Audiophiles Rejoice: The Nyquist–Shannon sampling theorem actually proves high-res audio is real and works

[kschmit2]

If a source has ultrasonic components resulting in distortions in the audible range, shouldn't a microphone pick up those distortions and thus the recording of this source - covering an audible range of approx. 20-20,000 Hz - would adequately represent the sound at the microphone position.

Is that understanding correct?

 

[waynel]

If a source has ultrasonic components resulting in distortions in the audible range, shouldn't a microphone pick up those distortions and thus the recording of this source - covering an audible range of approx. 20-20,000 Hz - would adequately represent the sound at the microphone position.

Is that understanding correct?

20 kHz plus a little margin for filtering is sufficient for all recordings for human consumption. If the distortion is in the recording chain then yes ultrasonics could produce distortion products in the audible region. It would have been better to bandlimit before the distortion or reduce the distortion to prevent this.

I don’t agree with 20 Hz as the lower limit as sounds below that frequency can still be felt.

 

[watchnerd]

https://link.medium.com/ktiNtRo1m8

I like ASR, would be interesting to hear what you think about this article.

My phono cartridge already does ultra sonics.

Does that make it hi rez?

 

[mnemonix]

This is simple enough to test with DISTORT app. There can definitely be audible components introduced when nonlinearity is applied to high-level ultrasonic signals.

So assuming all audio systems are non linear to some extent, "Hi-res" audio can only be a bad thing since it introduces the opportunity for unwanted audible components that couldn't exist if the utrasonic frequencies weren't there, whilst retaining nothing extra in the original signal above 20KHz that is actually audible?

 

[LTig]

If a source has ultrasonic components resulting in distortions in the audible range, shouldn't a microphone pick up those distortions and thus the recording of this source - covering an audible range of approx. 20-20,000 Hz - would adequately represent the sound at the microphone position.

Is that understanding correct?

Yes.

 

[pkane]

So assuming all audio systems are non linear to some extent, "Hi-res" audio can only be a bad thing since it introduces the opportunity for unwanted audible components that couldn't exist if the utrasonic frequencies weren't there, whilst retaining nothing extra in the original signal above 20KHz that is actually audible?

Right. I'm not going to hear a 30kHz signal even at 0dBFS, much less at normal audio levels, so why bother sending something through the circuit that can only add to the distortion? Here's that same (THD of -78dB) nonlinearity I posted earlier, except with the 30k and 31k signals at -35dBFS instead of -5dBFS as before. Notice the IMD difference level of -132dB at 1kHz:

 

[waynel]

So assuming all audio systems are non linear to some extent, "Hi-res" audio can only be a bad thing since it introduces the opportunity for unwanted audible components that couldn't exist if the utrasonic frequencies weren't there, whilst retaining nothing extra in the original signal above 20KHz that is actually audible?

Agreed that higher sampling rates than 44.1/48kHz are not a good thing for fidelity. Higher but depth during recording has merit though.

 

[Frank Dernie]

My phono cartridge already does ultra sonics.

Does that make it hi rez?

No it makes it a seismic sensor all of which produce spurious output below their proper transduction bandwidth.

 

[watchnerd]

No it makes it a seismic sensor all of which produce spurious output below their proper transduction bandwidth.

Extra reverb.

 

[Frank Dernie]

Extra reverb.

I presume you are joking?
All output below about 2x natural frequency on a seismic sensor is spurious, not extra reverb.
The extra reverb comes from mechanical and airborne transmission to one side or the other of the transduction system at audible frequencies.

 

[watchnerd]

I presume you are joking?

 

[MRC01]

Besides he left out Whittaker.

Which reconstruction formula are we talking about here: Whittaker-Shannon or Shannon-Whittaker?

 

[MRC01]

... I don’t agree with 20 Hz as the lower limit as sounds below that frequency can still be felt.

And heard! In my last hearing test, taken with headphones, my lower limit was 16 Hz. This is not unusual. We gradually lose some of our HF with age, but we keep LF acuity for life.

 

[skypickle]

Thank you to all who replied. I would have gotten it sooner if you said blue light and green light can never make red light.

 

[xaxxon]

From the article:


So can people really be trained to 'feel' the vibration above 20k? Not a musician / audio engineer, so I'm curious about the validity of this statement. Does that also mean they'll respond to me when I use a dog whistle?

Can you feel a 20khz signal? No. Not sure how you'd feel something higher than that then.

 

[xaxxon]

https://link.medium.com/ktiNtRo1m8

I like ASR, would be interesting to hear what you think about this article.

This guy is really confused what he is arguing and no one is disagreeing with the things that he says that are right.

No one cares about perfect, we just care about "indistinguishable from perfect" with human ears.