Tannoy: Profile of a Loudspeaker Collosus
First published in 1995 in Taiwan, here is a feature I wrote about Tannoy, which included a revealing interview with the company’s Technical Director, Alex Garner.
The company, its products, its sound
Guy R. Fountain founded Tannoy in 1926. The company manufactured rectifiers – for charging lead acid batteries – made from an alloy of tantalum and lead, hence the name Tannoy. It established its name as a loudspeaker producer by supplying all the re-entrant horn loudspeakers used for the public address systems on RAF airfields during World War II.
The Dual Concentric driver, whose arrival marked the company’s entry into direct radiating loudspeakers, appeared in the mid-1950s. It came about because Bertram Mills’ circus needed loudspeakers with a wider bandwidth than airfield horns and which could generate higher sound pressure levels. The Decca recording company subsequently adopted the drivers used in these for use in its studios.
Fountain, who had a considerable talent for marketing, cashed in on Decca’s use of his speakers by producing a domestic range, including models such as the Autograph and the GRF corner enclosure. Around the mid-1970s, when hi-fi began being manufactured in larger quantities than had ever previously been done, Fountain sold the company to Harman Industries.
After a management buy-out at the end of the seventies the company continued its work in the professional arena and stepped up its activities in the domestic hi-fi market. It has been active – and successfully so – in both ever since. Tannoy has consistently demonstrated considerable design flair and technical innovation. It has conspicuously avoided “me too” products and has manufactured loudspeakers whose aesthetic and musical presentation have helped them stand apart from the bulk of their mainstream competitors. At the top end of its range there are exotic models such as the Westminster. In the more affordable market it released the Sixes series, a range of speakers in striking hexagonal cabinets, which subsequently evolved into the Profile series.
If one quality characterises Tannoy’s “sound”, it is drama. As explained in the interview that follows, the company does not produce loudspeakers to provide background music!
Alex Garner
Technical Director, Alex Garner heads the design team at Tannoy’s Coatbridge headquarters. He has been with the company since 1976 and was instrumental in the management buy-out that regained control of Tannoy in the late seventies. Although he comes across as an easy going, affable character with a regularly exercised sense of humour, he’s also clearly a single-minded, highly motivated individual. His demands for the performance he expects from loudspeakers leave little room for anything but unavoidable compromises, although he’s sufficiently pragmatic to admit that a designer cannot ignore commercial imperatives and the laws of physics. He nonetheless has very clearly resolved views regarding what a loudspeaker should and should not do, something that’s reinforced by the strongly familial nature of Tannoy’s products.
Garner is well respected by his co-workers and industry peers. Steve Harris, a former colleague and now the Managing Director of loudspeaker manufacturers Mordaunt-Short describes him as “an absolute delight to work with from a marketing viewpoint. His strengths are that although he fully understands the market and commercial considerations he’s always driven by a desire to produce the finest possible products. I’ve been lucky to work with some outstanding designers but Alex is truly exceptional. If I had to name the two most brilliant people I’ve worked with it would be him and Ken Ishiwata (of Marantz), a man I hold in the highest esteem. On a personal level Alex is far too modest for his own good and, unlike many talented people, he’s certainly not a prima donna. He’s easy to work with and a thoroughly pleasant chap.”
This sentiment was echoed by Tannoy’s Product Manager, Ken Weller who added that “Alex is one of those rare people who fully understands the technical intricacies of speaker design and the black magic involved in making hi-fi enjoyable. What’s just as important is that despite spending many years in this business he’s still committed and enthusiastic – he’s not become jaded or cynical.”
And so to the interview proper…
MS: Where does the famous Westminster feature in Tannoy’s evolution?
AG: After the management buy-out, I spent a year organising the design team and we began work on the early Prestige range – the Sterling, Edinburgh and GRF Memory. These were aimed primarily at Far Eastern markets to demonstrate the craftsmanship and performance Tannoy could produce. We decided then to make a ‘statement’ speaker, one that encapsulated all the things that Tannoy was good at – horn loading, compression drivers, high output and wide bandwidth. The result was the Westminster.
MS: How does Tannoy’s history impinge upon the company today?
AG: It’s complicated, affecting the company and its products, but most importantly our customers. They have always been people who could appreciate our type of engineering. Customers in the period between, say, 1975 and 1985 were people who didn’t want the “standard” infinite baffle speaker of that time: they wanted something different – something more efficient, powerful and prestigious in terms of engineering content and craftsmanship. These people were traditionally in the areas around Japan – Korea, Taiwan and to some extent the USA – and so we were engineering for a non-UK market, which meant that we built up a distributor base in these areas that has been extremely useful to us to this day.
MS: Do professional and domestic buyers expect different things from a Tannoy loudspeaker?
AG: Yes and no. They’ re both looking for quality but there’s a small difference in what each type of customer needs. When a domestic customer buys a speaker for “pleasure” listening he doesn’t want it to be totally ruthless, whereas a professional monitor has to expose everything that’s going on. Effectively when you’re engineering for the domestic market you’re usually constrained by some sort of budget and so you engineer towards the customer’s viewpoint: he won’t be operating at high levels in a large room; his room acoustics are uncontrollable compared to a studio where you know that the RT time will be low and there won’t be awful echoes. You’re also more interested in the speaker’s off-axis response, which isn’t the case in a studio where the engineer is always sitting directly on-axis.
MS: Talking as a designer, tell me what makes a great speaker – in both performance and technical terms? And what makes a lousy speaker?
AG: Let’s take the lousy speaker first because that’s easy to deal with. The lousy speaker is usually made by a chap who thinks he knows what he’s doing. He buys ordinary OEM drive units, sticks them into a cabinet and then tries to sort out the horrendous difficulties the drivers have by playing with the crossover. This usually results in a crossover that’s either complex or wrong, so destroying any acoustic worth the drivers originally had. Cabinets are also vital and if the drivers are in a cabinet that’s not especially rigid or well understood then the speaker will be lousy. A simple test tells you whether a speaker is any good: if music sounds flat and boring the speaker is lousy.
To make a good loudspeaker you need control of all the elements it contains – the drivers, cabinet and crossover. You don’t have control over the rest of the hi-fi system but you need to design the speaker with an awareness of what’s likely to partner it.
Ultimately I like speakers that I can relax with and enjoy, designs that can reproduce the entire musical spectrum without being caught out. One basic test I employ is playing good BBC speech from a tuner. Although they do some odd things with microphones in studios these days, generally a microphone won’t introduce much coloration. It certainly won’t introduce the colorations that a loudspeaker will. If a speaker reproduces speech well, it’s off to a good start. I often ask young designers to stick their heads inside a cabinet and talk. This sounds crazy but some cabinets echo like buckets: how can you reproduce music using a box that distorts a human voice?
MS: What are the technical prerequisites for a great loudspeaker?
AG: It must have the classic flat – or reasonably flat – frequency response on-axis. By flat I mean that it has to translate the energy in all parts of the audio band to within 3dB. The response might rise or fall gently but it mustn’t have any gross undulations. Off-axis response is equally important because eighty per cent of the sound you hear in an ordinary room is reflected. It must be smooth and even so that as you move off-axis there are no major dips. This is where the majority of speakers go wrong in that you have directivity of drive units, the critical area being around the crossover point. When you listen off-axis they have a big “hole” in their response, right at the point where the music is.
Great speakers also don’t do things they shouldn’t. In other words, if you switch off the input suddenly nothing should come out of the speaker. Normally all sorts of hash and noise emerge, resonances at different amplitudes across the audio band. You have to ensure that the tail of the impulse response is small and that the Q of the energies coming out is low.
The impedance is also important. It’s no good coming up with a perfect loudspeaker if you can’t find an amplifier that will drive it! 3.8 ohms is a realistic minimum.
Phase, to which very few people pay attention, is another important consideration. I am convinced of the worth of having a particular type of phase response, one where the rate of change of phase with respect to frequency is constant. The speaker then acts simply as a time delay over the whole of the audio band – just as the air coming from a live instrument behaves. Our speakers aren’t perfect in this respect but they do have a relatively flat response. Because the DC unit is effectively a point source – and provided you don’t do anything silly in the crossover – you can make the speaker act like a constant time delay and provide great depth to stereo imagery with images “hanging in space” rather than the flat, two-dimensional images many speakers provide.
MS: I’m curious how you voice a speaker so that it has a flat response on-axis and when it’s in a room. Surely a flat speaker won’t stay flat when its output is reinforced at some frequencies and reduced at others?
AG: If you placed a grand piano in a room it would also suffer the same response variations because it’s a wonderful omni-directional source spraying sound everywhere. Much of the character of a live instrument has to do with that off-axis energy. A speaker has to preserve that off-axis energy and if its response is grossly non-flat it can’t do that. You end up with the room emphasising frequencies where there’s a lot of output from the speaker and not emphasising those where there’s no off-axis output, and it sounds odd.
This was one of the tricks Wharfedale’s Gilbert Briggs played back in the 1960s where he did live versus recorded comparisons. He chose instruments like the oboe where most of what the audience heard from the instrument was direct sound. The he replaced the oboe with a speaker and they heard much the same. That was extremely clever because you can trick people quite easily when you have everything else under control: unfortunately, when you’re selling hi-fi that’s not the usual situation!
A directional loudspeaker probably isn’t the ideal source unless you’re talking about a studio monitor or PA speaker, but for domestic hi-fi you do need an element of good off-axis energy. Arranging drivers along a line on one face of a cabinet compromises the speaker’s dispersion characteristics, which is why Tannoy favours using the point source Dual Concentric unit wherever possible. When we use conventional drivers we take great care to arrange them – and the crossover – so that their dispersion characteristics are as near ideal as possible.
MS: How about the delayed energy problem?
AG: There’s a whole world of different loudspeaker sounds caused by delayed energy, but the really important factors are what the driver diaphragm is doing and what the cabinet is doing. With the drive unit you need to ensure that the cone or dome’s termination and damping is optimal, so that all the energy travelling through the cone body is absorbed. It’s a very complex engineering situation.
The other area – and one that’s perhaps more important – is the cabinet. Typically you have, say, an eight-inch driver in a cabinet whose radiating surface is ten or twenty times larger. As the sound radiated from anything is proportional to the square of its area a tiny resonance in the cabinet will create a very large problem. You need to reduce the amount of energy the cabinet panels can store. There are two ways of dealing with this: limiting the amount of energy transferred into the cabinet from the drive unit, and enabling the panels to turn this energy into heat before re-radiating it. You do this though various cabinet construction techniques such as internal bracing and lossy adhesives.
MS: Are you saying that you could dramatically change the sound of a loudspeaker simply by assembling the cabinet with a different glue?
AG: Absolutely! It would sound totally different. Cabinets need to be rigid – up to a point – but they also have to be energy absorbing. For example, people make cabinets from inch-and-a-half Birch ply panels, with glued and screwed joints using rock hard adhesives, fit rigid internal bracing and then wonder why the speaker sounds awful. It’s because the cabinet is a high Q structure. It doesn’t have much delayed energy apart from at one or two points in the audio band at which the whole box goes off like a drum skin.
MS: What’s the nearest to a no-holds-barred design brief you’ve tackled?
AG: I’ve done a number, including the Westminster, which I designed because I knew that the waveform coming from a compound horn could be superb. Others were studio loudspeakers that needed to go very loud yet retain their sound quality – the Buckingham and the active Dreadnought. Truly though, there are no “no-compromise” situations. Engineering is always a compromise and, on a practical note, speakers have to be small enough to go through doorways and fit into the back of a Volvo estate car so that I can take them home to try them!
MS: What other manufacturer’s speakers do you regard as exemplary designs?
AG: Naturally I’m not going to advertise my competitor’s speakers here but I’ll admit that I’m very fond of the QUAD Electrostatic and I’m fascinated by certain omni-directional designs.
MS: Given the individual nature of many Tannoy designs do you support the notion of product ranges?
AG: Ranges are a commercial necessity because customers and retailers expect them but there are inherent engineering problems with them. For example, people expect to get more bass performance as they move up a range of speakers but a speaker’s LF response is a function of the diameter of the bass unit and the volume of the cabinet. As you move from a 160mm driver to a 200mm driver to a 250mm driver in progressively larger cabinets you don’t get dramatically more bass because the bigger cabinets aren’t disproportionately larger, which they need to be. They’ll handle more power but you need to change other aspects of the design, such as pulling back on its sensitivity, to gain the improved bass that buyers expect from the typical larger cabinet.
MS: What’s your primary role at Tannoy – a strategist, trouble-shooter, manager or a hands-on designer?
AG: Yes! I do a lot of planning at board and group level so that makes me a strategist. I’m definitely a trouble-shooter because I’ve been with the company so long I’ve encountered all the problems there can be with engineering and producing loudspeakers. I’m a manager up to a point, although I don’t manage in a typically hierarchical fashion. I like to give people room to develop their ideas, which fosters creativity. A hands-on designer? Not as much as I’d like to be. There are times when I yearn to attack a sheet of MDF with a Black and Decker!
MS: What technology do you think is most likely to transform loudspeaker design over the next decade?
AG: I imagine it will be a combination of cabinet design and DSP techniques. I also suspect that active loudspeakers will become more popular than they are now. Mainly, though, loudspeakers will continue to develop through evolution, especially when new materials spawn alternative approaches to manufacturing components.
