Hacker News Reader

Comment by Blackthorn | Hacker News Reader

Are you, per chance, a dog posting on the internet? Since 44.1khz sample rate is already past the range of the human ear, regardless of training.

View on HN

You need at least twice the frequency range for sample rate in order to represent the original signal. That's slightly misleading though, that's from the Nyquist-Shannon sampling theory and it's a mathematical fact but that is true for exact numerical samples, once you add in quantization that muddies the water a bit. Taken at the extreme, it's straightforward to see why a 1 bit quantization per sample at 44.1 kHz would not capture a perfect representation of some analog signal even if there's only a 1 kHz frequency component to the signal. If we instead decide to sample at 10 MHz but still one bit quantization, now that 1 kHz frequency component can be much more accurately represented even though we're still using the worst quantization possible. Don't think of quantization like a square wave or a step pattern, think of it as "the signal is closer to here than any other discrete value".

Now in terms of realistic audio encoding, 16 bit at 44.1 kHz is designed to be a faithful representation as far as human hearing is concerned. Can someone with a trained ear potentially tell the difference between that and 24 bit at 192 kHz? In a studio environment it's possible. Most audiophile claims are dubious and a blind A/B test catches them out on most of it but the Nyquist-Shannon sampling theorem does not directly apply to quantized samples, it's about exact samples and with quantization, sampling rate is intertwined somewhat with the quantization depth.

I don’t have great hearing, so I’m not sure I can really weigh in here (thanks punk concerts in my teens). I remember similar arguments around screens and 60Hz vs ‘the human eye’. I think a lot of people, myself included, can easily perceive the difference between 60Hz and something higher- given the right conditions. I would not be so quick to disregard claims of more sensitive hearing.

I would. It’s really simple.

The human threshold-of-hearing curve intersects the threshold-of-pain curve at about 20 kHz.

Above that frequency (or thereabouts) the sound has to be so loud that it will literally instantly damage your hearing before you can hear it.

This has been replicated across many studies for more than 100 years.

Flicker threshold is completely different. You can’t damage your vision by increasing the FPS, and it has always been commercially desirable to use a lower frequency because that is cheaper.

Would you agree that a trained human could identify artifacts produced by an imperect conversion process? If you lean "yes", then that's your answer: AD/DA is not a Rust function perfectly implementing the Nyquist theorem, it's a collection of physical components many of which introduce artifacts into the audio path. This thread is not about the theory of human hearing, the electronic components are literally imperfect.

They're no more imperfect than the pickups on an electric guitar, the assembly inside the microphone, the circuit in the compressor and everything else in the analog signal chain that exists long before AD happens.

Absolutely! All these examples have imperfect audio paths - that is the point.

But the central point is that there's no reason to pick on the digital elements in any particular way. Recorded music in 2026 is a pretty good recreation of the original acoustic pressure waves when it is intended to be, but (a) not perfect, even in the pure analog domain and (b) it is frequently not intended to be.

The central point is that AD conversion can and will introduce artifacts. DA process wil intrduce more artifacts. The "imperfect" is a huge range and AD/DA converters play a role in that. We are not talking about "golden cables" bs here, conversion does introduce measurable artifacts in the audio path. The more tracks you record the more artifacts you have. Can everyone hear them? Definitely no. Can they be heard - yes, I can hear the difference between an old Digidesign interface and Grace Design interface.

No, the central point is that the analog signal handling before AD introduces vastly more "artifacts" than the AD or DA does.

In addition, nobody cares about "measurable" artifacts (or rather, they should not). What matters are "audible" artifacts. We have measuring equipment that is vastly more sensitive than human ears (e.g. your recording equipment that can pick up signals far above 22kHz). What's measurable is not particularly interesting - what's audible is.

Artifacts do not sum linearly, because they do not originate from correlated sources (unless you're doing something rather unusual).

Glad you can hear the difference between two converters, but I trust you've tested it in a double blind setting?

Hm, no. The discussion was never about analog artifacts vs AD conversion artifacts. Both are present. And not sure why you use "artifacts", do you not believe the artifacts are real? How can the lowpass filter not introduce artifacts?

And absolutely - I blind tested coverters extensively. Mbox2, Black Lion Audio upgraded converters, UA, Prism.

Note: blind testing is not double blind testing. Scientists evolved double blind for a reason: blind testing doesn't remove bias.

Yes, the discussion was "never about analog vs AD". But my point is that I see little point wasting time on one set of artifacts (in the digital realm) that are tiny compared to those introduced in the analog realm. If there's a mouse and an elephant about to enter your home, you focus on the elephant, no?

The big difference, of course, is that "everyone" has convinced themselves that most/all of the analog artifacts, as big as they are, are somehow "tasteful" or "artistic", whereas the digital ones are just "math errors". I don't think is too helpful.

And look, if lots of people could get through double blind tests and still show they can hear aliasing or whatever the digital artifact du jour is, then I'd say "yes, absolutely, we need to be very aware of this and do everything we can to reduce or eliminate it". But as far as I can tell, this just isn't the case.

This is a more philosophical take… And I totally agree with you. I mix at 16/44.1 just for the record. I do not buy into the idea of gold plated connectors or 96 kHz mixing. My point was never about quality - I can hear the difference (the point!), doesn't mean for me personally > 44.1 is "better" or "worse".

To your main point: yes, all artifacts are just our learned, cultural, developed preferences. In the exact same way major/minor thirds were considered dissonant just a few hundred years ago - it's all a learned perception, not an absolute judgment.

I would go even further, doesn't matter whether people perceive aliasing as a major issue, it's no different from the U47 "warmth". You can't afford this, probably, as a software developer in a way, but at the most fundamental level any sound's - or artifact's - judgment is based on our our current diagram of "sounds nice" vs "sounds bad".

Can you give any examples of people identifying these artifacts in a/b tests?

Who has the best ears? What can they detect?

OK, so we are entering the stage of "can you provide a double-blind study link". I can look it up, I am not a researcher. Here is one: https://qmro.qmul.ac.uk/xmlui/bitstream/handle/123456789/134...

I know from my 20-ish year mixing experience that I can hear the difference when mixing. Is it good evidence? No. So we can agree to disagree then.

This is my first time seeing any research like this, thanks for the link.

I'm not disagreeing with you. I'm really curious about the limits of what people can hear, what can be taught and what is rare.

(I commented on this topic above/below in more detail.) Even with not-so-great hearing you would still be able to identify the difference (ie artifacts are pushed down, not up). Look up articles on the practical limitations of AD/DA converters and why the seemingly counter-intuitive claim that the difference between 44.1 kHz and above is noticeable, is actually a fully industry-accepted practical reality: aliasing, AD/DA lowpass filters, etc.

As I responded below, you are confusing math with physical reality. A true 44.1 kHz converter can't realistically capture frequencies ~18-20 kHz due to the limitations of filters used in the process. A perfect lowpass brick-wall filter just does not exist - they all introduce artifacts, which a trained ear can identify. You don't need to be a dog to hear the difference, just someone who does not assume that Nyquist theorem can be magically applied in the real world (and, ideally, someone who utilizes high quality converters with oversampling).

Is there evidence that a trained ear can reliably perceive these artifacts in a blind test of converters? I'd be interested in reading those links since converters typically oversample into the mHz range. At 11.29 mHz (256x 44.1 mHz), Nyquist will be at 5.64 mHz. Even the cheapest consumer converters are performing this type of oversampling.

If you are looking for studies, this one comes to mind: https://www.researchgate.net/publication/289039184_The_audib...

A quick search returned this PDF with a nice diagram of what aliasing looks like: https://download.tek.com/document/76W_30631_0_HR_Letter.pdf

To draw a design parallel: pixel-perfect design isn't something we are born with, noticing tiny details is a developed skill.

And yes, you are on point: oversampling is used extensively, but this just points at the exact issue: Nyquist theorem gave us a math algorithm, we still need to account for the electronic component imperfections. And then we are entering a different space of quality/precision/psychoacoustics/perception/etc. Meaning, not all converters, not all pre-amps, not all mics "sound" the same, even when they use same types of components on paper.

Oh, dear, that AES 2014 paper from Meridian (which was trying to push its controversial proprietary MQA audiophile system the same year) was widely criticized on audio forums when it came out, ranging from the rectangular dithering method to the use of a hard metal tweeter that could cause IM.

Do you have more convincing sources?

I don't. Do you? I am not a researcher. Saying that, do you have a double-blind study handy on MP3 256 vs 320 actual audible differences? If not, can you yourself hear the difference? If you can - it might be an illusion.

That extra 4.1 khz sample rate is for headroom for a low pass filter (and not necessarily a brick wall one). Leftovers or any such artifacts are below the noise floor, which is also an important part of the physical reality.

Would be happy to see an actual, real study to prove that humans can notice, but to my knowledge none exist that confirm they can. Not even any on teenagers or younger (the only group that can even hear close up 20khz).

Max representable frequency is half the sampling rate (nyquist-shannon theorem), which is still a bit above normal but IIRC the extra headroom has something to do with eliminating aliasing

Indeed. And what is the max frequency that a human can hear?

Depends on age of the listener, on average, 30 to 50 year olds hear a maximum frequency of 14 to 16 kHz.

Right. Which are quite below 1/2 of 44.1k!

Sure, but those are averages. I'm 30-ish, and my hearing doesn't cut out until somewhere in the 21kHz range. When I was younger, it was even higher. One of my roommates in college had one of those anti-rodent high-frequency noise generators, we almost came to blows over it.

The artifacts produced by pure 44.1 kHz convertion are aliased back down to lower frequencies. It's not about a theoretical human ear, it's about the actual physics of AD/DA conversion.

But the energies of the signal present above the Nyquist frequency (22050Hz in this case) are almost always incredibly weak, and double blind testing rarely shows any indication that humans can actually hear the aliasing.

Mixing process often involves hundreds of tracks, and if each introduces aliasing, this can become a problem. Some engineers do swear by "the final mix is 16/44.1 so why mix at a different resolution?" mantra - that's fine too.

This is false. Aliasing is not additive in any meaningful way.

Ok dude, you obviously never recorded anything. Twelve mics on a drum kit, 60 tracks of rhythm guitars, several bass guitar layers, vocals, backing vocals, electric organ, percussions, saxophone solo. Do you think recording them at 44.1 somehow creates a shared "cloud-based" aliasing artifact that I store in S3?

> Ok dude, you obviously never recorded anything.

https://ardour.org/ is my website.

Ha! OK I take that back.

Firstly, it's an amazing experience to randomly interact with people like you - I love and use your software. Hats off and thanks for what you offered to the industry!

But secondly, your statement makes even less sense to me: obviously artifacts do add up. Yes, not linearly, like any complex audio in general. But the more tracks with artifacts I have, the more artifacts I have overall. It's not like they cancel each other (outside of normal frequency cancellation).

Reply on news.ycombinator.com