You know the scene. You go out to a nice restaurant, wait 30 minutes for your table even though you made reservations, and when you're finally seated, realize you can't say anything intelligible to your dinner partner without literally being in their face, mouth to ear. Under these circumstances I'd recommend skipping any entrée with tomato-derived elements.
Recently, during one of these occurrences, my thoughts turned to power-the sound power creating the ruckus in the restaurant. A combination of the annoying content of the restaurant's background music system and the raised voices of my fellow diners as they struggled to converse all served to energize the highly reverberant and already noisy environment to the detriment of relaxed conversation. Of course it doesn't help that the background music is full dynamic range, so it's either mostly audible and then really loud, or mostly inaudible and then suddenly there. Either way, it's just more to add to the din.
In my soon to be reprinted treatise, "Sound Power: Who Has It, How to Get It, Why You Need It," I describe sound power as an "intrinsic property of the source," referring of course to a sound source. Sound power is the acoustic power emitted from a sound source, without considering its directivity, distance, environment, or any properties of the room.
The classic analogy in attempting to illuminate the concept is our friend the light bulb. A light bulb is rated in lumens, which describes the amount of light the bulb will emit-regardless of the environment where it resides. However, where and how you use it can produce very different results in the light levels measured or perceived. The light levels will measure differently based upon whether a fixture is used with the bulb (varies the directivity), the distance from the lamp (inverse square law), or the colors of the room (analogous to absorption and reflection). These same concepts hold for sound power as well.
In a reverberant space, the total sound power of the various sound sources will determine the overall reverberant level, which as you may have noticed, is often in competition with the level of the desired signal-your dining partner's conversation for example. This competition of signal to noise is the basis for the classic "N" factor found in the intelligibility equations. That's why to fully describe the impact of a device in a reverberant space it's not enough to only know the axial response, you should also consider the sound power of that device (and all of its fellow sound sources) as well.
Amazingly enough, that definition includes human sources. Yes, we can be (somewhat) quietly proud that we have sound power levels of our own. About 63 microwatts (0.000063 watts, based upon an SPL of 71dB @ 1meter, implicit Q of 2.5) for a slightly raised voice, not much to begin with, but hey a little here, a little there, and just like in the restaurant, pretty soon the levels start to add up.
How did we do that? Well, to begin with we usually know something about the sound pressure level of our device, even human voices. This is useful. Without going too deep, the sound power level and sound pressure level of a "standard source" (meaning one you have no hope of obtaining, but hey, let it go, read on) will be identical at a specific reference distance, power, and directivity. Those reference deals are 0.282 meter, 1 watt, and Q of 1. If we convert an SPL level, using the correction factors necessary, back to the SPL relative to those reference values, we will also have an estimate of our sound power level. In theory...which usually ends up being close enough.
Getting back to our restaurant environment, using the sound power data for the human voice, considering the number of patrons, and having some knowledge of the acoustic parameters of the room (volume, surface area, average absorption coefficient, room constant) it's possible to calculate the background noise levels simply due to people noise. And, by extension, one can also determine the amount of sound absorbing treatment that will be necessary to reduce that background noise level to something a bit more comfortable. In fact, you can probably do most of the calculations on your napkin while dining, as you can't converse anyway.