|
|
Backstage Class: In Search Of The Power Alley
Don’t blame the subwoofers, look at their interaction
By
José Brusi

Subwoofer cancellation is a relatively common occurrence, particularly
in open-air system applications. Essentially, low-frequency performance
can be lacking in certain points of the coverage area.
|

Figure 1: When the distance to both sides is the same, sound from
both sides arrives at the same time. As we are closer to one side
and farther from the other, sound from one of the sides arrives
before the other.
|
 |
Sometimes blame for this problem is placed on the subwoofers themselves,
but in reality, the underlying problem is that of the interaction
between subwoofer stacks located on each side of the stage. So let’s
shed some light on this problem and remove part of the “veil of
mystery” surrounding it.
WHAT’S GOING ON?
When loudspeakers are positioned on both sides of a stage the
normal configuration for live audio events there is always interference
between them. Assuming, of course, that both sides are carrying
the same signal.
The reason for this interference is relatively simple. As seen
in Figure 1, when the distance to both sides is the same,
sound from both sides arrives at the same time. But as we move closer
to one side, and farther from the other, sound from one of the sides
arrives before the other.
This time arrival difference produces comb-filter effects. Destructive
interference occurs when two signals, one of which is delayed in
time with respect to the other, are produced from two sources (loudspeaker
groups).
|
Note Figure 2, which shows the resulting frequency response of
summing two signals with a time difference of 7 milliseconds, equivalent
to 8 feet (about 2.5 meters). The spacing between the troughs (or “combs”)
increases as frequency goes down. Absolute cancellation occurs at frequencies
where the time arrival difference results in 180 degrees of phase difference
between the two signals, so these frequencies get wiped out.
|

Figure 2: Frequency response of two signals spaced 7 milliseconds
in time.
|
 |
As any frequency response anomaly, the audibility of comb filtering
is a function of the width of the troughs. These are wider for the
lowest frequencies, which makes comb filtering more audible for
those frequencies.
Comb filtering cannot be corrected by using equalization, since
there is absolute cancellation at the bottom of the troughs.
|
Only if one of the two signals is greater in level than the other will
the effect be diminished and the combs reduced in depth.
|

Figure 3: Single subwoofer side, modeled at 100 Hz.
|
 |
WHERE ARE THE COMBS?
If the time arrival difference were constant, we would simply compensate
by delaying one side and go home happy. However, each listening
position (except the center line) gets sound arriving from each
side with a different delay between them, which means that delaying
one subwoofer stack would just shift the sweet spot elsewhere.
Also, the combs are in different frequencies at different positions,
which means that the frequency response is widely different through
out the audience area. The problem is not nearly as significant
at mid and high frequencies. First, because the troughs are narrower,
and hence less audible, and second, because the zones created are
smaller.
With bass frequencies, one can literally walk the different cancellation
areas, whereas the high frequency zones may be so small that our
left and right ears do not share zones. Thus the effects of cancellation
are not as destructive.
|
In addition, sources are more directive as frequency goes up. This means
that when the audience is closer to one of the sides, they tend to be
(at least to some extent) inside the main coverage beam of one side of
the system and outside of the other. Comb filtering effects are reduced.
|

Figure 4: Left and right subwoofers (at 100 Hz).
|
 |
But with the low-frequency energy generated by subwoofers, this
situation easily can become a nightmare.
Combs are wide and very audible, while “kick and punch” just disappears
in some coverage locations. And, because the wavelengths involved
are so large, for example, 16 feet (about 5.5 meters) at 70 Hz,
these “bad” zones can be enormous.
Subwoofers are omni-directional below, say, 150 Hz, so the entire
audience area is within coverage of both sides of subwoofers.
Let’s be very clear: ALL types of subwoofers suffer from interference
effects no matter the type (direct radiating, folded horn) or brand.
|
The result is the “power alley,” which describes the fact that only a
narrow center line of a left and right subwoofer system produces interference-free
bass.
|

Figure 5: Left and right subwoofers modeled 13 feet off-center.
|
 |
PROVING THE DEAL
Let’s use electro-acoustic modeling to reveal this phenomenon. These
predicted coverage maps are very close to what happens in “real
life” open-air applications. Closed rooms add a reverberant field,
which tends to smooth coverage out, and room modes, which create
additional zones of their own.
For our modeling, we’ve placed a three by two stack of dual-18-inch-loaded
subwoofers to each side of a 27-foot (8 meter) wide stage. Our audience
area is 200 feet (60 meters) wide by 175 feet (52 meters) long.
|
The first coverage map (Figure 3) corresponds to 100 Hz. We have
activated just one side of subwoofers, to see what happens when all bass
emanates from a single stack. Sound pressure level expands smoothly in
a way that is close to the distance-squared law. So far, so good.
But problems come a-knockin’ on our door when both sides are activated,
depicted in Figure 4, measured at a point 13 feet (4 meters) off
to one of the sides. Also at 100 Hz, we can now see some kind of alien
hand with five fingers. These are maximum pressure lobes, where the phase
difference between the left and the right sides is small.
The areas between the lobes represent cancellation. Level readings show
levels down 20 dB from the lobes. That means our 100 Hz pretty much walks
out on us.
O.K., so at least some folks are enjoying good bass, right? Wrong. Areas
where 100 Hz is in phase also means that other frequencies are out of
phase. Thus there aren’t any “good” areas, except for that ubiquitous
center line.
|

Figure 6: Left and right subwoofers modeled 43 feet off-center.
|
 |
Naturally, some frequencies are more important than others when
dealing with bass guitar or bass drum, so our problems will be more
evident in some areas than others.
|
Figure 5 shows what is taking place in the same location as Figure
4 (13 feet off to the side). At the top, we can see that the closest group
of subwoofers arrives about 2.5 milliseconds (ms) before the other. On
the bottom, we see the resulting frequency response, where the first comb
occurs at about 100 Hz.
Let’s try another location in the listening area. Figure 6 shows
data that was modeled from a position more than triple the distance
43 feet (13 meters) and again, off to one side. Note how it differs
from the previous models, and also note the comb at 40 Hz.
|

Figure 7: The power alley that results with two sub stacks spaced
47 feet apart.
|
 |
WHERE IS THAT ALLEY?
For each frequency, we have a map with fingers. Then if several
frequencies are summed and plotted on a single map, the alley emerges,
because it represents the only locations where bass sums coherently
at all frequencies.
The width of the alley depends on the spacing of the subwoofers:
as the two sides of subwoofers increase in spacing between each
other, the alley will narrow. Figure 7 shows the alley for
two single subwoofer boxes spaced 47 feet (14 meters). It is not
as obvious graphically as it is audibly, but it can still be distinguished
as an approximately 13-foot (4-meter) band with more intense color.
|
The only practical way to avoid the low-frequency power alley is to place
all subwoofers in a single location. With smaller scale applications,
locating all of them in the center is usually not possible, but larger
shows often offer this option. In addition, larger shows may also benefit
by flying the subwoofers, again at the center position.
It’s important to understand what is happening with your subwoofers. This
awareness of the zones that occur, and why they occur, can help with both
system configuration and mixing decisions.
For instance, if the front-of-house mixing console is located in the center
of the coverage area, you might chose to go with a somewhat bass-heavy
mix. Off to the sides, destructive interference will kick in and establish
some frequency balance between the low frequencies and the rest of the
system.
And after all, it’s just not fair to blame your subwoofers for something
that’s not their fault!
José (Joe) Brusi holds a degree in Engineering Acoustics and Vibration from the Institute of Sound and Vibration Research at the University of Southampton (UK). He has worked with JBL Professional and currently manages DAS Audio’s Applications and Systems Group as well as the company’s AUTOPOL measurement platform.
September 2003 Live Sound International
|