1272 Chapter 34
coaxial loudspeaker (an actual Q should be used if
available). The value of N should be divided by two if
each listener can hear the direct sound from two nearby
loudspeakers and so on. Using the simplified method of
Eq. 34-6 will always provide a safe answer to Question
1 since it considers direct sound only.
34.3.3.7.2 Question 3: Can Everybody Understand?
The Alcons equation (Eq. 34-18) works well for a dis-
tributed system. For the value of N, divide the total
number of distributed loudspeakers in the room by the
number of loudspeakers producing direct sound to a lis-
tener. Thus, if each listener is in the direct field of two
loudspeakers, use a value of N equal to one-half the
total number of loudspeakers and so on. Use a value of
Q equal to the actual Q of each individual distributed
loudspeaker. A good estimate for a coaxial ceiling loud-
speaker is Q= 3. As with a central cluster system, try to
maintain a 15 dB SNR and keep distortion, hum, and so
on at a minimum for best intelligibility. In addition,
remember that the Alcons equation works best for rooms
with reverberation times of at least 1.6 s. In rooms with
a lower RT 60 , intelligibility is affected primarily by sig-
nal to noise.
34.3.3.7.3 Question 4: Will It Feed Back?
Avoid placing microphones directly under a working
loudspeaker. Provide switches to turn off loudspeakers
above microphones when microphone positions will
vary. Alternately, use an automatic mixer with logic out-
puts to automatically turn off loudspeakers above the
active microphone. For conference rooms and other sys-
tems with fixed microphone positions, use an automatic
microphone mixer with matrix output to create a
mix-minus output signal routing system that always
minimizes the signal from any microphone into a
nearby loudspeaker (see Section 34.6.5). Use the PAG
and NAG equations (Eqs. 34-19 and 34-20) with
Q=3/N (see discussion under Question 1 previously) if
the room has a significant reverberant component.
34.3.3.7.4 Question 2: Can Everybody Hear?
There are two basic patterns for laying out a distributed
ceiling loudspeaker system. They are the square and
hexagonal patterns, as shown in Fig. 34-30.
There are at least three variations of each of these
two patterns, as shown in Figs. 34-31 and 34-32. The
variations are in the spacing between the loudspeakers.
An edge-to-edge spacing places the loudspeakers so that
their coverage patterns just touch each other. A
minimum-overlap spacing overlaps the coverage of the
loudspeakers just enough to cover the dead spot in the
edge-to-edge pattern. A 50% overlap is just that; each
loudspeaker’s coverage pattern overlaps the pattern of
its neighbor by 50%. The result is that each loudspeaker
is completely overlapped by a group of its neighbors.
The choice of one of these patterns should be made
on the basis of the acoustics of the room, the ambient
noise, and the type of listeners and talkers. In a difficult
situation, such as might be encountered in a reverberant
space with significant ambient noise and some listeners
with hearing difficulties, a 50% overlap is indicated. For
business music (background music) an edge-to-edge
pattern may suffice.Figure 34-30. Square and hexagonal patterns for distrib-
uted loudspeaker systems. Courtesy Bosch/Electro-Voice.A. Square spacing.B. Hexagonal spacing.