Handbook for Sound Engineers

Sound System Design 1263

34.3.2.8.2 Choosing the Cluster Location

The ideal cluster location will probably be approxi-

mately above the primary microphone location; that is,

in a rectangular room, the cluster should be near the top

and at the center of an end wall. Compromise locations

are discussed in the following sections.

34.3.2.8.3 Evaluating the Cluster Location

The potential success of a proposed cluster location and

cluster design can be evaluated by answering “The Four

Questions.” starting with Question 4: “Will it feed

back?” Using available data and Eq. 34-21, answer this

question before moving on to choose loudspeaker types.

34.3.2.8.4 Choosing the Loudspeakers

The following discussions assume the design will center
around constant-directivity packaged loudspeaker sys-
tems for a speech or speech and music reinforcement
system.
Begin the design by choosing a loudspeaker for
coverage in the rear of the room. Use trigonometry or
one of the cluster layout methods discussed in Section
34.3.2.9, and determine the horizontal (side-to-side)
coverage angle required at the rear of the room. Using
more than one loudspeaker to cover an area is discussed
in the next section. Remember that the listener’s ears
are approximately 4 ft (1.2 m) above the floor when
determining required coverage angles. (This can be
simulated in most design methods by placing an imagi-
nary floor at 4 ft [1.2 m] above the actual floor.)
Once a rear coverage loudspeaker has been chosen,
use its Q value (from the manufacturer’s specification
sheet) and the known parameters of the room to answer
Question 3: “Can everyone understand?” by calculating
the articulation loss of consonants (Alcons) from Eq.
34-18. Assume N = 1, then do the calculation over again
using N = 2 and N = 3 to simulate the effects of adding
loudspeakers to the system. If Alcons is less than or
equal to 10% for each of these calculations, the system
design will work from the criteria of Question 3. If
Alcons exceeds 10% for any of these calculations, a
second cluster or a distributed system may be required.
(Note that some designers believe an Alcons of 15% or
less is acceptable.) It might seem logical, in this case, to
simply choose a loudspeaker with a higher Q value,
since this would reduce the Alcons. That loudspeaker
will also have a narrower coverage pattern, however,
and might not adequately cover the entire audience.

Using additional high Q loudspeaker will not solve the

problem either since they increase the value of N. Thus,

if Alcons is too high, about the only alternative to a

second cluster or distributed system is to reduce the

room reverberation time with acoustic treatment.

34.3.2.8.5 Aiming the Loudspeakers

If the Alcons is acceptable, choose additional loud-

speaker to cover the rest of the room, providing a yes

answer to Question 2: “Can everybody hear?” In many

rectangular rooms, only one or two additional (wider

angle) loudspeakers will be required.

The edge of the defined coverage pattern of a loud-

speaker is its –6 dB point. For this reason, it is common

practice to overlap the coverage patterns of the loud-

speakers in a cluster to compensate for the drop-off in

coverage of each individual loudspeaker near the edges

of its coverage pattern. A good rule for this practice is

to overlap two loudspeakers as little as possible to

maintain consistent coverage throughout an audience

area. Another good rule is to aim the overlap areas at an

unimportant part of the audience area such as an aisle.

Avoid aiming loudspeakers at hard rear or side walls

(which could result in echoes) and avoid aiming them

directly down at the microphone (which could increase

the possibility of feedback). Again, remember to aim

the loudspeakers at the listener’s ears (about 4 ft

[1.2 m]above the floor for seated listeners or 5 ft

[1.5 m] for standing listeners).

Now, re-evaluate Question 4: “Will it feed back?” by

performing the calculations for PAG and NAG

discussed in Section 34.2.3.4.5. The answers to the PAG

and NAG equations may be tempered by recalculating

Dc for an increased value of N. If feedback seems

possible, consider moving the cluster, or consider using

an automatic microphone mixer to keep NOM = 1 or,

when possible, teach the users to talk closer to the

microphone (which reduces Ds), see Section 34.4.3.

Using directional loudspeakers and microphones may

provide some additional gain before feedback, but it’s

best not to plan for this additional gain.

34.3.2.8.6 Powering the Cluster

The remaining question, Question 1: “Is the system loud

enough?” can be answered by choosing loudspeakers

and power amplifiers to satisfy Eq. 34-16 remembering

the criteria for head room and SNR.