Ahh, the mysteries of the wooden box, aka the loudspeaker. There are a couple of ways to consider loudspeakers.
On the one hand, you can simply think of them as yet another set of boxes in the signal flow, and actually, pretty generic boxes at that. Like, they're black. They're usually wooden (actually, MDF). Got a bunch of drivers in 'em. Expensive ones got amps. Big ones play. Ones that play in serious bars got metal grids on 'em to protect the drivers from beer bottles, flying debris and the larger varieties of insects and birds.
Or, you can think of them as laboratory instruments that are supposed to reproduce a former acoustic event as accurately as possible. The idea is that the speakers should have perfect reproduction, so that all the instrument and vocals sounds sound, like, totally real and also, like, totally awesome.
Or you can think of loudspeakers the way I do – as musical instruments. That is, devices that make music. When you do it this way, even if it feels weird at first, a lot of things begin to come a lot clearer as you go along.
When we think of them as musical instruments, we can generally regard the music that is played through loudspeakers as musical transcriptions (equivalent to a flute piece transcribed for piano, for instance). These are transcriptions of sound that was in other mediums (voices, orchestras, guitars, drums, etc.). What we want is for the loudspeaker to mimic those other sounds, so that what comes out of the loudspeakers is a satisfying and entertaining mimicry of the music that the other instruments made.
I've already told you a little bit about the role the room plays in this (TV Technology, June 13). Now we need to talk a bit more precisely about how the loudspeaker works.
A single loudspeaker is pretty straightforward. It is, as noted before, a box with a couple of drivers in it. One plays high stuff (the tweeter), one plays low stuff (the woofer). There may be a middle driver as well (the mid range).
Each driver can cover about three octaves of frequency (an 8:1 range). The wavelengths produced by each driver usually range from being much larger than the driver size to roughly the same size as the driver. When the wavelengths are much larger than the driver, the output radiation of sound is omnidirectional. When the wavelengths approach the same size as the driver (and its enclosure), the output becomes directional.
The amount of loudness (actually, amplitude) that can be generated by a loudspeaker is determined by the amount of air that can be displaced over time. At low frequencies, this means lots of air, which means a big loudspeaker.
The drivers are subject to a couple of other problems.
The first is that they are not very linear devices, which is to say that they are prone to significant amounts of distortion over a large part of their frequency and amplitude ranges.
The second is that their output is not uniform in all directions, for the reasons mentioned above and the physical characteristics of the drivers themselves.
The third is that drivers are dreadfully inefficient (between 1 percent and 10 percent efficiency at best), so they generate lots of heat and require significant amounts of power to work well. The heat, of course, makes them change behavior, so they sound different when they are hot than when they are cold.
And that leads to a fourth quirk, called power compression. For any given driver, as the input power goes up, above a certain point, the output power fails to keep going up at the same rate, so that the driver's amplitude range becomes compressed, particularly as it heats up.
Oh, dear, as Mother used to say. This means that speakers function as multiband compressors, at high levels. Didn't know that, did ya?
A final quirk is that loudspeakers use a fairly complex filter network, called a crossover, to divide the spectrum up and send the appropriate parts to the appropriate drivers, Sounds simple, I know. It ain't! The collective behavior(s) of the crossover, the drivers, and their radiation into an acoustic space turn out to be highly complex. It's a tough world out there.
What this all should make clear is that loudspeakers, are difficult, nonlinear devices that exhibit highly variable performance. They all sound different – quite different – and all of them interact with their environments differently. They are sufficiently tweaky that it is quite difficult to build two loudspeakers that are similar enough that we can't tell them apart.
This is, of course, exactly like other musical instruments. A good pianist has no trouble at all telling pianos apart, and will have his or her favorites, just like old friends. In my studio, the right main monitor has a couple of interesting little mid-range quirks I've come to know all too well. One of my nearfields has a slightly fizzy top end. Individual speakers have personalities and character.
Right now I'm deep into listening trials for a new studio monitor that I am involved with bringing to market. Our testing protocol at this point in the development process is to take the monitors into a variety of recording studios and get the engineers there to listen and tell us what they think. A typical listening day has our prototypes up on a meter bridge along with an assortment of other small monitors, plus the main monitors installed in the front wall of the room. Listeners compulsively switch back and forth between speaker pairs, trying different tunes, tracks and levels.
To me, those different speakers really sound like different voices. Changing speakers sounds to me like asking a different singer to sing the tune. It's that significant. In the brief few hours that we spend sussing out an array of speakers perched on a meter bridge, they all come to have quite distinctive voices, readily and easily identifiable, as in, "Gawd, the Blatters really sound forward in the midrange. And listen to that weird 3K edge on the Mitsukis."
Which is best? This is highly subjective, naturally, but in general it is safe to say that the speaker that most openly reveals the sonic gestalt of the recording, the speaker that is most easily "heard into," at the widest range of levels and revealing the greatest range of depths and details, will be the voice that is preferred.
So, what we really have to come to terms with is that loudspeakers are sufficiently individual and distinctive that it is really dangerous to think of them either as black boxes in the signal flow or precision laboratory instruments that will precisely reveal what the microphone captured. The physics says no, and our ears tell us no as well.
Next time, I'll tell you about the mysteries of stereo pairs of loudspeakers. Here's where it gets really interesting.
© 2001 by David Moulton
Get the TV Tech Newsletter
The professional video industry's #1 source for news, trends and product and tech information. Sign up below.