The final element in the chain of sound reproduction is the loudspeaker. Stephen Dawson looks at what integrators need to know, what technologies are on offer and how they interact with components in a system.

In the last edition of The Audio Philes, we spent a bit of time kind of deprecating the importance of high-power output from amplifiers. That’s because when used in the home, most amplifiers, most of the time, with most loudspeakers, are typically producing only a few watts of power.

That said, with some loudspeakers more power is required, and if you occasionally party hard, then you may need many more watts.

But we also noted that even if you double the power output of an amplifier, that will at best increase the actual output level by just three decibels. You can hear the increase, but it’s not a huge change. If you’re buying a system, consider an alternative: buy high-sensitivity loudspeakers. If one set has a sensitivity of 3dB more than another, then the result is roughly the same as doubling the amplifier power.

We will return to loudspeaker sensitivity shortly, including the shocking truth about their efficiency. But first, let’s consider the types.

Boxes

The most common loudspeaker designs are of course one or more speaker drivers contained in some kind of a box. The box isn’t just for convenience, it forms part of the loudspeaker and particularly controls the bass.

A large loudspeaker driver simply won’t produce much bass if it isn’t in a box. Just about all drivers, whatever their design principle, involve moving some form of diaphragm backwards and forwards, creating repetitive pressure waves in the air. As the diaphragm moves towards the front, it compresses the air in front of it and rarifies the air (reduces the pressure) at the back. Obviously, the compressed air will rush around to the back if it has enough time. At high frequencies, it doesn’t have enough time because the diaphragm starts moving the other way before a significant amount of air can undertake the movement.

But at bass frequencies, that’s precisely what happens. And that “leakage” of air reduces the strength of the pressure wave, weakening the strength of the bass. The lower the frequency, the weaker the output becomes.

So, loudspeaker drivers are installed in a baffle. That’s a flat sheet of something or other. It’s also the name we give to the front panel of a loudspeaker. The larger the dimensions of the baffle, the longer the path for the air movement that would weaken bass performance, so the deeper the bass that can be delivered. An infinitely large baffle – or at least one several metres wide – would provide plenty of bass, but these are clearly impractical.

Which is why the baffle is folded around into a box. And, indeed, if the box is sealed – no air can leak into it or out of it – then such a loudspeaker is sometimes called an infinite baffle model. And while it totally isolates the front and rear pressure waves from the driver, it introduces a different problem that also reduces the bass. As the diaphragm moves back, it’s trying to compress the air contained in the box, so the air acts as kind of a cushion. As the diaphragm moves forward, it tries to reduce the pressure inside the box, which increases the cushioning effect. The deeper the bass, the more the diaphragm needs to move forwards and backward, and thus the more the cushioning effect of the air contained by the box reduces output. Indeed, another name for such loudspeaker designs is acoustic suspension.

This isn’t as bad as it seems. Bass drivers intended for this application are designed to have a softer inherent suspension, since they’ll be relying on the cushioning effect of the contained air. The output reduction is relatively mild compared to most other designs: around 12dB per octave below the -3dB “shoulder”.

The more common loudspeaker enclosure design for high-fidelity loudspeakers has an intentional hole in the enclosure, allowing some air to move in and out of it. These are bass reflex designs. Of course, the hole isn’t just any old hole. They typically have a tube attached, and through careful design (using analytic techniques developed by engineers Thiele and Small in the 1960s), the tube has a length and volume designed to work with the internal volume of the enclosure and the “compliance” (kind of the inverse of the stiffness of the driver’s suspension) to produce a low-frequency resonant output. The net effect is to extend the natural bass response of the driver below what it would otherwise be.

The major downside of this is that the bass rolls off even faster below the point of that extended response, at 24dB per octave.

So, for similar driver designs and enclosure sizes, bass reflex delivers a more balanced bass response than an acoustic suspension design down to a certain point. But below that point, the output falls off rapidly, and eventually, the other design can produce some of the really deep stuff more effectively.

There are variations in bass reflex. Passive radiator designs use a carefully chosen sprung and weighted diaphragm to provide a similar functionality to bass reflex, while transmission line speakers use a very long, but narrow, labyrinth within the enclosure to provide a bass resonance. The now-defunct but long-lived line of Bose Wave Music Systems employed this to produce surprisingly effective bass from compact units.

Drivers

The devices that actually produce sound in a loudspeaker are drivers. The job of a driver is to turn an electrical signal into the aforementioned pressure wave in the air. The great majority do that by moving a diaphragm of some kind backwards and forwards in sympathy with the signal.

The most common form of driver is the dynamic driver. These simply use a coil affixed to the back of a (generally) cone or dome, immersed in a strong magnetic field, forced to move by the signal running through it. Variations include ribbon tweeters, which still use magnetic interactions, but employ these to make a concertina-like assembly squeeze air out. These only really work for higher frequencies, and in that role are generally highly regarded.

Somewhat different are electrostatic drivers. These employ two highly charged “stators”, which are perforated planes of conductive material, one in front of and one behind a large, extremely thin, plane to which the signal is fed, boosted to a high voltage. Instead of being magnetically driven, the force applied here is electrical (like charges repel, opposites attract). The thin moving material is too flimsy to withstand the stresses that would be imposed upon it by mounting in any kind of enclosure, so these are bipolar drivers: the sound comes from both sides, each fully out of phase with the sound of the other. So even though electrostatic speakers tend to have very large diaphragms, they are not so good in the deeper bass. Many electrostatics are hybrid, with bass being handled by a conventional dynamic driver in an enclosure.

Size matters

In theory, a single driver could produce all the frequencies – from 20Hz all the way up to 20,000Hz – generally required of a high-fidelity system. In practice, it’s not so easy.

If you feed a treble signal to a large driver, the sound tends to beam with very little dispersion (electrostatic speakers don’t have this problem, because the equal-but-opposite emanations from the rear of the speaker tend to diffuse the sound).

That’s why tweeters are small, typically 25.4mm domes.

But small drivers have to work very hard to produce deep bass. Bass involves moving a lot of air, and that can be done either by using a smaller diaphragm and moving it a long way, or by using a larger diaphragm and not moving it so much. Moving the diaphragm a long way – the movement is called “excursion” – increases distortion. So, more surface area for bass – whether using large diameter drivers or multiple smaller ones – is pretty much essential.

In short, it’s effectively impossible to make a decent-sounding full-range loudspeaker using a single dynamic driver. So, in virtually all high-fidelity loudspeakers multiple drivers are employed and each is fed the frequencies appropriate for its size. A two-way speaker has two or more drivers, some of which are used exclusively for higher frequencies and some exclusively for the lower frequencies. A three-way speaker adds a midrange driver, while it won’t surprise you that a four-way loudspeaker has four frequency bands handled by different drivers.

There are sometimes compromises. For example, 2.5-way loudspeakers have a tweeter for treble, one driver for both bass and midrange, and one or more additional drivers for bass alone.

Sending the right frequencies to the right drivers is the job of the crossover network, a collection of resistors, capacitors and inductors built into the speaker, which apply frequency-based filters. Crossover network design and component quality have a significant impact on sound quality.

Many high-fidelity speakers provide two sets of terminals to facilitate “bi-wiring”. The idea is that instead of one set of speaker wires to each speaker, you can use separate wires for bass and treble. When used, the conductive straps between the two sets of terminals are removed.

This is one of those clever-sounding ideas that’s really utterly ineffective. Decent speaker wire has vanishing low resistance, so both terminals remain connected, just via the two sets of speaker wire instead of the strap.

I sometimes think that the near ubiquity of bi-wiring terminals is because of confusion with bi-amping, in which separate amps are used to drive the high and low-frequency drivers. But these are actually entirely different. In bi-amping, the crossover needs to be implemented before the signal hits the power amplifiers, and the internal crossover network in the loudspeaker needs to be entirely disconnected.

Efficiency

I would estimate, based on the speakers passing through my testing room over the last couple of decades, that the average sensitivity of a high-fidelity loudspeaker is 89dB. That is, if the speaker is fed a 2.83V signal (that’s one watt into eight ohms), the sound pressure level measured at one metre is 89dB.

That sensitivity can be expressed differently, as efficiency. That is, for a given amount of electrical power sent into a loudspeaker, what percentage of it emerges as acoustic power? The answer for an 89dB-sensitive loudspeaker is… 0.5%.

Yes, that’s right. For the typical high-fidelity sound system, 99.5% of the power sent to speakers by an amplifier is simply turned into heat in the speaker. The rest is the sound that you hear.

What if you buy relatively sensitive loudspeakers? 92dB-sensitive loudspeakers tend to be so regarded. That works out to 1% efficient.

All this stuff is logarithmic. At the extreme end, the mighty Klipschorn loudspeakers are rated at 105dB sensitivity, which works out to 20% (although some reviews suggest that the 105dB figure might be optimistic).

The problem is what engineers call an impedance mismatch. Maximum power transfer is achieved when the source impedance and the impedance of the receiving device (air in this case) are the same. Horn-loaded speakers reduce the mismatch and thus tend to be more efficient.

But, as with all the other things mentioned in this article, there’s no hard rule for which is best.

This is why, in the end, the ear is so important in choosing loudspeakers.