Can you really hear the sound details over 20kHz?

[Blumlein 88]

Is'nt there a close up O scope time trace of a "high" speed real music transient using a transducer that can capture it? Like drum rim shot, metal stick on cymbal, or pick on string? Then A>D>A and compare. Scope all the way?

There are 176 and 192 khz recordings of cymbal strikes with microphones from Earthworks with the bandwidth to pick it all up.

I have taken those and down-sampled them or alternatively used a low pass filter at the native sampling rate. The bit above 20 khz (which btw is much less than you think) is not heard. You can't differentiate if it is present or filtered out by listening if you are doing so blind.

Like I've mentioned before, if this far down the road we are still unsure if the extra bandwidth helps or not there are only two possible answers. It might help a small amount for a few people rarely, and the difference is nearly insignificant or it makes no difference to anyone. Recording with 20 khz bandwidth is 99% sufficient or 100 % sufficient. We aren't talking big differences either way. If it makes you happy use 88.2 or 96 khz sampling. Beyond that is only fantasy at best. At worst those more elevated rates will sometimes cause problems you otherwise don't have.

 

[Kal Rubinson]

What they both agreed on was that frequencies above 20KHz may affect frequencies lower down the spectrum.

Do they mean intermodulation?

 

[rdenney]

Do they mean intermodulation?

Interesting article about the use of intermodulation of ultrasonics for nefarious purposes, to create upsetting audible frequencies. Also read the comments, which includes an engineer who conducted similar experiments.

https://hearinghealthmatters.org/he...ntermodulation-distortion-cuba-sonic-attacks/

Rick "noting that CD players and DACs at least attempt to filter out ultrasonics" Denney

 

[Wes]

The real question here is what he means by a "scientific definition"

 

[HiFidFan]

Do they mean intermodulation?

I really don't know, or more accurately, I can't remember. This was around 1996, maybe 1997. I just remember being struck that two respected sound engineers were discussing it as it was the first time I ever heard about the possibility that some frequencies might affect other frequencies. I always heard that any sound above 20KHz was meaningless, but that conversation opened up a new consideration (for me) that, well, maybe it isn't meaningless. I have no idea though and am not advocating either side of that argument.

Please go easy on me, I'm just a hobbyist with no formal background in sound engineering either with software or hardware. I don't even know what intermodulation is

 

[Lambda]

I don't think we can relay hear over 20khz.
I experimented with near ultra sound for data transmission and localization it has some creepy real world applications for cross devices tracking.

Anyways at high levels sounds over 20khz seem to introduce a unpleasant "strange feeling" to me. can totally be placebo!

But maybe the higher frequencys are important for sound localization and only processed subconsciously?
I have read study's saying some trained listeners can reliably detect interaural time difference (or ITD) of ~3µS

in extensive self experiments i was able to detect 1 sample delay at 96Khz so ruffly 10µs.
It is actually not that hard if you now what your listening for.

Maybe it can also cause some intermediation or other distortion?
AM modulated ultrasound seams to be abel to demodulate under some conditions?

 

[Blumlein 88]

I really don't know, or more accurately, I can't remember. This was around 1996, maybe 1997. I just remember being struck that two respected sound engineers were discussing it as it was the first time I ever heard about the possibility that some frequencies might affect other frequencies. I always heard that any sound above 20KHz was meaningless, but that conversation opened up a new consideration (for me) that, well, maybe it isn't meaningless. I have no idea though and am not advocating either side of that argument.

Please go easy on me, I'm just a hobbyist with no formal background in sound engineering either with software or hardware. I don't even know what intermodulation is

Intermodulation is when two frequencies interact to create a another frequency or frequencies that are not in the input signal. It comes from non-linearity in the electronic circuits. You get the original signals plus a small sum and difference signal of those two.

So suppose you had a couple of high level tones at 24 khz and 25 khz. Intermodulation distortion would create a lower level tone at 1 khz. This is the difference in 25 khz-24khz. It would also create tones at 23 khz and 26 khz. And at higher frequencies where these two sum. The one we care about is a tone with nothing audible to mask it created by two inaudible tones at 1 khz which we could hear. It isn't supposed to be there. When you extend the bandwidth well beyond 20 khz you increase the chance of something like this happening.

 

[HiFidFan]

Intermodulation is when two frequencies interact to create a another frequency or frequencies that are not in the input signal. It comes from non-linearity in the electronic circuits. You get the original signals plus a small sum and difference signal of those two.

So suppose you had a couple of high level tones at 24 khz and 25 khz. Intermodulation distortion would create a lower level tone at 1 khz. This is the difference in 25 khz-24khz. It would also create tones at 23 khz and 26 khz. And at higher frequencies where these two sum. The one we care about is a tone with nothing audible to mask it created by two inaudible tones at 1 khz which we could hear. It isn't supposed to be there. When you extend the bandwidth well beyond 20 khz you increase the chance of something like this happening.

Wow, great explanation. Thank you.

What you described may indeed jibe with the conversation between the engineers I heard all those years ago.

So, am I to understand that intermodulation always occurs, naturally? Also, is intermodulation something that can be purposely "shaped", I.E; has utility when mixing sound for example? And finally, if I am understanding you correctly, then ultrasonic frequencies can (do?), through intermodulation, affect frequencies in the audible range?

Edit: I think ? your post #181 answers my questions.

 

[danadam]

But maybe the higher frequencys are important for sound localization and only processed subconsciously?
I have read study's saying some trained listeners can reliably detect interaural time difference (or ITD) of ~3µS

in extensive self experiments i was able to detect 1 sample delay at 96Khz so ruffly 10µs.

You don't need higher frequencies nor hires files for that. Here are 16/44 stereo files with impulses delayed by 5, 10, 15, 20, ... µs: https://hydrogenaud.io/index.php?topic=107570.msg899713#msg899713

Here is abx log of someone detecting the difference in the 10 µs file: https://hydrogenaud.io/index.php?topic=107570.msg899732#msg899732

 

[Thomas_A]

Regarding 20 kHz, am 53 and my limit is 17 kHz today. I can hear the results of a poor DAC though, that may give difference tones in the audible spectrum for HF.

 

[Pio2001]

I was at a recording studio many years ago and this very topic was being discussed by a couple of the engineers. What they both agreed on was that frequencies above 20KHz may affect frequencies lower down the spectrum.

I have no idea if that's true, or has utility in regards to mixing a track, but I thought it was interesting.

In the context of digital recording, it is important to properly reject frequencies above the Nyquist frequency (22050 Hz for CD) at the analog to digital conversion.
In a DAC, it is done in the digital domain through oversampling. But in an ADC, these frequencies must be completely rejected before the A-D conversion with analog filters. That's why it is much easier to work with 88.2 or 96 kHz ADC.
Analog lowpass filtering between 24 kHz (0 dB) and 48 kHz (minus infinite) is easy and effective. Then, resampling to CD quality can be performed in the digital domain with proper brickwall lowpass filters between 20 kHz (0 dB) and 22 kHz (minus infinite), way easier than in the analog domain.

That's why recording engineers might want to use high resolution ADC, while high resolution DACs are not necessary.

 

[Thomas_A]

Besides intermodulation, you might hear tartiini tones IF the SPL is high enough.

 

[escape2]

Download this keys jiggling sound, and enable/disable a low pass filter at 20kHz, and try to hear if there is less "air" in the reproduction.
https://freesound.org/people/Anton/sounds/334/

According to this paper, keys jiggling have 68% of power above 20 kHz.
http://www.tnt-audio.com/casse/life_above_20khz.pdf

I even applied a 16 kHz low pass filter, and still couldn't tell any difference.

 

[Marmus]

So, in the time domain, do we have a trace of the initial hit (impact) of metal stick on cymbal or a close mike on a hard pick on a string, or rim shot is to see what the lead in to the resultant sine wave is from a high response transducer. Then, is that caught in any format? digital or analog? And lastly, if it is, if you manually smooth that, does the perceived sound change? I will try that but all my equipment is in storage right now.

 

[Lambda]

And lastly, if it is, if you manually smooth that,

like with a low pass filter?

does the perceived sound change?

this is the question here.

smoothed out in time the time domain is what a low pass dose.

 

[dc655321]

So, in the time domain, do we have a trace of the initial hit (impact) of metal stick on cymbal or a close mike on a hard pick on a string, or rim shot is to see what the lead in to the resultant sine wave is from a high response transducer. Then, is that caught in any format? digital or analog? And lastly, if it is, if you manually smooth that, does the perceived sound change? I will try that but all my equipment is in storage right now.

Not quite what you are asking for (why abuse a cymbal with a metal stick?), but @Blumlein 88's post here may have of some answers for you.

 

[Marmus]

YES, that was perfect. Thanks

 

[Marmus]

Ok, maybe I should learn of the horrors of cymbal abuse too. But as that thread states, folks (me) could think there is a sharp transient there.

 

[Wombat]

Not much there, even at half speed.

 

[Wes]