You can email Ralph about his articles, ask questions, comment or argue: ralph-at-homeentertainment.com.au

The Missing Links In Speaker Design - Chapter One - By Ralph Waters

Scroll down for subsequent chapters.

The loudspeaker industry has been pretending everything's pretty much under control for decades. They have, however, been avoiding the truth about loudspeaker performance, using specifications which mean very little and omitting the information that would tell you a great deal about what the loudspeaker really sounds like.

Let's start by looking at what a loudspeaker actually does. It is a device that converts impulses of electrical energy into a physical movement. This movement creates sound pressure which we interpret as music. The better the speaker, the more convincing the illusion.

The whole aim of a loudspeaker is to create an illusion that is as perfect as possible. The closer to the original the better. So that also means that the speaker ought ideally to recreate the original experience with nothing added and nothing subtracted.

One of the most amusing and least relevant bits of information which loudspeaker companies publish is the frequency response graph. This graph is supposed to tell the reader how close to the original performance the loudspeaker reaches. It is supposed to indicate how accurately the loudspeaker reproduces each and every frequency. Are bits missing? Are certain frequencies accentuated at the expense of others? The frequency response is supposed to reveal such issues. The aim is for a flat response which would mean that the loudspeaker reproduces all frequencies just as they were recorded with nothing highlighted above the original and nothing muted or suppressed compared to the original.

Sounds important? Well yes it is, but the main trouble is that the measurement is typically a single measurement taken at one meter on line with the tweeter. This means that the listener would hear what the graph indicates at only one position…crouching in front of the speakers with ears directly in line and one metre from the tweeter. I suspect that few buyers would listen at this position. It turns out that the family dog has the best listening position.

The trouble is that a loudspeaker's measured performance changes at every position. You change the position of the listener and you get a different result. It is absolutely impossible to characterise a loudspeaker's performance at one position only. What we really need are multiple measurements taken at various angles and heights and at distances that represent typical distances of the listener. Once these measurements are combined and weighted, then one can have a very good idea of what a loudspeaker sounds like. These vertical and horizontal measurements have been standard practice in the professional loudspeaker industry, but the hi-fi industry has neatly sidestepped this more meaningful approach. The fact is that it's not difficult to create a loudspeaker that measures well ( or flat) at one position. The real art of loudspeaker design is to make a loudspeaker that will sound good at any position in the room, or at least as many positions in the room as possible.

The second critical requirement of a loudspeaker is that it does not add ' stuff'. Anything added to the original by the loudspeaker is bad. Anything added is noise pollution. This pollution is called distortion. Distortion just means that the reproduced sound is flawed to some extent or another. The loudspeaker can introduce sounds into the performance not present in the original. These introduced sounds are unwanted and marr the realism of the performance.

Distortion can be divided into two main kinds…even and odd. Even harmonic distortion sounds kind of pleasant…like singing in the shower. You know its not real but it sounds musical. Well sort of. Even harmonics can sound pleasant enough, but nevertheless they do not sound real. Odd harmonic distortion on the other hand sounds unmusical and jars the listener. It sounds like an acoustic equivalent to stepping on hot coals or putting razor blades in your porridge. (Ralph, I've told you a million times not to exaggerate! - ed.)

There are other kinds of distortion too…like transient distortion. This distortion only comes to light on certain notes and on peaks of sound. So the speaker sounds well until the golden tongued opera singer hits a certain note and the speaker complains with a burst of nasty distortion.

The point is that these distortions are easy to measure and easy to hear. It is the single most important measurement available. It doesn't matter how good a loudspeaker measures in frequency, if it suffers from distortion of one kind or another, the performance will be marred.

Distortion is easy to recognise and often confused with how loud the speaker is playing. For example, the same person can listen to a brass band in full swing with pleasure and walk into a room with a loudspeaker playing at half the same volume and complain that the speaker is playing too loud. What they are saying is that the speaker is distorted. Any natural instrument can be played loud and still sound enjoyable.

You know a really good speaker by the fact that you can play it really loud without noticing that you are playing it really loud. You know a really good speaker by the fact that you can sit or stand in most positions in the room and find the music enjoyable and real.

The second hallmark of a really good ludspeaker is that you can listen for long periods of time with little or no fatigue. Fatigue is really the result of distortion. The more distorted the loudspeaker, the sooner the onset of fatigue.

And here is the crunchline…..most loudspeakers would be automatically thrown away if distortion levels were ever published. Most fail and some fail dismally. Even in the professional sound reinforcement industry, distortion measurements are rarely provided. Distortion levels of 3-4% are not unheard of.

And so whilst we obsess about a few thousandths of a percent of amplifier distortion we are blithely unaware of massive distortion which is created by the loudspeaker itself. The loudspeaker is for this reason, the weakest link in the chain and always will be.

So one key to realism and satisfaction in hi fi performance is finding a low distortion loudspeaker….one that sounds good in almost any listening position.

It's about time that loudspeaker manufacturers delivered measurements that were meaningful. Distortion and correct frequency responses are not that hard to create. I suspect that most are terrified that speakers some forty or fifty years old outperform modern speakers in this respect and are anxious to conceal the truth of the matter. In some ways loudspeaker design has gone downhill. We are seeing less design innovation today then forty years ago…but more about this in another article.

Some of the finest loudspeakers I have ever measured, were cheap and humble. I recall years ago a certain paper coned tweeter made by Pioneer. It was incredibly low in distortion and sounded good everywhere. It was a really nice device. Another great tweeter was the early AR tweeter . It looked like a cone tweeter but had a dome in the middle. The best of both worlds? But loudspeaker fashion has seen the demise of the cone tweeter. So spending a lot of money on a loudspeaker is not in itself a guarantee that you are buying a quality sound. Sometimes some cheaper speakers that are well designed can sound superb. I recently was gobsmacked at the quality of sound coming from a loudspeaker that my friend Warren rescued from the dump. (photo to come).

In the next article I will look at other things which contribute to a satisfying loudspeaker performance, and explode a few more myths.

(Ralph Waters is an experienced speaker (active and passive) and subwoofer designer, and was the driving force behind Richter Acoustics and Subsonic before moving on to other pursuits. He still experiments with designs and produces fascinating prototypes, but is not currently involved in manufacturing.)

The Missing Links In Speaker Design - Chapter Two - By Ralph Waters

Subtitle:Why Warren's simple speakers sound so good!

Preface by Geoff Forgie - Years ago in the 1990s, Ralph used to visit our store in Canberra to show us his latest Richter designs, sometimes at prototype stage, with bits and pieces hanging off them. He'd assess our responses and possibly adjust the final design to take any criticisms into account. I make no claim to having special hearing, in fact I've been losing top-end frequencies for a long time, not unusual as we age.

But at the end of this article Ralph recalls a listening session where "Geoff" makes an insightful comment on the sound. When I first read this, I thought "I don't remember this", so it must have been his other friend called Geoff, who lived in the same city. But no, apparently it was me, so my reputation as a legendary listener/analyser is made! Watch out, all you aspiring speaker builders. I'll be listening.

This article zeros in on one of the all-time great myths about speakers. In the shops we get asked all the time "are these speakers three-way?", as if this is inherently better. They even ask this about speakers which are obviously bookshelf size, and cannot be anything but two-way. Three way speakers are usually floor standing and as a matter of course can do deeper base than their two-way relatives. But there is no guaranteed advantage in a three-way design, as it makes the crossover more complex. Ralph explains why crossovers can be problematic in this Part Two, which I call "You Are Being Inducted".

Read on, it's already a hell of a ride, and Ralph promises further installments!

Went round to Warrens place the other day only to be completely amazed by the sound of the speakers he had salvaged off the tip. Warren is a typical country town 'character', but a bloke that defies all categorisation. He ekes out a living making wood furniture from old timber. He loves his dogs, and spoils them by taking them down by the creek for an overnight camp. His ute is about a thousand years old and he somehow keeps it on the road by salvaging old parts from here and there. Warren is a bit of a philosopher and loves to listen to lectures and talks and a bit of music now and then. He is quite deaf in one ear and needs a hearing aid but never uses one.

So what kind of speaker impressed this man, and me, and how does the underlying design teach us something?

This article seeks to explain why very simple speakers can sound incredibly good. Incidentally I hope it becomes apparent why a half deaf person needs the best sound to be able to hear properly though I cannot think of a good reason why you would really want to know this anyway. But follow me, and I'll explain.

With two good ears a person can compensate 'internally' for any deficiencies in a loudspeaker. But a person deaf in one ear cannot compensate. A person blind in one eye cannot comprehend depth of field properly and the analogy holds true for Warren.

A person with good hearing can actively and constantly adjust what he really hears to align with the sound he ought to be hearing. This is a very clever brain function that means that the brain is the onboard active equaliser and is on the go the whole time adjusting and compensating for a whole range of things. The brain knows how the music should sound and then adjusts the sound coming to the brain from the ears, so that it lines us with that internalised sound.

This is particularly true when bits are missing. The brain can insert missing bits and make a whole sound picture. But Warren cannot do this very well as he only has information from one ear that his brain can use and process. You see the brain uses information from both ears and then processes amongst other things the difference between the two signals to perform its compensating gymnastics.

So why do Warrens speakers sound so clear and so distinct?

Firstly, you will see that the speakers only have two drivers in them. A six inch woofer and a 50mm tweeter of the old fashioned cone variety. The six inch speaker naturally rolls off at higher frequencies and does not use any kind of crossover at all to help. The tweeter uses a simple capacitor to stop the low frequencies ( bass) from destroying the tweeter from excessive cone excursions. So we have a single component crossover. By good fortune it just so happens that the tweeter fits in perfectly with the roll-off of the woofer.

A "crossover" is that group of components inside the loudspeaker designed to make the woofers and tweeters all blend in smoothly and sound perfectly matched, and flatten out the response curve. Items used are capacitors, resistors and inductors. These components are in themselves the dirty grubs of the loudspeaker industry. They hide away like parasites eating away the goodness of the sound while most people are either oblivious to their existence, or if they do know a bit, think they are a good and healthy part of the speaker, and the more complex the better.

One of the paradoxes of the pharmaceutical industry is that anti depressive drugs have as one of their side effects - you guessed it, depression. And so too one of the depressing side effects of all the crossover components is that they create a certain amount of havoc with the sound. The very components that were put in place to make the speaker sound "balanced" are the same components that compromise the sound.

The first component we shall look at in this discussion is the simple inductor. An inductor is a long piece of wire that is often wrapped around a ferrite core. On better quality speakers the wire is wrapped around fresh air. It can look a bit like a donut in some cases.

This wire is often thin - perhaps only 0.8mm thick, and to do its job it needs to be long. So an inductor that is performing the job of "rolling off" a woofer can often contain thirty metres of wire in a two way system and much more again in a three way system. Imagine a thin wire stretched out thirty metres to which your speakers are connected. This is the picture of what happens inside your loudspeaker.

The wire has an unpleasant characteristic of resistance. The inductor may have ten times the resistance of the beautiful and expensive speaker wire that goes the last three metres between the amplifier and the input terminals to the speakers.

But the worst thing about an inductor is the inductance it creates. You see an inductor delays the sound. The bigger the inductor the greater the delay it causes for the electrical signal from the amplifier to get to the woofer. That delay means that instead of music sounding as if it is responding instantly, it always sounds a fraction delayed. That's because it is a fraction delayed. Or to put it around the other way, a speaker with no inductor or a very small inductor sounds alive and energetic. The speaker with the big inductor sounds lazy and uninteresting.

So Warrens' speakers sound very real. Very "there" . Since they are very present, the music is much more involving. The speaker seems fast and vital. The instruments sound more detailed and more realistic, more convincing.

But the true tragedy of the inductor is that not only does it delay the sound, the bass frequencies lag behind, whereas the high frequencies do not. So half the music sounds exactly right whereas the other half sounds as if it is not quite right. Somehow the speaker does not quite sound right.

I doubt whether this phenomenon ever reaches the threshold of consciousness with most people. They respond to a loudspeaker but may not know why. They have an intuition about the speaker but are unaware of the time factors that are shaping their personal responses to a particular loudspeaker.

Hi-fi buffs talk about a speaker's tempo, how "fast" it is. Really, what they are describing is how much time delay has been imposed on that speaker's responses.

Now here comes another rub…. The "better" crossovers use more components. You see the more components the more control you have over the finished result. And loudspeaker designers love to have a nice flat frequency response curve to prove that that they are doing their job properly. But with more components you get more and more delay and more and more interaction between the actual components themselves. You see, inductors store energy and so speakers with really complex crossovers often need amplifiers big enough to weld steel to drive all the components within the crossover. And the inductor is the worst offender. The inductor stores a lot of energy and then fights with the amplifier. The amplifier needs to be tough and unyielding in order to drive some loudspeakers crossovers. The inductor is like the rebellious teenager always wanting to do the opposite thing asked and always later than requested.

So a simple crossover requires a simpler amplifier. A simple crossover does not need the bully boy tactics of the monster amplifier to make it perform properly.

But a bully boy amplifier can never compensate for the time delay that the inductor creates. And so that is why even the best amplifier in the world cannot regain what was lost - and that is the perfect timing between all the frequencies.

So Warren's speakers - despite their limitations do some things really well. In fact they are superb in some areas and very satisfying to listen to. Warren did not know it but he stumbled upon a great truth when he said to me "Ralph these speakers sound really good". Yes they have no time delay and the little amplifier that drives them does so with perfect ease.

So the second lesson in speakers is that crossover networks pull apart sound and meddle with its ability to reproduce all frequencies in their correct time. These apparently small changes are undermining a potentially more satisfying listening experience. The perception of timing is almost at the intuitive level. Yet some can discern what is going on fairly readily. I recall with great clarity the day that I put an inductor on the tweeter circuit in an attempt to smooth the response. It worked. After a few minutes listening Geoff remarked "It sounds smoother but the high frequencies sound delayed in some way". Bingo, Geoff you perceived it correctly.

So Warren's speakers teach us that more is sometimes less. A loudspeaker's timing is as important I think as frequency response itself.

Off to milk the cow. See ya later.

The Missing Links In Speaker Design - Chapter Three - By Ralph Waters

Continuing: The Crossover Gang's Life And Crimes

In the previous chapter we were introduced to Ralph's neighbour Warren, whose speakers have become the stalking horse for Chief Inspector Waters, as he builds the case for the prosecution of poorly designed speakers, and the components that have become like a faustian bargain, offering powerful, magical fixes but stealing the soul. In the intro to this part, he adds to the picture.

"Well Warren loves a pretty gutsy, outdoors bush life. An outing for him is to take his dogs out for an overnighter down by the creek and sleep rough. He loves his dogs and they love him. He loves the bush and everything natural and beautiful.

Warren makes exquisite furniture from bits of dead timber he finds in the bush. A lot of blood sweat and tears plus a bit of elbow grease and hey presto another beautiful piece of furniture. More like a sculpture really. He made me a coffee table made from forty three separate bits of timber. Can you believe it?

Warren is not an audiophile but he does appreciate fine aesthetics. He does recognise clarity, low distortion and a big open sound. He can correctly identify an accurate frequency reproduction despite, or perhaps because of, his deafness in one ear. It makes him a more critical listener, as the brain has less chance to "gloss over" the bad bits. So the fact that his speakers have only a single component crossover network in the enclosure gives us a clue about the source of clarity and other good things."

In the previous articles, I started to show how the more components in a crossover, the more reactive and difficult the speaker was to drive. And the more time delays were introduced into the sound; tiny delays that really interfere in our listening enjoyment. We saw how Warren's speakers only had one crossover component - so they so not have the whole bunch of components that confuse and fragment the sound. We also noted that the bigger the crossover network becomes, the bigger and tougher the amplifier must be to fight the reactive load of the crossover.

Inductors not only delay sound, store energy, fight with the amplifier and create havoc generally, they also can and do change value when overloaded. Yes folks, they actually change the crossover points of the loudspeakers on transients. Hmm. When this happens it sounds like something went horribly wrong for a tiny moment or so. So, if I had to choose something to hate, I would choose inductors, slightly ahead of margarine or fluoridated water.

Capacitors also play a significant role. They have the incredible job of transmitting signals that keep on changing - at up to 25,000 times a second. After a while the capacitors can become tired and weary from all this frenetic changing of direction, and end up listless and lifeless, a condition which stops many older speakers from reproducing the more subtle aspects of ambience , clarity, fine detail and the immediacy of transients. This sort of lacklustre performance is easily prevented by using the highest quality capacitor - something that few loudspeaker manufacturers actually do. Resistors in the crossover are also problematic, but generally nowhere as critical to quality as inductors and capacitors.

But here's the rub: when using high quality capacitors and inductors, the crossover can still create havoc with the sound even if correctly designed. You see, the components interact with one another and the crossover itself has its own transient response. The crossover can induce peaks and dips not found in the response curves of the woofers, tweeters and midrange components themselves. Although this can, if carefully applied, help develop a "flat" or accurate response, the outcome can be the compromise of transient response.

This means that the components are unable to respond to the music without ringing and resonating on sharp peaks. Since music is comprised of a lot of sharp peaks - this is a serious issue. So a loudspeaker designer is constantly having to juggle the frequency response of a loudspeaker - that is, the ability to remain true to the original piece of music - with transient response: the ability to reproduce peaks and variations in sound pressure without ringing and introducing distortion.

So there we have it, the case for the prosecution. Crossovers delay the sound, interfere with transient response, induce distortion components of their own and overall are thoroughly nasty and horrible.

However, do not despair. It is easy to take a well designed loudspeaker and, instead of using passive components inside the loudspeaker such as capacitors, resistors and inductors, to replace these with an "electronic crossover" which suffers none of the above problems. The transformation is instant and dramatic. The speaker is live, fun, energetic, dynamic, and has much less distortion. It can be listened to for a long time before listener fatigue becomes an issue. More about those later.

To recap, Warren's speakers sound so appealing due to the absolute simplicity of the crossover network. One simple little capacitor…. that's it!

But there are other reasons too why Warren's speakers sound so good and the main reason is that it provides a BIG open sound. You may have thought that a big open sound is provided by buying a big speaker. Actually it doesn't work like this at all.

So in the next part of Why Warrens speakers sound so good, I shall be specifically looking at what makes a speaker sound Big and Open.

Enough for one day, time to service the tractor.

Chapter Four - The Problems With Multi-way Speakers

Or,Is More "More"? - By Ralph Waters

Well it is, more or less. A picture of waves on water is very instructive when dealing with loudspeaker design and in comprehending how sound moves through the air.Go down to the creek and chuck a rock into the water. It puts a circular pattern that radiates evenly from where the rock hit the surface of the water. And believe it or not that's how sound radiates through air. Well sort of. A simple example will illustrate this point. Take the sound from a triangle. It sounds pretty much the same from any angle: front, back, side. Imagine a circle of sound radiating from a sound source of any kind - musical or not. The following diagram shows how the sound travels (looking directly down from above the sound source).

In a loudspeaker we usually have multiple points all of which are radiating sound. These components are called tweeters, midrange and woofers (or more generally speaking - transducers or "drivers"). In a loudspeaker the frequencies which emanate from each driver are different. The woofer handles the low notes, the midrange the middle notes, and the tweeter the highs. On some speakers there's a super tweeter, which extends the response of the speaker well beyond the normal hearing range into what is called the ultrasonic range. Many speakers are now used in conjunction with a subwoofer. So a loudspeaker system can have as many as five sources of sound , each radiating outward. Plus the bass coming from the port if there is one, so that makes six.

This is where the problem comes in. Unlike the rock in the water, we now have multiple rocks, and as the circular patterns radiate outward, each pattern interacts and disrupts the pattern of the other transducers. If there is only one rock, we have a perfect circle. The more rocks, the messier the result.

Not only that, but each transducer is separated by a physical distance. Just as multiple rocks splash into different places different distances apart, so too the transducers each radiate at different places. A two driver speaker has a smaller problem to deal with than a multi driver speaker. There are more sources of sound - all of which interact with one another. And the multi speaker system has bigger distances between the individual drivers.The sound from the tweeter reaches the listener first, followed by the midrange and then finally by the bass notes.

We have a messier acoustic pool the more rocks we throw in. Somehow the listener has to catch each band of frequencies and reassemble them to make a piece of music. On the downside, multiple distances changes the time that each signal arrives at the listeners ears, each driver has a crossover which also applies various time delays. The human brain must somehow do massive calculations, adjust the anomalies and come up with a final result that makes sense. Thank God that the world's fastest computer sits between our ears!These issues affect all multi way speaker systems. They all suffer from this inescapable problem. No exceptions. Nothing can get around this problem using conventional technology.

With all this going on, one wonders why anyone would build a multi way speaker system. Well, believe it or not, there are good reasons why designers opt to travel this path, but none of them overcome the basic physical issues that we have just described.

The fact is that loudspeaker design is always trading one thing off against another. Improvements in one area are always offset by costs and tradeoffs in another area. The loudspeaker designer is always weighing the costs versus the benefits and resolving various issues in a complex and multi dimensional equation .

Is there any theoretical solution to the problems I have described? Well yes there are. A point source speaker will not suffer from these issues. What is a point source speaker? Well, they take three forms. The first is a simple twin cone speaker. All the sound comes from the one device, with no separation of time or distance at all. The second point source is a coaxial speaker where the tweeter is mounted in the centre of the bass driver, sitting just in front of it. There may be some difference in time - depending on what happens in the crossover network, but there will be no separation in distance…they both share the same axis.

The third type is where the tweeter is actually contained within the magnet structure of the woofer. The classic old fifteen inch concentric Tannoy range are a good example of this approach. In these three instances all the high mid and low frequencies share the exact same axis so there are no anomalies in sound caused by differences in distance between the tweeter and the woofer axis. The first pair of speakers I ever owned were a point-source kind made by Coral and they remain one of my very favourites. There is something eerie about speakers with a point source characteristic.

Those speakers that have struck me as wonderful have always been, if not point source, approaching this ideal by the very small distances between the tweeter and the woofer. So for me a well designed 4-5 inch two-way speakers always have an allure.To improve on the equation you want point-source speakers where the woofer has absolutely no crossover network at all. These designs mean there is absolutely no delay in sound caused by the crossover at all. This is maddeningly hard to design and involves finding a woofer with exactly the right frequency response. I have designed several speakers like this but it has taken a mammoth effort to design the actual woofer to do what is wanted without an electrical crossover network.

How is it done? Well it's very involved, and rquires the use of physical rather than electrical components to shape the response of the woofer into the desired area.

OK, what are the benefits of a multi-way speaker system? A multi speaker system is one in which the sound is divided up into more than two drivers. A three or four way system is an example of a multi-way speaker. And yes there are definite benefits.

Although I have been critical of multi-way designs, there are some clear advantages which cannot be ignored. Firstly the fact is that a woofer with its massive and heavy cone is no good at reproducing high frequencies. Conversely the tweeter with its tiny and super light cone is absolutely useless at bass reproduction. The more the sound is divided up, the more the individual drivers can be perfectly matched to the frequencies needing to be reproduced. There is nothing more natural and startling in its speed and accuracy than a light cone and small coned midrange driver.

Speakers using a midrange sound faster and more accurate. They also have lower distortion. I recall vividly years ago when I used a super light cone four inch woofer made by Fostex for a midrange unit in a three way system. I made it for a friend and the result was scarily real. The fact is that the driver was perfectly suited to midrange frequencies. It was super, super fast, very low in distortion and open, "airy". The music was realistic. Even in a multiway system the exact selection of drivers is crucially important. But more on that topic later.

I have a penchant for speakers which use a super tweeter. Somehow speakers which can travel right through to incredibly high frequencies have a supernatural sense to them. They make me shiver with delight. I have a sense that something is somehow better and different but I cannot pinpoint why or how. Sorry if this sounds mystical, but it is tied up with the reproduction of harmonic structure, which is well beyond the limits of this discussion. But there is a scientific and measureable reason why a supertweeter improves the sound of the entire speaker.

Conversely, I have always enjoyed a subwoofer that can accurately reproduce sounds to 25Hz. I am not talking about the ridiculously pumped up subwoofers that abound these days, but a carefully design sub that really reproduces music rather than noise.

The Richter Evokator was one such animal. I recall vividly recall a demonstration that I conducted at the Audio Trends store years ago. I used a small two way speaker made by Brad Serhan from Orpheus Loudspeakers called the Minotaur. It was a five inch two-way speaker that sounded great everywhere but lacked bass extension, authority and that feeling of power. Add a good subwoofer and the result was staggering the music sounded more dynamic, more exciting and more real. But not just in the bass. Turn off the sub and the speaker sounded flat and disappointing by comparison. The audience were suitably impressed, as they ought to, as the demonstration was of the cheese and chalk variety. The subwoofer somehow made the entire sound spectrum more exciting dynamic and believable.

I guess that the only conclusion I can reach is that the more a speaker reproduces the full audio spectrum, the better it should sound. And more drivers should in theory solve the problem of how to reproduce sound with lower distortion and a more open dispersed sound.But then you have to subtract from this, the confusion to the brain created by more drivers and the inevitable and inescapable time delays of more drivers and the mental confusion the listener has in recompiling the separate bands of sound into one integrated piece of music.

Speaking personally, I have constantly contradictory and mixed feelings about many multi way speakers. I often think that they sound great in areas such as lowered distortion, wider frequency range and better dispersion but somehow the overall music often sounds confused to one extent or another.

In the next chapter I will look at one of the most neglected topics in hi fi speakers - that of dispersion. This is a real eye opener as very few loudspeaker manufacturers talk about it or measure it, yet like distortion it is one of the most important aspects of loudspeaker design. In amongst all of this you may take issue with some of the things I have said or have some questions. Please feel free to contribute, or just ask away. Selected emails will be published by our editor Geoff Forgie. Send to Ralph-at-homeentertainment.com.au

Anyway the heat of the day has gone I might go down to the river. This time of the year it flows swiftly and at a temperature of 11 degrees it air conditions the river banks beautifully.

Editor's Note: Phew! It's not the heat that had me worried, but as an enthusiast for large multi-way speakers with plenty of bass I thought for a while that I had it all wrong. But in the end, Ralph admits that they can do good things.

Chapter 5 - Dispersion (Pt.1) - Is the Sweet Spot Really That Sweet?

The role of dispersion in creating a satisfactory hifi illusion.

Some audiophiles take a misguided enjoyment in setting up their seat for that 'sweet spot" where the speakers sound just right. The very fact that the speaker has such a specific sweet spot is a flaw not a virtue of design.

Warren's basic two-way AR speakers sound big. Really big. Despite their modest size they are clear at almost any point in the room. The sound seems to "fill the room", as some reviewers are fond of saying.

A good speaker should sound clear and accurate anywhere in the room. That means as the listener travels 'off axis' or away from the centre of the speaker's forward direction, it should sound just as good as right in front of the speaker. Further, sitting or standing should make no difference at all to the way the speaker sounds. The sweet spot phenomenon is a giveaway that the speaker is "beaming" rather than "dispersing".

What we are talking about here is the loudspeaker's ability to spread all frequencies throughout the room in an even manner. Some speakers do this very well and others do a lousy job. The fact is that most speakers sound totally different, even unclear, the further one travels away from that "sweet spot". There are a number of why this happens and it turns out that the loudspeaker designer is wrestling with some physical problems over which he has little or no control.

The first of these is the diameter of the woofer. It is an inescapable and non-negotiable feature that larger woofers cannot disperse higher frequencies. Take for example a twelve inch or 300mm woofer: it will fail to disperse frequencies over about 1200Hz. After this the frequencies rapidly become beamed into the centre. A ten inch or 250mm woofer reaches a little higher to around 1500 Hz and an eight manages 1800 Hz. A six inch will be OK to about 2400Hz and a four inch will disperse well to around 4000Hz. What happens with these limitations?

Consider a typical two way speaker comprising an eight inch woofer and 25mm dome tweeter. The tweeter crossover point is typically 3,000 Hz and the woofer can only manage to disperse the sound effectively to 1800Hz. That means that in the most critical band of sound, the speaker creates a narrow beam of sound. Instead of a big full and open sound, it becomes narrow and congested - and there is no remedy. The speaker can be played loud, but it will not, subjectively, seem to fill the room in the most vital vocal area. Loudness and completeness of sound are two quite different things.

The very first speaker that really impressed me was an AR6. It was 1970 and I went shopping for my first commercially built loudspeakers. I did what most people did and read all the reviews and listened to hundreds of speakers. I was hit by the remarkable openness eveness and natural quality of the AR's. They were cheap and small in size and eventually my eyes and wallet led me to a much more complicated loudspeaker which had rave reviews and more drivers and a much more impressively sized box. It was the first of many poor hifi buying decisions. I should have bought with my ears not my eyes.

Harking back to what I said above about beaming, the AR6 was an eight inch two way and you might think it might suffer from poor dispersion. But no - AR had a different concept which solved the problem of dispersion. The woofer was crossed over at 1800 hz - exactly where the woofer would normally start to beam. A two inch cone tweeter took over from that point. In the centre of the tweeter was a small dome that enabled the dispersion of the AR6's to remain open and wide throughout the whole audio spectrum.

Sadly the cone tweeter has been abandoned for small dome tweeters which by and large do a better job with higher frequencies, but cannot be crossed over at such low frequencies - they are not robust enough. When you cross a dome tweeter over at low frequencies the dome is too close to its resonant frequency and a huge increase in distortion occurs. In addition the dome tweeter struggles to handle the vast increase in power, and poor reliability ensues.

If you look at an on axis frequency response of a two way speaker with a dome tweeter, you see a nice flat response. But go off axis or off centre and the response deteriorates rapidly. There is an exception to this. If you use a five inch or four inch woofer, you can get the best of both worlds and enjoy open highs lows and mids. Well the bass may not be big deep or impressive but everything is open and clear. Ironically one of my favourite speakers was a four inch two way speaker called the FairyTales. This was an unmitigated commercial flop, which is sad because I think it had one of the most natural and open sounds of any speaker I have ever heard.

In the next chapter we'll look at propagation patterns, vertical and horizontal. Some instruments have a spherical propagation pattern which is in mnay ways ideal. Arguably, the ideal speaker is one that can sound the same up down, side to side anwhere…exactly the same. This means that the listener can be lying on the floor, sitting on top of a ladder or hiding in the corner, or halfway up the stairs, and still enjoy the perfect listening experience.

But alas, most speakers have huge problems in this area. The graph here shows that with most speakers , lower frequencies spread out naturally and high ones tend to beam. In the graph we can see how the same driver spreads the 1Khz signal wide , whereas the 5Khz signal is contained within a much smaller area. The trick of the loudspeaker designer is to keep this envelope as wide and as high over as much of the frequency range as possible.

A theoretical solution to this is to divide the frequency range up into smaller bits and have these bits reproduced by smaller and smaller drivers. For example a tweeter with a 12mm dome will not start to beam until well beyond the range of human hearing. Smaller cones or domes naturally spread the sound. Larger ones don't.

So the multi-way speaker often does have a terrific dispersion, although as I have pointed out in the previous chapter, often at the cost of timing and accurate imaging due to the fact that the image is spread over a range of sources rather than one single point. The vastly more complicated crossovers introduce a swathe of difficulties and issues. Having said that, correctly executed multi-way speakers can sound really superb. Historically, I guess my favourite speakers have been 12" three way designs, but I hear many poorly designed multi-way speakers, and it's a bit off-putting.

Remember, if you are going to a three or four way design, prepare yourself to pay a good amount for an amplifier capable of "capably" driving ( yes the repetition is deliberate) the speakers. I think the message is that more drivers does not mean better. It can sound good in many areas, but do not make any prior assumptions that it will sound better. How to fix them? A multi-way speaker with an electronic rather than passive crossover should always sound better. I'll be talking more about these later.

A convincing speaker always has an out of the box sound created by high dispersion. A real sound is always a sound that is not limited in size. One final mention of Warrens AR speakers: they sound so big because the cone tweeter takes over from the woofer before the woofer starts to beam or focus the frequencies.

So there you go, one set of speakers salvaged from the Bingara dump can teach us heaps about loudspeaker design. With 39 degrees in the shade outside, I think I'll just sit tight and enjoy the air conditioner. Way to hot for me.

Editor's Note: Over at our HiFi & Home Theatre Page, I recently wrote about old speakers I like, and an amazing number of them do turn out to have some wide dispersion tricks up their sleeve.

The Truth About Loudspeakers - Chapter 6 - Dispersion (Pt.2) - By Ralph Waters

If you could visualise a picture of sound in colour and then paint a picture of what each loudspeaker sounded like, it would be a very informative exercise. Believe it or not, that is exactly the data all professional loudspeakers used for "sound reinforcement" (such as those for theatrical, concert and even sportsground venues) come with. What we really need to know about a loudspeaker is how wide and how high the sound is spread once it leaves the loudspeaker. Below is a picture of a loudspeaker. It shows the loudness of a loudspeaker viewed from the side.

The next picture shows the loudspeaker viewed from above. The different colours represent loudness. The hot colours show the loudest sounds, the cold colours the softer sounds (relatively speaking). The professional loudspeaker manufacturers regard this information as essential and basic, but somehow in the hifi world, this information is never supplied. I believe that this information is just as important in the home hifi realm as it is in the sound reinforcement industry.

In a large venue, the speaker systems are deliberately designed so that they achieve good dispersion of sound, so that a listener in any position within an auditorium will hear pretty much the same sound as another position. Bad sound means complaints and that means lower profit, so it is imperative that every seat sounds as good as possible.

To this end, professional loudspeakers are usually designed to have a limited vertical dispersion and a wide horizontal dispersion.

What we need to be reminded of is that every loudspeaker has a completely different frequency response at different angles. For example in the series of graphs below, and reading from top left to bottom right, we see the horizontal dispersion of a loudspeaker and we can see that the higher the frequency the narrower the beam of sound coming from the loudspeaker in a forward direction from the loudspeaker.

Whereas at the 20-40 Hz level , the speaker is truly omnidirectional…the bass is equally loud in front, at the sides and even behind the loudspeaker.

That is why the traditional loudspeaker response curve normally supplied with hi fi speakers is basically meaningless…it is measured in one position only ….at one watt on axis at one meter. Whereas most people do not listen on axis. Most of the time people are scattered around a room and yet still want to hear the music clearly.

I've said before that the one watt on axis response at one meter is great for the family dog but useless to everyone else. That curve really tells you nothing.

How is the loudspeaker dispersion pattern controlled? There are two main methods that a loudspeaker designer employs. The first is to apply a horn loading to a tweeter or midrange unit. This is standard practice in the pro market but rarely seen in the hifi market. Most hifi speakers do not have horns, so the hifi designer doesn't normally have this tool at his disposal unless he's determined to make it so. I have, however, seen a growing number of shallow horn loaded tweeters in the last couple of years.

The next tool is the actual arrangement of drivers on the baffle of the loudspeaker. Believe it or not the pattern of drivers actually has its own signature pattern of dispersion. This is seldom talked about in the hifi world and is a sorely neglected topic.

One very strongly promoted configuration is so called D'Appolito. The D'appolito array is a speaker design pioneered by Dr. Joseph D'appolito in the early 1980s. A D'appolito array consists of a single tweeter with two midrange drivers arranged above and below the tweeter in a vertical alignment. An extension of this concept is to have a three way system with woofers mounted at the top and bottom, such as in the Duntech Sovereigns.

The first thing you notice about this configuration is the incredibly wide stereo image. Horizontally they disperse in a quite wonderful way. But the trouble is that the vertical dispersion is truly horrible. If one is sitting above or below the tweeter height, the highs are greatly diminished. So a satisfactory listening experience can only occur if the listeners ears are at precisely tweeter height. There are some people that can put up with this kind of fussy behaviour but it I find it totally unsatisfactory.

I love a loudspeaker that sounds great anywhere and everywhere.

The situation gets really weird when we come to centre speakers in a surround sound setup. With a centre speaker you are really wanting a wide image so that the audience's sweet spot is not limited. Using a D'Appolito configuration centre speaker, nine times out of ten the speaker is laid flat on its side. This makes the image high but not wide. In fact the wider the centre speaker the more narrow its image becomes. In other words the four driver centre sounds tiny in width compared to the two driver unit…just the opposite of what we would intuitively believe.

This is precisely the opposite of what a designer should be trying to achieve. Why do designers do such a silly thing? I guess I am as guilty as hell on this one as it is clear that centre speakers must not intrude in front of the picture. The solution is to lay the speaker flat underneath the flat screen. It solves the problem about where to put the centre speaker but acoustically it sucks. So the flat centre speaker is a commercially driven idea…rather than a sound acoustical solution. It makes them more saleable but does not provide an optimal audio result. I am ashamed to say that I made many speakers like this, so compelling is the "logic" of this situation.

Every pattern of loudspeakers has its own special dispersion pattern, but guess which speaker has the best from an acoustic point of view? Well yes, no prizes at this stage…it is the speaker which emulates a point source…the coaxial or twin cone. Now how many centre speakers have you seen with that design? Very few, such as some models by KEF. There should be more.

Anyway, for the next chapter the topic is why speakers used to play well softly, but don't anymore. You think I'm kidding don't you? But I am fair dinkum, I plan to tell you how fashion and size have destroyed modern speakers' ability to sound good when played at low volumes. That's all for this week - the wife has a list of other jobs for me to do!

Email me: ralph at homeentertainment dot com dot au. Questions, agreement, argument, whatever.

The Truth About Loudspeakers - Chapter 7 - Softly, softly - By Ralph Waters

Editor's Note: there's an open invitation to readers to email Ralph on any of the topics he has covered. I think this latest one raises some very interesting points, and I'd be surprised if there weren't a few emails coming in this week to either argue or agree.

I've said before that a loudspeaker should sound great when played loud. This is a sign of low distortion. A distorted loudspeaker will be very fatiguing, aggravating and even painful if played loud. But an even more important characteristic is its ability to sound good when played softly. The fact is that in modern loudspeaker design, the ability to play loudly has often been at the expense of being able to sound good when played softly.

Yes, I know this sounds like a giant dose of nostalgia and sentimentality but quite a few speakers of yesteryear sounded better, and sometimes vastly better when played softly.

You may ask who cares how a speaker sounds when played softly? Well the fact is that much or even most of your listening is very likely conducted at between one quarter to one half a watt input power. And the fact is that speakers of forty years ago sounded miles better at this level. Let me explain.

I have been listening to loudspeakers all my sixty years of life. From the age of eight or so, it became an engaging passion, and pulling apart speakers and listening to the individual components provided a source of endless fascination. So not only did I listen to loudspeaker systems, I spent lots of time listening to tweeters, midranges and woofers in their raw form. And it was incredibly useful and informative to me as a speaker designer. You see it became pretty obvious to me that in the quest for power handling, something got lost in the equation. Both the sound of woofers and tweeters were degraded compared to the older designs at low levels. During the 1960s, most amplifiers could put out no more than about fifteen watts. Some of the valve amplifiers were struggling to put out five watts per channel!

Typical amps put out eight or ten watts at the most. Ten years later that quickly increased to thirty or even fifty watts. Today some relatvely cheap amplifiers can put out one hundred watts or more. Power became much cheaper.

The increase in amplifier power caught most loudspeaker manufacturers with their acoustical pants down as existing technology could not cope with this extra power. The voice coil is the place where the electrical energy is dissipated in the loudspeaker and in the sixties the voice coil consisted of very thin wire wound around the thinnest flimsiest paper called a former. When the wire was overheated, the paper burnt and the speaker ceased to function. Very expensive on warranty claims!

The new era of solid state amplification demanded bullet proof power handling. The answer to this dilemma arrived in two main ways…firstly the wire coating was improved to withstand higher temperatures and the voice coil was made of aluminium which did a far better job at dissipating heat. For good measure heavier wire was utilised which added more power handling. The Taiwanese really got on top of the power handling issue and by the eighties even the cheapest drivers coming from Taiwan were more or less indestructible this side of deliberate abuse. With tweeters, ferro-fluid was added to the voice coil gap and this increased their power handling dramatically.

During this period of time, there was another factor that added to the insatiable quest for increased power. There was a relentless move toward smaller and smaller speakers. In the sixties large speaker boxes with a full-range driver were regarded as the norm. In the seventies big multi-driver boxes thrived … with boxes of twelve inch speakers in abundance. Even twin twelves and fifteens, would you believe? The loudspeaker was supposed to make a big statement, and floor-standing models were everywhere. However from 1980 to the present time, desirable sizes for speakers gradually diminished. This was in the context of larger and larger homes, which is a paradox in itself.

Smaller boxes, though, need a lot more power to get up and go. Yep, a large box is a very efficient device for generating deep large bass. Getting deep bass from a small driver in a small box is only possible by making the cone of the driver much heavier and putting enormously more power into the driver. This phenomenon is a firm fact of physics. There is no way to negotiate or bend the rules…small boxes are less efficient at producing bass than big boxes and that's that. Halve the box size and you need four times the power to achieve the same result in bass performance. Halve it again and that means a sixteen times greater need for power handling.

So the quest for greater power handling and the need for ever smaller boxes meant that the drivers themselves had to dramatically change in the way they were constructed. No longer were the super-light paper cones favoured. They went out and were replaced by heavier paper cones and then even heavier plastic cones. Now, cones have graduated to very sophisticated carbon fibre and Kevlar construction. The voice coil formers gradually became more robust and the wires became heavier and heavier. In fact the total mass of the cone of a loudspeaker made in the last decade may be more than double the weight of its counterpart from the sixties. The tweeters too changed from light paper or silk cones with paper voice coil formers to much heavier gadgets.

All of this had a really terrible impact upon soft playing performance. I first realised this around 1989 when I was playing around with some ferro-fluid tweeters. I noticed that when I played them softly all the soft detail in the violins was missing at low levels. The air which gave string quartets life and breath was gone. Another aspect of ferro-fluid degradation was found in the ambient detail of the tweeter. In any recording space, it is the softest echoes that provide us with the clues about the size of the recording studio. These ultra soft echoes are crucial in a sense of realism and as they spell out the fact that the musicians are located in size and space. What did the ferro-fluid do to this micro information? Well they just erased it completely. So tweeters ceased to become revealing at low levels. Whilst their ability to handle high levels of power increased, their ability to resolve small details at softer playing levels went out the window.

The same degrading trend happened with woofers. The change in construction of the woofers had a tremendously deleterious affect upon bass realism, tonality at low levels. The mass production of high power handling loudspeakers, signifcantly degraded some beautiful attributes of woofer and midrange musicality when reproducing soft passages of music.

With higher power handling came new methods of construction. Woofer cones used to be made by hand. The paper mix was carefully designed and placed into a mold. The better speaker manufacturers left the mix to dry naturally. This resulted in a naturally pulpy soft cone that was intrinsically low in distortion. Later came steam drying of paper mixes and this could be very effective if the paper mix and cone shape were correct. Reams could be written on paper mixes alone, but suffice it to say at this stage that manufacturers such as Magnavox Australia and Plessey Australia made some of the finest loudspeakers in the world in terms of low distortion and soft playing ability.The secret was in the pulp from the unique Australian hardwood. Another manufacturer that really did it right was the original Coral corporation in Japan. Fostex still make sensational drivers using this older approach.

Older methods used hand assembled cones and voice coils, and tighter tolerances could be obtained. This in turn led to more efficient performance in the magnetic gap, whereas machine-made loudspeakers rely on larger tolerances. I found this out in the early eighties …testing lots of loudspeakers. I found that some mass produced woofers with very large magnets actually performed worse than the locally made and produced and hand assembled Magnavox woofer which had a considerably smaller magnet.

The mass of the woofer cones increased dramatically in a bid to get deep bass out of a small box. This required much larger magnets to drive the heavier cones, but even with the more powerful magnet circuits the modern drivers tend to be far less efficient. However, there comes a time when the limits of material strength are reached and then there are problems of distortion and poor frequency response. Modern production required mass produced speaker cones and this implied plastic materials. Heavy plastics require reinforcement such as fibreglass and Kevlar or carbon fibre. But these materials never performed anywhere as well as the vintage method of woofer production.

Another factor which led to degraded sound was the change from paper or linen surrounds at the edge of the woofer cone to rubber or neoprene. The new league of speakers had to be stiffer and less compliant than the old speakers so that they would perform properly in small boxes and so the old generation of 'high compliance speakers' vanished. Linen and paper have a low moment of inertia. This means they have very low initial resistance to movement, while rubber and neoprene have a much higher resistance to movement. This automatically means that a woofer with a rubber surround requires a substantial more power to get it moving, and the whisper-quiet details, the intricacies and delicacies of low level ambience are wiped out. In other words all the subtleties that create a better illusion of reality are gone.

You may not notice any change if you play Pink Floyd flat stick, but if you listen quietly to a chamber ensemble, you will notice the difference immediately. Linen edges and paper edges were able to resolve a mass of detail because they responded to the tiniest signals quickly and easily. Would you use thin silk or rubber washing up gloves if you wanted maximum sensory information?

So the industry traded soft playing ability and big boxes for small boxes and high power capability. In the process some very important things were lost, the main one being the ability of a loudspeaker to resolve detail at low playing levels. We think, quite erroneously, that modern speakers must sound better than speakers forty years old, but the reverse is quite often true. Some of the better speakers of that vintage sound remarkable…and in the area of soft playing usually outperform the modern loudspeaker.

Next Chapter: why those old speakers often fell flat, and the technology from Australia that changed everything!

The Truth About Loudspeakers - Chapter 8 - Anechoically Speaking - By Ralph Waters

Geoff had some reservations about some aspects of my last article. He felt I was being a bit too kind to older speakers, and points out that many old ones sounded pretty awful. I have to modify that statement. Many older speakers sound more than awful, in fact most older speakers sound putrid. There were not too many decent speakers around in the sixties and seventies and there were very good reasons why that was the case. More of this in a moment.

So let me backtrack and qualify my opinion a little more closely and a little more carefully. The soft playing ability of speakers of that era was particularly obvious with a certain kind of woofer that is not manufactured any more. This driver was known as a "high compliance" driver and usually featured either a linen or paper surround. A distinctive of these drivers was that they needed a box some three to five times larger than modern drivers. Due to the light cone, they were fast and detailed, informative, tuneful and often had extended high frequency response with low distortion.

The enclosures required however, were marriage breakers. I recall too vividly the time I made my first enclosure for my high compliance woofer and it was eight foot tall. My long suffering wife refused to have it in the house - which turned out OK since they did not fit through the door anyway.

My love affair with speakers that require large boxes has never ended, but now they are restricted to more reasonable sizes as you can see from the photo. [hey, what about those cute little Fairytales you loved last time - large or small, you love them all. It's a blonde or brunette argument - ed.]

The big box high compliance drivers had some serious advantages and alas some serious drawbacks as well. The advantage is that you obtained incredible efficiency of bass. Several watts of power were enough to draw complaints from a neighbours a block away. The drivers themselves only had to move minute distances and still put out thunderous output with very low distortion. On the downside, they could not cope with large amounts of power; they just couldn't do the larger excursions (movements) if given a wallop from a big amplifier.

Many famous manufacturers made these kind of drivers for a decade or so and then the demands of more power handling with smaller boxes spelled their demise.

Goodmans, JBL, Altec, Plessey. Wharfedale, Coral, Magnavox Australia, Celestion,Tannoy and Foster are some manufacturers that come to mind.

During this era, the gentle art of hand assembly, specially picked fibres for cones and a lot of subjective analysis led to some rare and beautiful drivers. But to be honest, the results tended to be haphazard. Two Coral drivers in my collection are standout winners in the low distortion stakes. Their tonality and accuracy is rivetting. They are better than almost all of the modern drivers in my collection. Yet ironically two of their sister drivers are very much on the horrible side.

Lets get back to the question of why so many speakers of the sixties and seventies sounded so horrible.

The simple fact is that modern loudspeakers have benefited enormously from advances in computer technology. Years ago a speaker company needed sophisticated anechoic chambers to make correct measurements in. This was limited to the well heeled loudspeaker companies such as JBL and ElectroVoice and some English companies.

These companies made loudspeakers which were head and shoulders above other companies and the design effort was clearly evident in the sound quality and their growing popularity. If you didn't have a special chamber, it was very hard to design a loudspeaker to sound any good at all by mere guesswork, or trial and error.

However, in the early seventies something very amazing happened technologically. The use of computers was a total game changer and allowed very small companies such as my own Richter Acoustics, to make extremely sophisticated measurements without the use of an anechoic chamber.

In an instant, the computer would generate a specially shaped noise and compare the noise with the input of the microphone. One could deliberately exclude the effects of room reflections and effectively obtain an anechoic measurement. Not only that, but the computer could allow you to see many other facets such as distortion at the push of a button. All measurements were instantaneous. Gone were the days when it would take you up to two or three minutes to make a single measurement. To my knowledge, I was either the first or one of the first people in Australia to use this technology and it gave me an incredible advantage to be able to make thousands of measurements in a single day. It allowed swift product development if you could maintain concentration yourself. I used the MLSSA system, but there were a number of different systems available.

One of the marvellous measurements that the computer could now make instaneously was the waterfall plot. I will go into this plot in much more detail later, but it was in these plots that the difference between good and marvellous speakers and terrible speakers was revealed. I spent hours pouring over these plots and comparing them to the subjective experience. The waterfall plot and the distortion curves were the two measurements that told me most about how a loudspeaker would sound in the real world.

The other piece of technology that would forever change the world of loudspeakers was with the definitive research conducted by an Australian, Neville Thiele of the ABC. Before this science was developed the performance of a woofer was a hit and miss affair. Before it was screwed into a box the designer would not have any positive idea of what size box, whether it should be sealed, what size vent, etc. All of that was to change, and radically!

Thiele studied the work of J. F. Novak, and established a system of mathematical equations that predicted the behaviour of a woofer before it was even put into a box.

Another Australian, Benson, refined this work with the use of computer simulation. His doctoral student Richard Small went on to further expand and develop the fundamental principles set out by Benson and Thiele. Small then went on to head up research at KEF and later at Harmon.

It was my personal privilege to study Richard Small's papers and to correspond with him. So it was this marvellous new technology that was developed in Australia that completely changed loudspeaker design forever. To this day, any properly designed loudspeaker always comes complete with its Small/ Thiele parameters.

Complete with my cassette based program fed into my Commodore computer, I was able to accurately predict the behaviour of any woofer by a series of simple measurements.

The Sydney based Australian companies Magnavox and the Melbourne based Plessey were quick to take up on this technology and were the worlds first driver manufacturers to make available these marvellous and almost magical parameters.

So in a few short years, any small backyard operation could enjoy the same development tools as the most advanced and well funded companies anywhere in the world.

There was no excuse for poorly designed loudspeakers and the level of design and quality of finished loudspeakers was greatly elevated in a small period of time.

Naturally, Australian loudspeaker manufacturers had a competitive edge now over imported speakers and it was at this time that I happened upon the hifi world and happily designed speakers using this amazing new technology. Many of those speakers were designed in conjunction with Magnavox Australia.

So the era of poorly designed speakers was over?

No not a bit of it. Perhaps the worst era in the history of the loudspeaker was about to emerge. The era of the banger box. More sordid, nasty, exposing tales about this "dark arts" industry next time…. I have to shift the cows into a different paddock.

The Truth About Loudspeakers - Chapter 9 - Bangers & Mash - By Ralph Waters

One of the saddest chapters in hi fi history was acted out in the seventies and eighties. Here in Australia it was a devastating era for designers and for customers interested in well designed, good sounding speakers. I'm referring to the era of the "banger box".

Since the technology for designing good speakers was available, there was really no excuse at all for the tidal wave of crummy speakers. Despite all the technical know-how that was there for the taking and using, this was the worst of the worst times for loudspeaker quality.

Let me explain. Stereo Hi Fi hit the Australian scene like an express train. Australians have always led the world in accepting and adopting new technology. Whether that technology was mobile phones, DVD, iPads or digital cameras, Aussies always see the attractions and are happy to give it a go. The adoption of stereo hi fi was no exception.

The seventies saw an explosion in the number of hi fi shops in Sydney. In just three adjacent streets you could count twelve or thirteen shops. They were always chock-a -block. Everyone wanted a new stereo system. Initially the shops were stocked with the best English Japanese and American brands. These were usually fully imported, but later the loudspeaker cabinets were often made in Australia and then the complete speaker was assembled and packaged. The cabinet-making companies looked at the business in loudspeaker sales and many of them decided that they wanted a slice of the action.

Initially this trend was not too bad. Some of the better cabinet makers imported driver kits from reputable names such as Audax, Foster and Peerless. However the rot quickly set in and these were quickly replaced with a plethora of mainly Taiwanese components. The durability of these components was appalling to begin with , but quickly improved so that they were stable enough to last. But the loudspeakers had zero design input. The Taiwanese learned to emulate the cosmetic appearance of the reputable speaker company products, and soon the market was flooded with good looking imitations. They sold by the truck load. Just one hi fi store in Miranda used to sell up to three truckloads of speakers per week…. or more.

The scale of manufacture of these speakers was hard to comprehend. I knew of four major factories in Sydney alone, with space of 1000m2 or more, making speakers twelve hours per day. Some of the factories had automated lines which stacked, inserted and cut the boxes automatically.

Everything about these no name speakers was truly terrible…The thinnest of chipboards was used. Two component crossovers, with capacitors that were as tiny as those in a transistor radio, were drafted into service. Absolutely no design effort was involved. I was asked to measure some, and +/- 24 dB in frequency response was not unusual. ( + /- 3dB is considered to be acceptable). The drivers were selected according to outward appearance and cost and that was all. The results were as you could have expected….a boomy, noisy box that produced torrents of terrible sound. In this mindset the bigger the box the better, so gradually twin twelve woofers appeared and then a magnet-less (passive radiator) substitute on one to cut down costs.

Bangers with impressive names like "Sweden Acoustics", "Scandinavia Sound" etc. appeared. Invariably they were labelled with deceptive and bombastic names: 'Studio Monitors", " Reference Series" for example. Each hi fi store had their own in-house brand that they bought in quantity or which they assembled themselves.

In a feeding frenzy, many speakers were sold from the backs of vans to unsuspecting buyers who thought that they were buying an expensive studio monitor for a price that was a steal. (This continued to happen on a smaller scale right up until the early 2000s - ed.)

Sadly the main stream reputable manufacturers joined the happy throng, with brands such as Marantz, Technics adopting this philosophy.

The biggest hifi stores were built around the incredible profits available from sales of bangers. In fact a chain of stores originally from Melbourne that trades over 2.7 billion dollars today was founded firmly on the foundation of banger sales.

The biggest hi fi chain in Sydney was firmly founded on banger sales.

The upside of this development was that customers were fed aural sludge for a decade and then decided to buy 'something better'. I was fortunate to enter the market with my Australian company Richter Acoustics at this stage and we seriously challenged the supremacy of British and American speakers as the only legitimate well designed loudspeakers on the market. This also corresponded with an intense desire by Australians to 'Buy Australian', a most welcome development, at least for a while.

Since that time, the banger has all but vanished. The monstrous twelve inch three way has thankfully, pretty much disappeared from the marketplace. Their undead still haunt the online secondhand sites!

But instead we have substituted mp3 fed tiny speakers with nondescript sound at best. Have we progressed, or simply swapped one aspirational deception for another?

The age of the banger was perhaps the biggest commercial scam in Australian history. Fortunes were made on the back of this idea. The vast majority of the speaker makers have vanished from the market. A couple have morphed into something more reputable. Thankfully this sad era is behind us.

Next issue I will look at horn loudspeakers. ….Horn loudspeakers seem to invite the psychotic to the exotic and perhaps even the erotic. The strange behaviours of horn devotees is well known to insiders, but perhaps not entirely documented for the world at large. To be continued.