Loudspeakers & coverage

A good audio system sounds the same in every seat: loud enough to understand without straining, and never noticeably louder in one place than another. Getting there is mostly geometry — how sound falls off with distance, how wide each loudspeaker spreads it, and how far apart the speakers sit.

The inverse square law

Sound from a small source spreads out over an ever-larger sphere. Double the distance and the same energy covers four times the area, so the level drops by about 6 dB for every doubling of distance in open space.

That is the core problem with a single loud speaker at the front of a long room: the front row is blasted so the back row can hear. Indoors, reflections fill in some of the loss far from the source, but they don't help intelligibility — reflected sound is exactly what smears speech.

How loud will it be?

Speaker data sheets give a sensitivity: the level measured on axis at 1 m with 1 W of input. From there, every doubling of power adds 3 dB and every doubling of distance subtracts 6 dB.

A ceiling speaker rated 90 dB (1 W / 1 m), tapped at 4 W, gives about 96 dB at 1 m. A seated listener 1.8 m below hears about 91 dB on axis — well above the level speech needs.

Coverage angle

No loudspeaker radiates equally in every direction. Its coverage angle (or dispersion) is the cone within which the level stays within 6 dB of on axis. Outside it, the sound falls away quickly — and coverage usually narrows as frequency rises, which is exactly where the consonants that carry intelligibility live.

For a ceiling speaker, that cone hits the listening plane — seated ear height, about 1.2 m — as a circle:

A 90° speaker in a 3 m ceiling covers a circle about 3.6 m across. Raise the ceiling and each speaker covers more floor, but the listener is further away and the level drops.

Laying out ceiling speakers

Distributed systems put many small speakers overhead so that every listener is close to one. How far apart they sit is a trade-off between evenness and cost. Three spacings are common:

• Edge to edge — circles just touch. Fewest speakers; noticeable dips between them.
• Minimum overlap — circles overlap enough to cover the corners of a square grid.
• Edge to centre — each circle reaches its neighbours' centres. The most even, and usually the choice for speech-critical rooms.

Speech systems in meeting rooms

In a conference room the loudspeakers mostly carry the far end of a call and program audio. Two extra concerns apply:

• Keep speakers away from microphones. Every metre helps the echo canceller, and directional mics should have their nulls pointed at the nearest speaker.
• Don't overpower the room. Far-end voices should sound as if the person were sitting at the table — clear, but no louder than local talkers.