Ben Duncan continues his look at monitoring speakers

Last month, we dipped into the physical reality behind those innocent looking loudspeaker sensitivity figures. The accompanying table 1 gives SPLs for a variety of power inputs over a range of typical sensitivities. The percentage efficiency of the transducer in converting amplifier watts into acoustic energy is also indicated to add new perspective to the difference between say, 83 and 92dB sensitivities. You can check these figures off against a manufacturer's specification sheet and so attain a better idea of how 'loud' a particular speaker will go in comparison to another.

However, you can't expect the levels quoted to be accurate in an absolute sense. First, the acoustic can raise the apparent SPL with, say, standing waves. Then, speakers may have special nearfield characteristics in which their acoustic relationships are abnormal. There are also compression effects, and last, the speaker's dispersion characteristics which come into play. And it's these latter three that we'll look at next.

The Nearfield

This is the region in front of the speakers where direct energy predominates, as opposed to sound reflected from the boundaries of the room. Figure 2 shows (albeit in a very simplistic fashion) how the reflected sound or 'reverberant field' swamps out the direct energy as the listening position becomes more distant. Of course, the implied line on the diagram where the nearfield ceases and the reverberant one begins is a fiction: in reality there's a continual progression. A degree of reverberant energy is essential for realism, but at the same time, reflected sound is highly coloured, so it's important for the more accurate direct sound to predominate. A sensible compromise is to ensure that reflection is suppressed between the speakers and the listening position, whilst reverberation is given free rein beyond this point (Figure 3).

This treatment is known appropriately as 'Live end - Dead end' (LEDE), as pioneered by Don Davis, an American acoustics consultant. Leaving the speaker-room interface as a future topic, the important thing to note is that nearfield behaviour is crucial in respect of the small control rooms and consequent close-up listening position common to musicians working in Europe. Moreover, abnormal nearfield behaviour isn't just a function of the room, but also something a speaker exhibits in its own right.

In particular, the nearfield behaviour of large, horn-loaded bass speakers includes a region of between two and twenty feet from the speaker mouth (always assuming suitable room dimensions) where the sound wavefront coalesces. The bass in this zone can be very coloured, and there may seem to be 'no bass' despite the fact that the walls are shaking! So, unless a bass bin is a type with bass that 'comes together' at short distances (ie. no more than three or four feet from the mouth) it's unsuitable for control room purposes.

Direct radiator speakers of course don't exhibit this peculiar 'Near-nearfield' region, neither do Thiele (pronounced 'Teel') cabinets, nor midrange and top horns to any significant degree. Another nearfield bass bin anomaly to look out for is when the apparent source of, say, the low midrange travels up the mouth of the horn. In other words, as you approach the speakers, the apparent source of the sound over a specific range of frequencies will shift. All bass bins do this to an extent - it's a tradeoff balanced by the spectacular spaciousness and depth they offer - but nevertheless, the effect shouldn't manifest itself to any annoying degree at typical listening positions.

Compression

As you drive a speaker harder, the temperature of the voice coil rises - to surprisingly high levels. At full power, it will be around 200 to 300 degrees C (depending on the coil-former, material and the adhesives technology employed). At these temperatures, the voice coil's copper wire has approximately double its room temperature resistance. The effect has four ramifications.

One, it reduces - indeed, typically halves - the power input to the driver, so affording a measure of self-protection. Next, it gives rise to dynamic colouration if one driver in a 2 or 3 way system has limited power handling capability, and goes into thermal limiting well before the midrange (and it's invariably either the bass or the top-end driver that succumbs first), causing the tonal balance to be badly upset.

This effect has had some people worried about hearing damage, when, after several minutes of loud monitoring, more and more top-end has to be put on to maintain balance. Although SPLs capable of hearing damage do cause the ear's HF response to shut down, it's more often the tweeters that are suffering excess power! And of course, this is an easy thing to check for when you evaluate a potential purchase.

Thirdly, because real speakers exhibit capacitative and inductive reactance (ie. frequency dependent resistance separate from the change of DC resistance brought on by cooking the voice coil), thermal limiting doesn't result in an equal reduction of SPL across the spectrum. For example, at the upper end of each band covered by direct radiator speakers (ie. just below the upper crossover point for each driver) the load seen by the amplifier is largely inductive, so high power levels won't increase the overall impedance to the same degree as lower down. The result (a surfeit of hard, biting upper-midrange) is well known... and not kind to the ears either.

Lastly, thermal limiting effects largely set the maximum power input, and in a tautologous manner, the absolute maximum SPL indicated for speakers with high temperature voice-coil technology will be typically 1 to 2dB lower than the 'power input versus sensitivity' figures imply. Thus, when input power is raised by X dB, the acoustic output rises by a figure less than X - it's for this reason that these thermal effects are known as 'compression'; the audible effect being just like that of the rack mounted effects of the same name.

Dispersion

For sound coming directly from the speakers, the driver's dispersion pattern seems unimportant, on the assumption that siting the monitoring to aim directly at the listening position is common sense. When two or more people need to listen, the off-axis sound becomes more important, but it's important in any case, as off-axis energy goes to make up most of the reflected sound and the reverberant field.

From another angle, dispersion should be near enough equal at all frequencies, otherwise the stability of the stereo imaging will suffer. The best 'stereo' comes from large electrostatics and certain horn-loaded units, and it's no coincidence that these offer a relatively constant directivity over a wide band of frequencies.

Dispersion of between 90° and 130° is agreed in some quarters to be the optimum for stereo, but wide dispersion per se is less important than a directivity pattern which doesn't alter radically with frequency. Below some 250Hz, our hearing loses its sensitivity to direction, so the above comments has little relevance to bass speakers.

Dispersion is most crucial in the midrange, and bearing in mind the poor HF dispersion of direct radiator speakers with large cones, satisfactory imaging in the upper-mid demands small speaker diameters - say 8" or less. You can now see why micro speakers, with say, 4" drivers and otherwise execrable characteristics sound so good - if only in stereo!