Has anyone been able to define well or measure differences between vinyl and digital?

G’day MikeyDee

My wife and I had the opportunity to attend a small Mozart chamber orchestra in the castle in Salzburg Austria a few years ago. That’s where sitting close within 6-15ft the small audience had a wonderfully intimate experience of dynamic range. No Hi-Fi experience can reproduce that intimate experience in a double thick stone castle, huge rectangular room with pillars etc.

Perhaps @fair , can enlighten with at least 2 or 3 of these research papers he claims are hard to find? A new paradigm with 3 decades of research that legitimately calls into question all current signal processing and hearing knowledge should have many available sources to reference.

Still I wait for this. A meta analysis of purely digital sources, some too old to be relevant due to hardware limitations and others with experimental flaws, does not support your hypothesis let alone suggest there is any new paradigm. I do appreciate the repartee as it demonstrates the vinyl argument.

This is just like the tube discussion.  Even though there are significant, identifiable differences between typical tube amplifiers and SS amplifiers with good design practices, differences that are highly audible, every discussion devolves into a debate between those who point out those differences and those who believe in some unseen, unmeasurable property that "must" exist.

@cleeds

No, Fourier Analysis is not a theory. It's a theorem; it can be shown to be perfect with math. Unless you can show a fault in the actual math, all of your hand waving and word salad is for naught.

Fourier Analysis, as a theory, is a subset of Harmonic Analysis theory. It has dozens of theorems. If you are interested, pick one of these books - Harmonic Analysis on Amazon - and see for yourself.

A handful of the theorems is routinely used by DSP practitioners. For instance, Parseval's Theorem can be used for quick estimation of THD contributed by a device, via comparing pre- and post- waveforms.

Again, I'm an analog guy. But if we seek better digital, it can't be done without understanding how digital works. I'm all for hi-res and everything it takes to get the best out of it. But let's not pretend that somehow the underlying premise of digital audio is somehow broken. It isn't. Its problems are elsewhere.

One of the problems, as I see it, is simplification, excessive popularization, and at times even vulgarization of science, which became widespread in the Western world during the several past decades.

I guess I have to go that route as well, given the circumstances. So here it goes. A mathematical theorem is like a part of a legal contract: its words have precise meaning, often unexpectedly different from their everyday meaning; and it has small print.

As it relates to the Sampling Theorem, the phrase "contains frequencies" means something quite different from what everyday common sense would make one assume. And the theorem itself is just a paragraph in a long contract, with lots of small print.

Let's say you have a health insurance contract, and it covers your teenage son too. He rides electric bicycle. God forbid, he gets in an accident on the bicycle, hits his head, and requires expensive urgent care and rehabilitation.

Naturally, you assume that the insurance will cover it all, and you feel safe in belief that you'll be only out for deductibles and copays.

Suddenly, your insurance company sends you official letter saying that this event isn't going to be covered by them, because during the accident, according to a police report, your son was violating The Law.

Perhaps he was riding without a helmet. Perhaps he was riding on a walkway in a town with an ordnance prohibiting that. And there is a small print in the actual insurance contract stipulating its provisions null and void if injuries were sustained in the process of violating The Law.

As it relates to the Sampling Theorem, the signal has to fulfill very precise obligations, before the theorem can guarantee its accurate capture and reconstruction.

The meaning of the words describing these obligations is precisely defined elsewhere in the theory, in other definitions, lemmas, and theorems. Nothing wrong with the theory per se, at all.

Practical music doesn't fulfill these obligations, and thus the Sampling Theorem only works approximately. How well it works can be calculated too, using other parts of the theory, yet this is far more involved, and the answer is signal-dependent.

The vulgarization, in the case of CD format marketing, was in omission of the facts I described in the previous paragraph. Was it done on purpose or through honest mistake? I don't know. 

@thespeakerdude

Perhaps @fair , can enlighten with at least 2 or 3 of these research papers he claims are hard to find? A new paradigm with 3 decades of research that legitimately calls into question all current signal processing and hearing knowledge should have many available sources to reference.

Not what I meant. I meant that advances in signal processing and understanding of how mammalian hearing system works were significant over the past three decades. Yet not all of the advances are reflected in engineering practice yet. 

CD format, as an example, was developed prior to that, and its designers couldn't take advantage of these advancements. Sadly, this format and its derivatives remain leading by volume for lossless online streaming.

Certain common handbooks on DSP and auditory science haven't been updated yet either. Leave alone the mass of practitioners who still work in the paradigm expressed in these handbooks.

I'll give you a couple of examples. First, a gradual understanding, over about quarter of a century, of the role of so-called Octopus Cells in the functioning of the hearing system: https://www.google.com/search?q=octopus+cells+hearing.

Second, influence of new micro-surgery and robotic devices and techniques, which became widely available in the past two decades, on the hearing system research and medical practice: https://www.google.com/search?q=cochlea+microsurgery+hearing+research.

Still I wait for this. A meta analysis of purely digital sources, some too old to be relevant due to hardware limitations and others with experimental flaws, does not support your hypothesis let alone suggest there is any new paradigm.

That's why I prefer to believe in results of meta-analysis. Some of the smaller-scale experiments support the hypothesis, some others don't. All are imprecise in one way or another. Yet with a larger array of data, statistical inference starts working with sufficient, quantifiable precision.

From where I stand, this meta-analysis shows that the old paradigm, supporting the notion that 16/44.1 can encode perceptually transparently any music for any human listener, is experimentally disproven, and thus the paradigm itself is falsified by the evidence.

I do appreciate the repartee as it demonstrates the vinyl argument.

You are welcome. Winning a gratitude, however small, of a tough intellectual opponent, is one of the best rewards one can hope for in a discussion.

This is just like the tube discussion.  Even though there are significant, identifiable differences between typical tube amplifiers and SS amplifiers with good design practices, differences that are highly audible, every discussion devolves into a debate between those who point out those differences and those who believe in some unseen, unmeasurable property that "must" exist.

I lost interest in this particular discussion quite long ago, after I was able to replicate the "tube sound" inside a common DAW, using easily available software plugins. I occasionally listen to tube amps owned by others, including very expensive ones, but keep coming away underwhelmed.

In a way, the proponents of the tube and other highly distorting amps are right that there are some properties of such amps that evoke hard to measure phenomena inherent in the human hearing system, yet these phenomena can be activated by other means too.

For instance, the Missing Fundamental effect. It makes an amp sound more warmly-bassy to some listeners, which can be pleasant on some source material. Yet, this effect is routinely used in Bluetooth boomboxes, via quite simple DSP algorithms. The downside is of course masking of midrange, so it is beneficial only for some music genres.