There are two situations where higher sample rate makes sense to me. SoundFX where you slow down a bird song to sound like Godzilla.
Plugins that don’t use internal oversampling. Some Guitar amp plugins, sound radical different at double sample rate. Im happy at 44.1 but whatever You feel best, is the best for You.
It is interesting the Nyquist-Shannon (and the myriad others like Whittaker and Kotelnikov, who came to the same result) Theorem is bandied about with much abandon when discussing sample rates.
The most significant thing about it is that it contains an integral from minus to plus infinity. That is what is called doubly infinite and therefore only applies to sounds which are continuous and infinite both backward and forward in time. Guess what? No sound in nature is doubly infinite (they abruptly start somewhere = single ended) and don’t last very long (not infinite).
The counter theorem?
The Cheung–Marks theorem undermines the absolute interpretation of the Nyquist-Shannon theorem that many seem to apply to real world audio recording.
From Wikipedia:
Generally, the Cheung–Marks theorem shows the sampling theorem becomes ill-posed when the area (integral) of the squared magnitude of the interpolation function over all time is > not finite> . “While the generalized sampling concept is relatively straightforward, the reconstruction is not always feasible because of potential instabilities.”
and:
Although true in the absence of noise, many of the expansions proposed by Shannon become ill-posed. An arbitrarily small amount of noise on the data renders restoration unstable. Such sampling expansions are not useful in practice since sampling noise, such as quantization noise, rules out stable interpolation and therefore any practical use.
In other words, it ain’t perfect in an imperfect world!
So, if it isn’t perfect, how much imperfection can we get away with?
Unfortunately, it seems hard to get past the blind acceptance of the N-S theorem being infallible in practice to find research that is willing to examine the correlation of the levels of imperfection ‘inherent’ in various types of real-world sounds with what sample rate is sufficient for each of them.
I would imagine that fairly constant synthesiser sounds would be the closest to the non-existant ideal to which the N-S theorem ONLY applies. But what of acoustic instruments recorded in free air, where non-linearities (non-infinites) occur in abundance as all the sounds decay rapidly and bounce around inside and outside instruments.
And what of FX that contain a lot of non-infinite processing?
Relevance to DAW use?
Now, just to be clear, it is a given that one simple sample rate conversion does not seem to be enough for most people to be able to tell the difference in any of the A-B tests. But that finding is ONLY of relevance to what format the delivered material is provided.
Also, the issue has nothing to do with ranges of human hearing per se, other than making sure that it is covered.
For DAW use, the issue is about the myriad non-linear processes (FX like compressors and reverb) that may be applied along the journey from recording to that final deliverable.
The famed mastering engineer, Bob Katz, while stating in his Mastering Audio that he has built digital filters that make 44.1k sound good, in Paul Gilreath’s The Guide to MIDI Orchestration he recommends recording at 96k, or upsampling to that as soon as possible, and keeping all processing at that until ready to downsample at the very end to the target deliverable sample rate. That was in 2006, when 96k was affordable, but I wonder if he might have revised that since, as 192k is much more affordable now.
The real clincher for me in going 192k was that Universal Audio upsampled most of their emulation plugins to that internally, which, given that they don’t brag about it because it basically undermines the number of plugins that their cards can run, would seem to indicate that they needed to do it to guarantee fidelity of their plugins.
We are recording sparse vocals and acoustic instruments, so want fidelity in capturing the sounds ‘around’ the instruments and voice. We are using 192k to be engineeringly conservative. In the absence of definitive research (and not the abundant supercilious and dismissive heresay) to the contrary, we are covering our bets by employing the best equipment we can afford at its maximum rates.
Sane discussion please!
Please, can we have sensible discussion about this. I want to know, rather than assume worst case as the basis for my decisions.
As that YouTube video cites, audio perception is highly subjective, so having preconceived ideas just leads to self-fulfilling conclusions. That does not mean we should just bypass engineering principles as we still have to create the same audio-physiological illusions regardless of the state of mind-feeling of listeners. That is, our illusions still have to work when they are not in self-deception mode! That really means they have to be physiologically subliminal, not just conciousness-subliminal.
But we still can’t fight nature, and win 
No matter what, or in who you believe, we still have no mechanics in our auditory system to hear anything above 20kHz (around birth, going down with age).
But our brains are easily fooled.
Please don’t make this more difficult than it is.
No matter what, or in who you believe, we still have no mechanics in our auditory system to hear anything above 20kHz (around birth, going down with age).
I hear Ya! (Ha!)
Your posts and references sound very scientific and logical and I am not disputing the ‘hard core facts’;
but at the same time there is some part of me that says in the future we will find that the earth is not flat. 
And that humans can sonically ‘perceive’ things that we should not physically be able to ‘percieve’.
I call it ‘quantum hearing’. And not just the higher frequencies but with very low ones as well.
And like —m. faraday with mucho experience and years of field work; I have no math to prove it.
But I truly believe it to exist.
{‘-’}
Then microphones and loudspeakers that can produce these upper frequencies will need to be invented first.
Human perception of not just sound (audio)
The entire electromagnetic spectrum.
I don’t know if recorded audio will ever be able to catch everything.
I remember being at a lecture when some harmonic singing was being demonstrated. Now, it requires being able to create particular resonances in the head, and has some strange-sounding noises required to get the process going.
I was on the PA desk and when it came to the demonstrations, the presenter told me to cut the audio so that their voices were not augmented in any way.
Well, what came out was what I can only describe as ‘the angels coming home to roost’ with the sound seemingly coming from everywhere in the room equally, totally filling it with ‘angels’ voices’.
It was absolutely amazing and I recommend that if you have the opportunity, listen to it LIVE, as even the presenters’ own recordings capture none of the magic of what happened live. However, if you can tell the direction of where the sound comes, then it is not being done properly, so do not judge it solely on such a sub-par experience.
Our Neumann U87ai and Jeanne Audio JA-251s easily reach well beyond 30k, as viewing recordings of voices and guitars in RX3 shows harmonics up there.
I don’t know where the Tannoys actually get up to, though I could record them with the mikes!
For me, I’m only able to get up to just below 10kHz, so I am reliant on RX3 and my wife’s superior hearing range to make sure there are no audio anomalies above that in my mixes!
However, I think trying to justify higher sample rates, or not, on the basis of hearing range is a red herring. The issue is really about the accuracy of the substantive in-band frequencies.
+1 ^^^
Hey… I think you might be on to something here! 
Al
The best plug ever made.
Yer ears!
And you can’t go out and BUY better ones.
i disagree
They need way too much maintenance to be called best plug’s ever.
Need to clean them at least once a week.
- they are not consistent in their signalling to the brain and processing is often related to a person’s health when having certain illnesses. (like flue)
So in fact they are very “unpro” and too unstable to rely on in my opinion.
kind regards,
R.
Most will blindly follow unfortunately.
IMO this is the problem. Everyone thinks they are not biased. We have all read about the different theories that can bias an audio test. The only scientific way is to have a neutral 3rd party do the tests in a controlled audio environment, eliminating all kinds of subtle variables, where there is little chance of any bias. And I would guess 95% have never done this. So yes, it goes back to trusting your ears…but anyone who proclaims they did some “unbiased” tests…I’m not buying it unless the tests were administered by experienced 3rd party.
As for the 3rd party, this is one time you don’t want to visit that $5 per song ME you found on Craigslist.
I have a audiophilic friend who was preparing himself for months to buy a new interlink from Siltech.
Months of reading reviews and also on forums kept him of the street. Once he had it, he was so pumped up that in a way he was very biased when A/B ing with the old interlink cable, he said everytime, “YES it sounds better”. I was sitting next to him and was talking him to the mouth “Yes I think I hear a difference”. The truth was that I really didn’t here the difference and If there was any I could easily say that it probably was a increase of 0,001 % in good or bad favour. Bottomline is that your mind is the biggest infleuncer.
The reason I was talking him to mouth was to not go into endless (and useless) discussion over what I was hearing and concluding (perception is subjective to me!)
Man I always found it hard to have discussion with audphiles, espacially when trying to convince him that the recording he was listening to was made with cheap ass cables and mics and mixed with digital effects… 
I repeat: Whatever you hear has to be within the human hearing range.
That’s why I don’t discuss this anymore. I try to tell that we can’t fight nature. If you really want to discuss the sample rate debate, please talk to any Auditory Doctor worth his grain of salt before buying into all misinformation and myths on the internet.
I happened to discuss this with some Auditory Doctors when my mother attended a research group for Tinnitus at the biggest hospital in my home town, country even.
She have had Tinnitus for many years and participated voluntarily in this research after consulting me. She knew I had more than average interest in the subject, after I finished my Sound Engineer education in the mid 90’s (started recording about 1979-80).
PS. For anyone who believe we can “feel” the frequencies above our hearings range, please follow the Fletcher-Munson (phon) curves and figure out how much energy you have to provide, to bring it up at an even level with some lower frequencies (we still can’t hear it, but may feel it).
The only way these high frequencies can be used in medical equipment is due the very hard handed high-pass filtering, using all the energy in specific frequency areas.
If we provided the same amount of power into the lower frequencies (we have no amps that can feed that amount of power across a full frequency spectrum within a musical context).
If we could have provided that power played back, we had burned up and/or exploded.
Think of why we can get hot in a ultrasound treatment. Then think of what we had “felt” when that power had hit us in a full frequency musical context (talk about Wall of Sound )
Don’t make this harder than it is. Don’t fight nature. It is only good old physics and math, and human auditory system limitations.
I finish: However clever you are, or good you are at cut’n’paste, and how many “mumbo jumbo” words you use:
Whatever you hear has to be within the human hearing range. Agree?
Unless You are an elephant, but it would be making it hard to control the mouse.
You are probably correct however somehow I’m still not convinced.
Everyday unorganized sounds yes.
But when it comes to music both recorded playback and in live performance.
I believe there is more to it than simple ‘hearing with yer ears’.
That missing ingredient is why humans get ‘touched’ by the sounds.
{‘-’}
Not to reveal any particular location but this hospital… Is it in Oregon? I was there in the early 1980’s when this Tinnitus research was initiated.
Tinnitus, by the way , while well off topic is a truly miserable affliction
The Nyquist (et al) theorem does NOT apply to real-world sounds as it ONLY applies to waveforms that are known over all time, forwards and back.
The Cheung–Marks theorem seems to indicate problems when the functions are not infinite and in the presence of noise.
I don’t pretend to understand all the maths, but I am not willing to accept statements incorrectly applying theorems as ‘fact’ to stymie valid discussion.
Also, the hearing range discussion is a irrelevant. Trying to limit scope to it, just because YOU think that is all that sample rate is relevant to, then using it to dismiss valid discussion of the whole topic is disingenuous and disrespectful.
It’s just that, as often, that the scientific worldview and the magical / wishful thinking of esoterics clash here.
So, here are a few facts, accept them or not, it doesn’t matter - they stay facts:
- Every waveform can be represented as the sum of sine waves (Fourier theorem - which is proven)
- To represent a sine wave of frequency f, which is the fundamental to every other waveform, a signal rate of 2f is necessary (which is the Nyquist theorem - which is also proven)
- Human hearing has an upper range of (at most) about 20 kHz, there is no way for humans to detect frequencies beyond this point, we are simply not equipped. Try to receive ultra short wave (FM radio, or “UKW” as we call it in german) with a pure short wave radio - you’ll fail. Epically. Same for your ears… they are an amazing piece of biology, but they simply can’t detect anything
above 20 kHz.
There is only one valid conclusion to all of this:
= A signal rate of 40 kHz + (some space for a super steep, high quality low pass filter) is enough.
And, since the people who build the audio devices KNOW all of that, there is a second conclusion, which I leave up to you to figure out.
A little hint: you can even apply some Mark Twain here.
The ONLY thing that is proven is that it is valid at MORE THAN (not exactly) twice the highest required upper frequency, BUT ONLY if the waveform is defined to plus and minus infinity (that is, doubly infinite). And the Cheung–Marks theorem (that is, proved) seems to indicate that being non doubly infinite and the presence of noise upset even this.
In other words, the Nyquist theorem sounds nice, and appears good to use for an argument for going for minimal specs, but it does NOT apply to ANY real-world sounds under ANY real-world conditions, so to continue to use it for ANY justification for one’s DAW sample rate choices is dubious.
I’m not saying I have answers here, but I refuse to accept INCORRECTLY applied theorems as EVIDENCE for ANYTHING.
Now, can we get down to real-world discussions so that we can actually understand what we are really dealing with in DAW work?
Hi Patanjali and it is good to see you here.
I divorced Pro Tools and bought Cubase last week and just today I registered on the Steinberg forum…and saw an old friend!!!.
Now this question you have asked about the best sampling rate to use is a beauty and I have actually done some tests on this.
I didnt do the standard listening test. I used sin waves and recorded them at different sampling frequencies.
What I noticed is that as the sample rate was increased up to 192Khz the sin wave was produced very accurately. However, at lower frequencies (and even at 96Khz) the sin wave was distorted on the recorded waveform. I understood this to mean that though the frequency of the signal was not altered, its harmonics were. Pitch the same but tonally changed.
The implication is that the higher sampling rate may sound better in the upper frequencies but can we hear those frequencies… probably not.
So to answer you question I say yes, you get a more accurate recording at higher sampling rates but no, you cant hear it.
Cheers
Lachlan
PS
I tried to send yo an email last week but it got sent back. Please email me with your new email address.