Thermodynamics, System Coverage, And The Cost Of Lunch

Lunch is not free, and energy isn’t magic.

Please Remember:

The opinions expressed are mine only. These opinions do not necessarily reflect anybody else’s opinions. I do not own, operate, manage, or represent any band, venue, or company that I talk about, unless explicitly noted.

thermodynamicsWant to use this image for something else? Great! Click it for the link to a high-res or resolution-independent version.

Before diving into this topic, I want to be very clear on a few points. First, this kind of discussion is a bit “above the pay grade” of small-venue folks like myself. Second, there’s a lot of theory involved, because I don’t have anything in the way of deep, direct, hands-on experience with it.

Ready a grain of salt to take with all this, okay?


The pro-audio world sometimes likes to behave as though thermodynamics is less of a harsh mistress than it really is. That is, there seems to be a semi-willful ignorance regarding energy and where it goes. This can lead to a sense of there being some sort of free (or reduced cost) “lunch” when it comes to the directivity of a system. The problem is that lunch is always served at full price. If you want sound to only go where you want it to go, you also have to deal with the laws governing the behavior of that audible energy.

Achieving useful, desirable directivity with an audio system was traditionally the purview of wave-guidance. In other words, horns. You channel your sonic flux through the horn, and (within the physical limits of the horn), you get certain advantages. One benefit is better “pressure transfer” to the world beyond the driver. Another nice bit of help is greater directivity. With a horn of the correct overall size and flare rate, you can focus sonic energy (within a certain passband) into a defined radiation pattern.

When horns and horn-cone hybrid boxes are used with the intention that their natural, physical directivity prevents them from interacting too much with each other, what you have is a point-source system. In such a setup, the hope is that any particular listener is overwhelmingly hearing only one source per passband…or, even better, hearing all passbands from one source. (This only has so much feasibility, especially where low-frequency material is concerned.)

As the ability to use more boxes and more electronic transformation has expanded, people are doing more and more with system processing on arrays. The enclosures involved in these arrays also have natural, physical directivity. They are also very likely to use some sort of horn for the high-frequency section. Unlike a point-source system, though, the idea is that you actually are supposed to hear the boxes interacting. This interaction can be controlled on the fly by way of changing box or driver amplitude and delay. If you want one kind of coverage, you tweak the system to interact in one way. If you suddenly decide that you want different coverage, it’s theoretically possible to simply tweak some parameters and get your change.

This is very nifty. Managing everything with actual, physical horns is a heavy, large, and predetermined sort of affair. Processing changes, in contrast, are flexible and physically lightweight. (The math, on the other hand…) “Nifty” is not “magic,” however, and this is where some people get tripped up.

The Lighting Analogy

Bear with me for a moment, as we do a foundational thought experiment.

Let’s say you have a stage light. You turn it on, and it works nicely, but you have light energy hitting something you don’t want to hit. The nice thing about your fixture is that it has shutters. You adjust the shutters so that the light no longer falls on the undesired area.

Question: Did the light falling on what you actually wanted to hit become more intense as you shuttered the beam?

No, of course not.

The visible-light radiation from the fixture hit the shutters, and was largely exchanged into heat. The luminous flux wasn’t redirected through the business-end of the fixture and mystically redirected – it was absorbed and converted. The relevant thermodynamics of the system are fully in play, and inescapable. The “cropped” energy was simply prevented from reaching a target, and that energy stopped being useful as visible light.

Now, let’s take a different approach. Let’s say you could avoid hitting an unwanted area with the light by a different means: Optics. You put a lens with tighter focus into the system, and restrict the beam-width that way.

Did the light falling on the object become more intense?

Yes, all else being equal.

The lens took the entire output of the fixture and focused that flux into a smaller area. The maximum possible fixture output remained usable.

So, what does this have to do with sound?

Focus Vs. Cancellations

In an effective sense, a horn is acoustical “lensing.” It’s a way to focus sonic energy from a driver (or drivers) into a defined space, physically giving you the directivity you want.

The flipside to this is a large, highly processed array of sound sources. Given enough drivers, enough processing, and enough time, it seems entirely feasible that a system operator could get the same coverage pattern as what would be found with point-source boxes. What has to be remembered, though, is that “lunch” has a required cost. The thermodynamics of the two approaches are not the same at all. Like our hypothetical light and tight-beam, hypothetical lens, the highly focused horn is energy efficient. A single driver (or set of drivers) have as much of their acoustical output as possible put to use solely for covering an audience.

The big, technically advanced array is energy inefficient, because it doesn’t use a physical object to focus its coverage. Instead, it requires the interaction of more energy. If you want to create an acoustical pattern through interference, you have to combine the output energies of multiple audio-output units. There are many shades of grey to take into account, of course. Even so, in the most extreme case, cancelling the output of a 1000 watt driver may require the use of another 1000 watt driver. The energy consumption of the resulting system is 2000 watts plus inefficiency losses, but your usable sonic output has not necessary doubled – remember, you’re using one driver to cancel the other for purposes of pattern control. At the physical point of that cancellation, the usable sonic energy is 0, even though the system is still consuming a large amount of electricity. It’s the same as shuttering the light. The sonic energy is merely being made unusable in a certain target area.

…and there’s a tendency to try to forget or “talk around” this. Marketing departments especially love to come up with fancy terms for things, even when those terms make no sense. Some of these highly processed systems are called impressive things like “complex point source.” The problem is that there’s no such thing. As soon as the idea for the system is to have large, intentional, audible interaction and interference across multiple units producing wideband audio, we aren’t in point-source-Kansas anymore, Toto.

There’s nothing wrong with that. Systems that have their coverage managed by way of processing and multi-box interactions are a great tool for versatility. You always bring the same gear and deploy it in basically the same way. Having exactly the right boxes for a needed point-source solution is much more possible when you’re doing a permanent, custom-built install. I’m inclined to believe the folks who claim that point-source will always measure as being more clean and coherent, but I also believe that measuring well isn’t the end-all, be-all in a discipline that has so many trade-offs.

The solutions are different, their appropriateness is situationally dependent, they are not thermodynamically equivalent, and someone is going to have to buy lunch.