The Brave New World: Loudspeakers All Around!

As I noted last month, audio authority John Woram has said, "Audio is a hallucination where the sufferer believes that he or she hears music coming from a small wooden box."

Woram went on to observe, "Stereo is a more serious form of the hallucination, where the sufferer believes that he or she hears music coming from a point in space between two small wooden boxes."

Alert readers will recall that last month we talked about that quirky but quintessential audio device, the loudspeaker. We noted that it is, in and of itself, singularly quirky and difficult. And by that quirkiness, the loudspeaker makes a mockery of the idea of precision or accuracy of reproduction. Too many problems with drivers, crossovers, directionality and, dare I even say it, rooms!

But there’s also a lot more to it. Loudspeakers exhibit, when played in pairs or groups, a remarkable and unprecedented behavior that is at such odds with nontechnological acoustics that it absolutely confounds our hearing. What’s even wilder is that we seem to be oblivious to the existence of that behavior, right in front of our noses, even while we’re wallowing in the pleasure we derive from it! Go figure!

Consider it this way. All normal sound sources including all conventional musical instruments vibrate freely in space, emitting waves of pressure change into the space. When I say they vibrate freely, I mean exactly that. Each source vibrates as a function of its mass, tension and size. Each vibrates independently of all others, and there is essentially no linkage between the vibration of one source and the vibration of any other source.

Our hearing makes use of this behavior. We are immensely sensitive to the minute variations in frequency that occur between sources, which we call beating, chorusing, phase-shifting and flanging. This behavior, whatever we call it, is the basis for how we hear that there are multiple and different simultaneous sound sources.


When we listen to a single sound in a reverberant room, all the reflections of that sound have the exact same rate of vibration (albeit later in time) as the source. The term we use to describe this is "phase-locked," and it has important implications for our hearing. In fact, we identify all "early" (less than approximately 50 ms) reflections of a given source with that particular unique source because of their particular unique phase-locked relationship with that source.

When, on the other hand, there are two sound sources, we have two families of phase-locked sounds that are "beating" with each other. Even if they have "identical" timbre and loudness, we have no trouble determining (a) that there are, in fact, two sources, and (b) where they are in the reverberant space, thanks to those independent families of phase-locked sounds.

This is where the behavior of loudspeakers gets weird. Loudspeakers are, of course, driven by an electrical signal that consists of alternating current flow in a wire (actually two wires). The direction and magnitude of current flow determines at any given instant in time the position and motion of the loudspeaker’s drivers.

It is no problem at all to send such a signal (we usually call it a "mono" signal) to multiple loudspeakers in parallel. When we do, all the loudspeakers will move in phase-locked synchrony, unlike any group of non-technological sound sources.

The effect of this is both stunning and outrageous, from a cognitive standpoint. Think about it! The cognitive "signature" of any given sound is derived from the phase-locked relationship of the direct sound and its early reflections. With an array of phase-locked loudspeakers, however, we now have a multitude of physical sound sources spread out in space that collectively have the psycho-acoustic "signature" of a single sound source. The result is that we perceive a "single" sound whose perceived location is determined by the collective relative positions of the loudspeakers. This "single" sound is what we call "the phantom image."

There is another way of thinking about this: the loudspeakers serve as the first "early reflections" of a (phantom) sound source whose direct sound we don't hear. Because our brain is good at filling in the missing blanks, it "infers" where that phantom source must be and that "inference" is what we actually perceive, or think we "hear."

This means that loudspeakers have two different behaviors. As single sources, similar to any non-technological sound source, they behave in an entirely logical and predictable way – as discrete sound sources. However, when we have an array of loudspeakers – meaning two or more speakers hooked up in a stereo or multichannel system – those loudspeakers will behave as "the first early reflections" of any recorded sound source that is sent in common to multiple members of that array.


It gets especially wacky and wonderful when we include recorded early reflections, ambience and reverberance in our signals. Now, we have information abut not only a "phantom" sound source, but also a "phantom" sound space. Our stereo or multichannel array of speakers, in combination, serves as a kind of "cognitive window" into that exotic imaginary space. It gets even better when we get the playback room interface right – then the playback room also supports the illusion of a "phantom space." Doesn’t get much better than that!

This is, of course, key to both stereophony and surround audio production, both for film/video and for music. It helps to understand how the speakers do their part in this, and to focus on the qualities of speakers that maximize this effect, to wit:

• First, the speakers need to be well matched in sound quality.

• Second, they need to have really good frequency response.

• Third, they also need to have at least reasonably good power response, especially in the horizontal plane. There are two parts to this: low frequencies depend on room absorptivity and symmetry in placement, while high frequencies depend on special strategies for gaining wide dispersion, plus room symmetry as well.

• Finally, the speakers all need to be equidistant from the listener and the listening sweetspot, as well as symmetrically distributed around the room.

This is what I’ve been working on recently, trying to design and build loudspeakers that do all these things well while also remaining affordable. Over the next several months, I'll try to explain the details of that work, to the extent that I can. What I can say for sure is that loudspeakers are getting better; I know, because I've been there and heard them!

Next month, though, we’ll take a fresh look at the quality of real broadcast audio. It ain’t pretty!

Thanks for listening.

Dave Moulton is trying to do too many things at once. You can complain to him about that or anything else at either or