without infinite bandwidth) - the wider the amps bandwidth, the "better" the square wave it can reproduce. (and yeah you can easily reach the limits of typical measuring instruments this way!)

Umm. Why exactly would anyone want a “better” square wave by including more of the ultrasound harmonics? So it looks better on a scope display? It’s like arguing that we need higher bandwidth circuits... to get higher bandwidth.

 

[ElJaimito]

For recording/processing purposes 32 bits would be great! - but for reproduction/playback purposes, 16/44 hits the spot.

Well we all -- by now, even Pro Tools users -- have 32-bit floating point as our default recording settings, no? But good luck with a 32-bit recording interface, that is way below the basic noise level of the universe, where we live, anyway...

 

[EJ3]

Exactly!

But if the manufacturer specs include a wide bandwidth (say 100Khz) - then you know that the slew rate & transient response will be OK (within the audible range)

IT's still nice to have DAMPENING FACTOR AND SLEW rate, a lot can be inferred from those 2.
And different years of the same make/model amp cam have enough differences particularly in DAMPEMING FACTOR) that can make a difference in whether or not you are going to use it as a subwoofer amp for your custom made subs). I believe that SLEW RATE is also a factor in the low end and how well the DAMPANING FACTOR works in controlling the starting & stopping part of the sub-woofers movement.
Lashto, while he is concerned about the deep low end, seems even more concerned about the high end (where both the SLEW RATE AND TRANSIEN RESPONSE are IMHO more important than the DAMPENING FACTOR).
I could be wrong, I have been wrong before. If you don't believe me, just ask my wife.

The reason that I put out the information on my Dahlquist M-905's is that the claimed FR is 40 Hz- 24 KHz (no parameters) and the test revealed 26 Hz-20 KHz +-2 DB.
But why wasn't isn't the 20 KHz-24 KHz that the manufacturer clamed ever tested when the claim is there?
Whether we can hear it or not, it would be nice for it to be tested (even if it's just to hold the manufacturers toes to the fire).
When Amir tests an AMP or a DAC, the test doesn't stop at 20 KHz!
IF a speaker manufacturer makes the claim of a higher than 20 KHz response, it would be nice if the speaker will actually get to the specified frequency and with what parameters.
Perhaps some of my friends whose ears are less compromised than mine, could here it (and that might be why they may think that my speaker sound like hell (when I thing that they sound great) because I haven't attenuated the very upper part of the speaker response that I don't hear.
Or maybe they think my speakers are awesome because they hear the response that goes that high.
But without knowing if anything is actually up there or what's up there at that frequency, I can't make any relative adjustments to what someone else might be hearing at that frequency.
If they are at my home and they find the upper frequency annoying, I could possibly make adjustments that they could hear & make it better or eliminate it.

 

[lashto]

Why would one want 100kHz if it isn't in any recording and when it is it will be at very, very low SPL.

to paraphrase, that is what I would call The Circle of the Deprecated 70's

The circle 'starts' with some not-so-true ans seriously deprecated assumptions like "humans only hear 20-20" or "music/sound are ears-only experiences" ...
.. continues with "therefore we don't need mics or recordings beyond-20-20"
.. goes into the music studio which truncates 20-20 because it's easy and they 'know' that "nobody needs beyond-20-20"
.. reaches the audio industry as "there is nothing beyond-20-20, why would we build speakers/etc that go there"
.. enters the home of the customer who cannot even test anything by himself (cause there is barely any test material or gear.)
.. and then it goes back to the 70s

And so everyone happily lives everafter .. thoroughly enjoying the limits of the deprecated 70's tech

 

[pablolie]

... .. thouroughly enjoying the limits of the deprecated 70's tech

Clearly that's what you have chosen to do for yourself. To each, their own preferences...

 

[lashto]

There's just one problem.
...
If you can't hear it, you can't hear it.

I did not follow the "we can't hear it" parallel discussion.
This thread is about what we can hear: i.e. at the very least 8-28. And the very real/measurable/proven effects of 'inaudible'.
May I ask what exactly is the "can't hear it problem" and what makes it ontopic ?

And a more generic question for the 'opposition' in this thread. Your 'sneezing' at everything new was and is very audible. You already made it clear that you are against any possible form of progress or change (in terms of FR/20-20). Fine with me, anyone has the right to be as negative and anti as they wish. And good luck with that!
But just curious: is there anything you are actually FOR?

 

[solderdude]

Aside from CD there is no sharp upper limit to 20kHz which is your complaint.
O.K.... FM tuner was limited to around 16kHz.
The bottom limit is mostly caused by transducers (speakers)

Studio masters are rarely recorded at 44kHz but more often at 96/24. Only the masters sent to CD press are brick-walled to 20kHz (down sampling).

Most speakers do not 'brickwall' above 20kHz. There are some that start to roll-off above 16kHz.

The fact that most testers/reviewers don't measure below 20Hz and above 20kHz is another matter. Also the fact that some specs show 20-20kHz limits also is another matter.

Below 100Hz or so the 'body' also senses SPL, but only at louder levels. For this you need speakers and they are not very good below 20Hz at high SPL (which is what is needed).
Headphones/earphones can reach lower but ... no input from the body.

You assume there is 'something' in the auditory system that takes over above say ... 26kHz (because you want 100kHz) but there is no evidence what would do that and how that enters the brain at the low levels that are present in music.

You keep repeating the same mantra ... audio is not limited between 20-20kHz.
Yes, for young people with excellent hearing > 20kHz is possible ... with test tones at high SPL.
<20 Hz (when it is a pure sine which it never is) is detectable at high SPL even with the ears blocked.

For audio the most important question is does it contribute to experiencing music ?
20Hz (-0.5dB) seems enough and makes a 16Hz organ note still audible.
20kHz (-0.5dB) is more than enough for music enjoyment.

 

[lashto]

Clearly that's what you have chosen to do for yourself. To each, their own preferences...

hmm guess my post wasn't clear enough ... one major point is that the beyond-20-20 option does not exist for 99.99x% of people.
There is no 20-20-or-not choice, anywhere.

There is only a group of researchers/experts who managed to DYI their way into beyond-20-20. And they came back saying "go for it, it is wonderful"...