Golden ears

:: science, doomed

Or: Hi-Fi and the death of truth.

High fidelity audio is now dominated by people who think they have ‘golden ears’ and are able to hear differences between audio components which have been physically impossible to detect and almost certainly do not exist. Unsurprisingly these people refuse to consider any experiment which could reliably reveal whether they can, in fact, detect any difference. Often the same group of people will make absurd claims for systems which have very audible lack of fidelity — distortion — such as records and valve amplifiers, which shows how far removed they have become from reality1.

There are plenty of areas where subjective opinion is the only useful thing: my opinions about some of the guitars I own are extremely subjective, as are my opinions about various movies, bands, books and a huge number of other things2. But subjective opinions about things which are objectively measurable either agree with the measurements or they are wrong: if you try to live on the Moon without oxygen you will die within a few minutes, and believing you won’t will not keep you alive if you try: there are no alternative facts, there are only errors and lies.

Not all objective facts can yet be measured: for instance until 2015, while we believed gravitational waves to be a real phenomenon, we only had indirect evidence for them3. Perhaps the differences that the golden eared claim to detect between Hi-Fi components are real, but not yet measurable other than by their golden ears. If that was the case then there is still a good way of detecting whether they are real: carefully controlled, sufficiently blinded4 comparison experiments. If someone claims they can detect a difference between two things then you do a careful experiment which will reveal whether they can, while removing any possible bias due to the subject, the experimenter or anyone else involved. If it turns out that they can, then you know there is something to be measured and you can try and work out what it is. Even if you can’t measure it you know there is some objective truth there.

The golden eared reject such experiments.

The golden eared delusion is not itself significant: who, really, cares if a bunch of rich cranks believe they have magic ears? But the antiscientific thinking which underlies it is very significant. Once someone fails to understand that the differences they ‘hear’ are in their own minds, rejects experiments which could show this then they have decided that they can believe whatever they want to believe about Hi-Fi: there is no objective truth. And if you are free to believe whatever you like about Hi-Fi, why should I not believe whatever I like about vaccination, climate change or how many Jewish people, Roma and others the Nazis killed? Once objective truth is dead, it’s dead: you don’t get to say it’s only dead in specific domains. The golden eared delusion is one small path to the decay of any notion of objective truth itself which we are now seeing.

It’s interesting to understand how the mechanisms of truth decay work: if we understand them perhaps they can be reversed. So it is interesting to try to understand how how and why this particular pathology arose. Fortunately this is pretty easy.

Ancient history

Initially there was a period during which there was rapid and very audible progress in music reproduction. This started before 1900, and came to an end sometime in the late 1980s, with some tail-off after that. I have a Quad 2 / 22 / FM2 in (now) better-than-new condition and even with a very good source it’s really noticeably worse than a good modern amplifier. Even after amplifiers got good, there was lots of progress in turntable, tonearm and cartridge design right through the 1970s. However even then some early signs of the later pathology appeared, such as S-shaped tonearms which started because of confusion for which spurious justifications were later invented as they looked so pretty.

Throughout this period you could get better Hi-Fi by getting newer Hi-Fi. Manufacturers loved this of course, and a certain kind of person also liked it. My friends and I spent an altogether inordinate amount of time obsessing about Hi-Fi in the late 1970s, and everyone knew someone whose father (inevitably father: Hi-Fi was then and almost certainly still is a male hobby) had some really expensive and beautiful system on which we could listen to the latest Yes album. Some people even probably thought it attracted girls: in my experience, if it did, it didn’t work nearly as well as guitars, motorbikes or just being a decent human being did.

The wall

But then it all gradually hit the wall. Successively various components of systems got sufficiently good that, while they could still get better, they could no longer get perceptibly better. Loudspeakers were the only significant exception: see below for more about this.

During and after this event something related also happened: very portable audio reproduction systems arrived and became pervasive. Almost everyone who wanted to listen to music started using these systems rather than enormous Hi-Fi setups. Their quality was at first very poor, and any system based on in-ear headphones has limitations even now, but for people whose goal was to listen to music rather than to play with toys, they were more than adequate.

These two changes meant that large numbers of people who would formerly have bought Hi-Fi systems to listen to music bought portable systems instead., and people who still wanted to listen to music on a traditional Hi-Fi and had formerly upgraded their systems regularly to improve the quality largely stopped doing so as there became no reason in terms of quality of sound reproduction to replace components. This was very bad news for the Hi-Fi industry.

Well, the inevitable happened: many Hi-Fi companies went out of business. I have no idea how many Hi-Fi companies there now are compared to 1980 or how their financial value compares, but fewer, and less.

Some companies survived by simply servicing the relatively small remaining market as new generations of customers appeared: not everyone wants to use their parents’ or, more likely, grandparents’ Hi-Fi and in any case electronic components don’t last for ever.

Some companies realised that, for many people who are interested in Hi-Fi only part of their interest is in achieving the best possible sound reproduction: a very significant reason they like Hi-Fi is because it can be beautiful. It is undeniable that a well-engineered turntable or valve amplifier is a very lovely machine in much the same way that a vintage sports car or a 1950s jukebox is. There is absolutely nothing wrong with the desire to own beautiful things and very well-engineered things are often beautiful. A reason which is, I think, rather less good than the desire to own beautiful things is the desire to own status symbols: very expensive objects which will be recognised as such by other people. So some companies started to produce Hi-Fi which was explicitly designed to be appealing in this way.

Some companies — very often the same ones who were producing very beautiful Hi-Fi — made a business of being willing to maintain their own products for a very long time: when you bought something from them you knew it could be repaired almost indefinitely and you were therefore willing to pay a high price for it5.

A way under the wall

Other, less honest, companies realised they could exploit four secrets that had been lurking, usually slightly below the surface, in the minds of their customers since the beginning.

The first secret is that quite a lot of people suffer from what is called ‘gear acquisition syndrome’ — GAS for short. GAS was first described among musicians and it involves thinking that ‘if only I had a better (guitar, amp, effects pedal, …) I would be able to play much better and would become the hugely successful rock star I know I could be.’

In other words, GAS makes you think that what’s stopping you being a great guitarist is not lack of talent or unwillingness to practice, but lack of the right gear. GAS is a form of displacement activity where instead of dealing with the real problem — that you’re not very good, that you don’t practice — you spend endless ours obsessing over what gear to buy and in fact buying gear. And GAS doesn’t stop: once you have the expensive guitar and you still can’t play like Jimmy Page, well, it must be because the modern ones aren’t up to much — you need a 50s or 60s one. And the tape echo simulator you have isn’t good enough — nothing but a real tape echo (valve, not solid state mind you, the solid state ones where never up to much6) will do. And on it goes, endlessly, eating money and consuming the time you should spend practicing in looking at adverts & reading reviews.

GAS applies to Hi-Fi as well: rather than just listening to music people start obsessing that it would all sound much better if only they had better Hi-Fi. And it doesn’t matter whether it actually would, or even if it could sound any better: GAS is still driving you to buy more, ‘better’ Hi-Fi7, spending time on that which would be better spent listening to more music.

The second secret is that people like to think that they are special. Everyone likes to think that they are somehow gifted: one of the hard things that happens to almost everyone as they grow up is realising that, in fact, they are pretty much the same as everyone else. Some very few people really are gifted: Mozart was gifted, Einstein was gifted, Picasso was gifted, Jimi Hendrix was gifted. But most of us have more-or-less the same gifts as everyone else. And this is obvious really: if everyone is gifted, or most people are gifted, then, well, those gifts are just ordinary.

Most people learn this truth eventually, but no one enjoys learning it and some people don’t learn it at all. In some cases this failure to learn is very toxic: Donald Trump and Dominic Cummings are current examples. Even people who understand that they are not, in fact, special are susceptible to suggestions that they are. This is why rock stars, dot.com billionaires and film stars are such horrible people: they spend their lives surrounded by sycophants who are endlessly telling them how special and important they are. Almost everyone will eventually break under the pressure of such flattery: even if they didn’t start by believing they were, well, a bit special, they will end up doing so after enough people have said they are.

So if you tell someone, often enough, that they have the special gift of golden ears then, unless they have a very deep-rooted understanding of why they don’t, in fact, have golden ears — why golden ears can’t exist — some of them will start to believe they have. Because they’re special, and they have special ears: of course they do.

The third secret is that human sensory perception is both unreliable and subject to bias. There are very many examples of this: many of the most famous ones being optical illusions of various kinds. A good example is the chequer shadow illusion, in which two areas which are in fact identical shades of grey appear to be different shades. Another good recent example is the famous viral dress phenomenon from 2015, in which different people perceive the same dress as being coloured either black and blue or white and gold. There are many, many others: human visual perception is clearly simply not reliable.

But all these examples are optical illusions: perhaps hearing is special and is reliable and not subject to bias? This would be extremely surprising, but without evidence it can’t be ruled out as a possibility. Well, there is lots of evidence. A famous example is the McGurk effect: this is something that will be familiar to anyone who has watched dubbed films, or films in which the audio is not completely in sync. What happens is that, if you are watching someone speak then they actually produce one phoneme but the sound corresponds to a different phoneme, you can end up hearing a phoneme which is neither the one that they really said nor the one that the sound corresponded to but some third phoneme. If someone says /ga-ga/ but the sound of /ba-ba/ is played over it you can end up hearing /da-da/, for instance. Shepard tones, which seem to endlessly rise in pitch are another example8. There are many more examples of auditory illusions.

Perhaps the most well-known auditory illusion of all is stereo. When you listen to a correctly set-up stereo speaker system and a well-engineered recording, the impression that sounds come from instruments at well-define points between and often much further away than the loudspeakers9 is compelling for most people. But no sounds are coming from those points: your senses are fooling you.

None of this should be surprising: as our senses did not evolve to provide reliable, repeatable information free from bias: they evolved to make sure that we heard some terrible monster approaching from behind before it ate us. And if, occasionally, we hear monsters which are not there then that’s a lot better than being eaten. What would be surprising, in fact, would be if our sensory system was reliable.

The fourth and greatest secret is that people want to believe in magic. I’m sure there are some people who want to believe only in the things that science can explain, but almost everyone wants to think that, somewhere, there are elves and dragons, water-spirits and wizards. In 2003, the BBC conducted a survey to find the best-loved books in the UK. Of the top ten books, six involved magic of some kind; of the top five, four did. The best-loved book of all was The Lord of the Rings, which has also been found to be the best-loved book in Australia, Germany and the US, and has sold more than 150 million copies: about one copy for every 50 people now alive.

Almost no-one wants the world to be a place where there is no magic of some kind: even people who ‘don’t believe in magic’ want to believe in things like faster-than-light travel and time travel which are magic dressed-up as science10. I want to believe in magic: wouldn’t the world be a better place if there were elves, river gods and goddesses and magical objects? Of course it would.

And this desire for there to be magic runs pretty deep. Perhaps we can’t have elves and genii locorum (or, perhaps, we can, somewhere just out of sight), but can’t we still have magical objects? I know that there are no magical objects, but still I own a beautiful valve compressor using a ‘new old stock’ military valve. Is it better than a really good digital compressor? Certainly it is not, but I want to believe it is. And I know that my beautiful ES–175 is just machine made of wood, metal and bone and not even a particularly good example of one, and that it’s easily replaceable. But I would risk my life to rescue it in a fire, because some part of me believes that it is made of wood, metal, bone and magic.

The four secrets are:

  • the human desire for an endless succession of better objects (GAS);
  • the desire of humans to believe that they are special and have special abilities not granted to other people;
  • the unreliability of human sensory perception and the ability to bias that perception;
  • the human desire to believe in magic, and particularly magical objects.

Now it is easy to see how these can be exploited by unscrupulous Hi-Fi companies:

  • offer an endless succession of ‘better’ Hi-Fi components, thus satisfying customers’ GAS;
  • persuade customers that they are special and have golden ears able to hear the differences in sound that ordinary, lesser, humans can not hear;
  • rely on the unreliability of human auditory perception together with poor or no experimental controls to do this;
  • provide components which purport to be magic in all but name — special speaker cables, special capacitors, turntable plinths made of lignum vitae, Hi-Fi components made long ago which are purported to have magic properties and so on.

And, for some Hi-Fi companies this has worked very well indeed. The market is necessarily fairly small because magic objects don’t come cheap and, well, if the ordinary people had access to them the argument that the people they were selling to have golden ears would fail.

And, well, why does it matter? It’s obvious why it happens — companies need to stay in business, customers need to justify GAS, believe they are special and that magic exists, and lack of experimental discipline can be used to achieved this. And it’s obvious that the customers are gullible fools: but it’s their money, why should anyone else care?

The death of truth

They should care. They should care because people are not compartmentalised. People who believe that they have special powers in one area tend to believe they have special powers in other areas. That makes them deeply unpleasant people — it’s hard enough dealing with the arrogance of people who really are gifted: dealing with the arrogance of people who only think they are gifted is a horrible experience.

But that is only a tiny part of the problem: people who don’t accept properly-controlled experiments in one area will be less likely to accept them in other areas; people who think that they can ignore scientific method in one area will tend to ignore it in other areas.

But you don’t get to pick and choose: in the areas where science works, it works, and if you say it does not work in an area where it applies what you are saying is that, well, you get to choose when to believe what it tells you or not based on what you want to be true. The end result of this is that people will start to think that they can just get to choose what they think is true as suits them: if it is convenient to them for something to be true, well then it is true, if it’s not convenient, then it is false, even if it’s the same thing that was true yesterday.

Once you open Pandora’s box, then you open Pandora’s box: what comes out is whatever is in it, but also everything that is in it. You don’t get to say you only want some of the contents.

And we have opened Pandora’s box: we live in a world of alternative facts and made-up truths, a world where the very notion of truth is in the process of being destroyed. Many of us are now ruled by people who think that truth is whatever is convenient to them this week, because many of us think that truth is whatever is convenient to us this week. Truth is dead.

But it’s not: there is, in fact, a world outside your head and that world does not care what you think is true or whether you think you have special golden ears. That world cares only about what is true. A virus does not care what you think: it cares only about what is true. The physics, chemistry and biology of the planet’s climate does not care what you think: it cares only about what is true. The virus will kill you whether or not you think it can, and the consequences of what we are doing to the climate will kill billions of humans no matter how hard they pretend it won’t. There is no magic, there is only truth, and the only way to discover that truth is carefully controlled experiment.

The fabric of myths and lies on which modern Hi-Fi is built are a small part of the death of truth, but they are a part of it. If you think you have golden ears while conveniently choosing to reject any experiment which could tell if you really have, if you believe in the magic properties of certain components, or if you are merely involved in selling these myths and lies to people who believe then you are partly responsible for this catastrophe. And I will not forgive you.


  1. I am personally very fond of the sound of records and of valve equipment, and have constructed a valve hifi amplifier as well as owning another. But I am fond of them because I enjoy the lack of fidelity they introduce. I do not pretend that either thing provides particularly accurate reproduction. 

  2. Somewhere, some reductionist AI person is saying ‘but these can be reduced to objective facts about the state of the brain’. Well, yes, but ‘reducing’ something to the state of a system which is so complex no human can understand it in any detail (this is obvious: if a human brain has enough state to store a complete copy of another human brain then, by recursing, it is clear that it must have an infinitely large state space) is not reducing it to anything objective in any useful sense. 

  3. The first direct observation of gravitational waves was made on the 14th September, 2015, although not published until 11th February 2016. 

  4. These experiments are normally called ‘double blind’ since neither the subject not the experimenter knows enough to bias the result, consciously or not: they are both blinded. I prefer the term ‘sufficiently blinded’, which covers experiments where there may be more than two parties involved. What I mean by ‘sufficiently blinded’ is that no-one involved has information in advance which would allow them to bias the outcome. 

  5. For a long time Quad — formerly the Acoustical Manufacturing Company — made a business doing this: they may still. These two ideas — producing objects which are very beautiful, which serve, emphatically, as status symbols and being willing to maintain them indefinitely — has of famously been exploited in the photography field by Leica 

  6. Related to GAS is the idea of some lost golden age in which everything sounded better — somehow, the sound achieved with a specific model of tale echo, or a particular studio compressor made in tiny numbers in the late 1960s has never been equalled. It is not acceptable to consider that the sound achieved on records using the magic equipment might have been more due to the genius of the sound engineer than the equipment, because that would mean that your recordings sound bad because you are not very good, which can never be considered. 

  7. GAS is, in some ways, a malignant version of the ‘beautiful object’ motivation: the never-ending sequence of guitars, amplifiers and effects pedals that a musician with GAS buys are ever more exotic and beautiful, as is the never-ending sequence of Hi-Fi components that a person with GAS acquires. Of course, in both cases, the sufferer continues to fool themself that this is not what is motivating them. Again, Leica is a good parallel: if you are thinking of spending more than ten thousand pounds on a camera and lens and you think that it will make you a better photographer the you are a fool; it would be better to admit that you ate doing it because you want the beautiful object or the status symbol. The sufferer from GAS either does not recognise this or truly believes that it is not the case. 

  8. This is related to a device sometimes used in electric music called a ‘barberpole phaser’ which gives the impression of an endlessly upward or downward sweeping phaser. The same effect can also be achieved with flanging. 

  9. I find the stereo illusion works very much less well with headphones. But, well, it’s an illusion: perhaps other people hear the illusion much better than I do with headphones. 

  10. Almost certainly: FTL travel translates directly into causality violation, which is extremely bad news.