|
Thirteen
years after inventing the world’s first successful automatic
feedback control device, Sabine, with its patented system, retains
its technological lead.
Since microphones and amplified loudspeakers were first paired, acoustic
feedback has lain in ambush, ready to sabotage sound systems, ear
drums, listener enjoyment, and sound engineer egos with equal malice.
Until Sabine came along, the various cures and preventative measures
for feedback elimination were problematic, expensive, complicated,
fraught with compromises in sound quality, and/or just plain didn’t
work.
The ravages of runaway feedback continued unabated until 1990, when
Sabine’s breakthrough brought the world its first practical
automatic feedback controller: the FBX Feedback Exterminator. In the
years since, Sabine’s products have proven to be easy to operate,
reliable, and unrelentingly effective in providing 6-9 dB increases
in feedback-free sound system gain with no audible effect on sound
quality.
Our proprietary technology makes that claim even truer today than
it was 13 years ago. Continuous advances and improvements mean that
Sabine’s family of products delivers the purest audio and finest
protection from all varieties and levels of feedback. |
|
| By Doran Oster, President |
Ever
since Lee DeForest invented the first vacuum tube, engineers have
walked the tightrope between feedback and system gain. The purpose
of this guide is to give you the tools to get all the gain you need
without the agony of feedback. We’ll start with a common-sense
discussion of the techniques sound engineers now use to control feedback
to get the most gain and clarity out of their sound systems.
|
Our imaginary work bench
Imagine a mic and speakers set up in a tiny shower room. Clap your
hands. The sound reverberates back and forth between the tile walls
and floor. Just a touch of the volume fader fills the room with screeching
feedback. Now move our sound
system out to an open grassy field. Clap your hands. There is no echo.
The speakers are well away from the microphone and there are no reflections,
so now we can really crank up the system without a bit of feedback.
Most sound systems have characteristics that fall between these two
examples, but examining the extreme cases makes it easier to understand
the more common in-between situations.
|
Back to top
|
|
| |
|
What is acoustic feedback?
Feedback is the loud ringing sound that occurs when the sound leaving
a speaker is picked up by a microphone and re-amplified again and
again. (See Fig. 1.) The cycle repeats until the feedback reaches
the system’s maximum loudness or until someone turns down the
volume. Virtually every sound system that has a microphone and a speaker
in the same room is susceptible to feedback. |
Click images
to view
Fig.
1 |
Back to top
|
|
| |
|
Which frequencies feed back?
All acoustic systems have distinct resonant frequencies. Regardless of where you thump a guitar’s
top, it always responds with the same tone. This is the natural resonant
frequency of the guitar. It is the frequency where all of the instrument’s
components vibrate naturally as a unit. In sound systems, these resonant
points are the frequencies where feedback occurs. Each of the system’s
components, including and especially the room itself, has its own
set of resonant frequencies. Each component adds together to produce
the total system’s resonant frequencies. It is almost impossible
to predict which frequencies will feed back without first “thumping”
the system, but you only have to turn up the amp for them to rudely
reveal themselves. The frequency that feeds back first is the one
that requires the least amount of energy to excite the resonance.
If you remove the first feedback frequency, the next feedback frequency
will be the one that requires the second least amount of energy, and
so on. |
Back to top
|
|
|
|
Controlling feedback
In order for feedback to occur, the amplifier has to be turned up
enough so that sound from the speaker re-enters the microphone louder
imaginary experiment, feed-back easily occurred in the shower room
because the sound leaving the speakers did not dissipate very much
before re-entering the microphone. But when we move the speakers away
in the open field, the sound energy dissipates as it radiates away
from the speakers. If there are no surfaces to reflect the sound back
to the mic, the sound quickly loses energy, dropping to one quarter
the energy every time the distance from the speakers is doubled. By
the time the sound finally reaches the microphone, the sound energy
is weaker than the original sound, so there is no feedback.
From this example we deduce the Prime Directive of Feedback Control: Keep the
sound emanating from the speakers away from the microphones as much
as possible.
Here are the most common tricks of the trade for controlling feedback:
- Stand close to the microphone.
Speak loudly and clearly so that you do not have to amplify the
sound too much.
- Each open microphone has a chance
to feed back. Mute or turn down the gain of any microphone that
is not in use. Noise gates can be helpful for this.
- Mount the microphones in fixed
positions. Moving the microphone around on the stage increases
the chances that the microphone and the speaker will form new
resonant paths.
- Use cardioid or hyper-cardioid
microphones, and point the mics away from the speakers. They pick
up much less sound from the backside of the mic, which protects
against monitor feedback. Be careful not to put your hand on or
too close to the microphone’s screen, since this can cover
the ports that enable the heart-shaped (hence cardioid) rejection
pattern.
- Place the speakers in front of
the microphones so there is not a direct path back to the microphone.
- Aim the speakers so the sound does
not reflect directly off a wall back into the mic. You can estimate
the speaker’s dispersion pattern (the area that is directly
“sprayed” with sound) for the mids and high frequencies
by imagining rays of light radiating out of the speaker’s
horns. If you can see the center part of the horn, you are probably
in the dispersion pattern. Lower frequency sounds tend to radiate
out in all directions from all sides of the speakers.
- Make the surfaces of the room as
sound absorbent as possible to reduce sound reflections. Use acoustical
absorbing tiles in the ceiling, put down carpeting, and hang curtains.
In the real world of most performance spaces, you cannot always
follow these anti-feedback techniques. Lead singers insist on
pointing the monitors directly at the mic. Worship leaders insist
on the mobility of a wireless microphone, and nightclub owners
will not likely carpet the dance floor and hang velvet curtains.
Even after you’ve tried all these tricks, you may still not
have enough gain and clarity to satisfy the audience. Do the best
you can, and then go on to the next level of feedback control:
equalization
|
Back to top
|
|
| |
|
Equalization
Equalizers (EQs) are sets of filters, or volume controls, for different
parts of the audio spectrum. Since the earliest days, sound engineers
have used equalizers for two distinctly different purposes: 1) To
improve the tone quality and balance of the sound, and 2) To control
feedback for extra gain and microphone mobility. Some types of EQs
are best at shaping the tone and other types are better at controlling
feedback. It may seem paradoxical to add filters to a sound system
in order to increase the gain. But if you can use extremely narrow
filters to turn down the frequencies that are feeding back, you will
be able to increase the gain of all the other frequencies for a total
net gain. There are essentially three categories of equalizers: graphic,
parametric
and adaptive parametric.
|
Back to top
|
|
| |
|
Graphic EQ
Graphic EQs are basically a set of volume controls for individual
sections of the audio spectrum. The earliest music equalizers were
the bass and treble tone knobs. As technology advanced, these filters
were narrowed to give more precise control. Today, the industry standard
is called a 1/3-octave graphic equalizer, which has 31 individual
volume controls spaced 3 per octave.
There
is a common misconception in the industry about 1/3-octave EQs that
is important to this discussion. Many industry veterans incorrectly
presume that 1/3-octave EQs use 1/3-octave wide filters. If this were
the case, the EQ filters would not be wide enough to create smooth
curves. Instead, they would produce a notched frequency response that
would make the EQ use-less for shaping the sound and useless for controlling
feedback frequencies between the sliders.
Actually, most manufacturers use 3/4 to 1-octave wide overlapping
filters placed on 1/3-octave center points. These wider filters provide
the necessary smooth frequency response. (See Fig. 2.) It’s important
to understand that the term “1/3-octave” refers to the spacing
of the sliders, not the filter width.
Graphic EQs are excellent for shaping the sound, and they are fairly
simple to use. However, using one-octave wide EQ filters to control
feedback invariably causes an unnecessary decrease in the gain and
fidelity of the program. It’s easy to see that if feedback occurs
somewhere between the sliders, you will have to pull one of those
EQ sliders down pretty far to eliminate feedback. That pulls out plenty
of your program, too. On the other hand, you’ll get considerably
more net gain and
much better sound quality if you use wide graphic EQ filters for tone
control and insist on narrow filters for feedback control. (See Fig.
3.) That’s where parametric EQs come in. |
Click images
to view
Fig. 2
Fig.
3 |
Back to top
|
|
| |
|
ParametricEQ
In the quest for perfect sound, engineers developed very narrow tuned
filters for controlling feedback points in auditoriums. In the early
days of sound reinforcement, these filters were custom made to a specific
frequency and width for a specific application. Now there are a number
of commercially available parametric filter sets that allow engineers
to dial-in the width, center frequency and depth of the filter.
The problem with parametrics is that they’re expensive, they
require a good deal of expertise and auxiliary equipment to tune properly,
they require constant retuning whenever the room acoustics change,
and they are far too slow and cumbersome for catching feedback that
occurs during the program. |
Back to top
|
|
| |
|
Adaptive Parametric: The FBX®
Solution
The Sabine FBX Feedback Exterminator® is the next step
in the evolution of feedback control. The FBX® is essentially
a self-tuning parametric
EQ. It constantly monitors the program, searching for tones
that have the overtone signature of feedback. Once feedback occurs,
the FBX® automatically places a very narrow, constant-width
filter directly on the feedback frequency and lowers it just deep
enough to eliminate the ringing sound.
The FBX out performs other EQs five ways: |
1. The FBX finds and eliminates feedback automatically before and during
the program.
2. The FBX’s narrow filters eliminate feedback without losing
the fidelity of the sound.
3. The FBX is fastest. It typically finds and eliminates feedback
in less than one second.
4. The FBX gives the most gain. Use wide-filter graphic EQs for
controlling the shape of the sound and narrow FBX filters for controlling
feedback, and you’ll typically achieve a 6 to 9 dB increase
in gain compared with using the EQ alone.
5. Increase wireless mic
mobility.
|
Back to top
|
|
| |
|
What about that 6 to 9 dB increase in
gain?
Gain increase from equalization really depends on the characteristics
of the sound system and the room. Returning to our imaginary system
in the shower room, the sound bounces off the hard tile surfaces and
reflects back into the microphone with only a slight touch of the
volume slider. If you filter the first feedback point, you can only
increase the volume fader a touch more before the second feedback
occurs at a new frequency. Even if you filter six different resonance
points, you may only achieve 1 or 2 decibels of net gain because there
are so many low-energy resonant paths.
When we set our system in a large open field and the speakers are
far away from the microphone, we really have to crank it up before
we hear the first feedback. We would need an enormous system to drive
six feedback points. In this system, damping six feedback points could
easily deliver well over 15 dB net gain! |
Back to top
|
|
| |
|
How much gain do you achieve with the six FBX
filters?
Six resonance points worth — whatever that happens to be in your
unique system. You can maximize your gain by following our anti-feedback directives and by learning more about
how the FBX® filters
work best for your situation. |
Back to top
|
|
| |
|
Microphone Mobility
Mobile karaoke and wireless microphones present a special feedback
challenge. It does little good to set a number of filters for a mounted
microphone if you plan to carry the mic around the stage to different
locations. Each position on the stage has its own unique set of resonant
frequencies, so the filters that control feedback in one location
will probably not provide much help in other locations. You are faced
with a balancing act. If you insert too many filters in the system,
you will hear a degradation of the sound quality. If you set too few
filters, you will not have enough mobility or gain. In this case,
it is usually best to walk around the stage area until you find an
area where feedback is a particular problem. Then place one or two
feedback control filters to take care of that location and repeat
the process in the next few areas. FBX®
filters add less gain to mobile systems than to fixed microphone
systems, but they add a significant increase in the usable area while
preserving the natural clear sounds. |
Back to top
|
|
| |
|
Feedback Control During the Program
One of the most powerful features of the FBX is that it can eliminate
feedback during the program. FBX filters come in two types: fixed and dynamic. Both filters are placed
the same way: Feedback is detected, and the filter is placed just
deep enough to eliminate it. The difference comes after the filter
is placed. Fixed filters remain on the initially detected feedback
tone — they do not move. These filters provide the initial maximum
gain before feedback and are set automatically during setup. Dynamic
filters can release and move to new feedback frequencies and are for
adaptive feedback control during the performance.
|
Back to top
|
|
| |
|
Hearing is Believing
To hear the difference for yourself, insert an FBX in your sound system
and bypass it. Mount the mics on stands to fix their positions. Remove
as much feedback as possible using your normal method with just the
graphic EQ. Next, lower the volume, bypass the graphic EQ, and activate
the FBX. Now slowly raise the gain of the system until at least six
FBX filters have kicked in. Next, turn down the mics and play your
favorite CD through the system. Alternately listen to the system with
just the FBX and then just the graphic EQ. You will hear the FBX provides
much clearer, brighter and louder sound. If you do not have immediate
access to an FBX, run this experiment with a graphic EQ alone. You
will be amazed to hear what it does to your sound. |
Back to top
|
|
| |
|
|
|
| GLOSSARY: Definitions of "tech"
terms |
|
Decibels
We have the ability to hear an amazing range of loudness. People placed
in an absolutely quiet anechoic chamber eventually perceive the sound
of air molecules hitting their eardrums. On the other hand, people
working near jet engines hear sounds a billion times more powerful.
Engineers have developed a convention that economizes the calculations
of such an enormous range of values. This convention describes these
changes in terms of decibels (abbreviated dB) named in honor of Alexander
Graham Bell. Many non-technical people find the different uses of
the term decibels confusing because it seems to have so many different
meanings. For example, decibels are commonly used to describe the
loudness of a sound, the change in loudness (or gain) from one time
to another, for changes in signal voltage, and a number of other technical
measurements involving the power ratio of large numbers. While we
gladly leave these calculations to the engineers, it is helpful to
realize that a change of 1 dB is equivalent to a 27 percent change
in power.
With this in mind, we realize that turning up the system gain by 3
dB increases the power approximately 100% (27% x 3). In other words,
turning up the amp from 400 Watts to 800 Watts adds about 3 dB to
the system gain. Wow! Does doubling the power from 400 Watts to 800
Watts make it sound twice as loud? No! A three decibel change sounds
only slightly louder. In general, you have to increase the power about
10 times (or 10 dB) to make the sound seem twice as loud. When engineers
describe the loudness of a sound in terms of decibels, they are comparing
the sound pressure level of a particular sound com-pared to an international
standard. Fig. 4 gives several common reference points. |
Click images
to view
Fig.
4
Typical
Frequency Response |
Back to top
|
|
| |
|
Gain
Gain is a measure of the change in power (or loudness) in a sound
system. For example, turning up the amp causes an increase in gain,
while moving away from the speakers causes a decrease in gain. By
convention, gain is expressed in decibels |
Back to top
|
|
| |
|
Feedback
FEEDBACK describes what happens when a loudspeaker disperses sound
back into an amplified microphone, and at a level sufficient to allow
one or more frequencies to ring out of control. Feedback can occur
at any frequency, but is especially painful at mid to high frequencies.
The specific frequencies that feedback in a particular situation depend
on the acoustics of the environment, the placement of the microphone(s)
and speaker(s), the response characteristics of the sound system components,
and the volume of amplification. Anyone who has operated a sound system
or attended a conference or a concert is familiar with feedback and
its unpleasant consequences! |
Back to top
|
|
| |
|
Net Gain Before Feedback
Many people measure their increase in gain by the amount they push
up the mixer's calibrated slider. But if adding gain causes feedback,
you will have to cut the gain of the feedback frequency at the EQ
in order to add gain at the mixer. A more accurate concept could be
called NET gain. It is the amount of gain you achieve pushing up the
mixer slider, minus the gain you lose lowering the EQ sliders. NET
gain is the gain you realize in front of the speakers as measured
by a sound pressure level meter. That is the gain that matters.
|
Click images
to view
Net
Gain=Mixer minus EQ |
Back to top
|
|
| |
|
Frequency Response Curves
A frequency response curve is a graph that shows the gain of a component
or a group of components at different frequencies. Fig. 6 shows the
frequency response of a typical equalizer with the 1,000 Hz slider
pulled down 12 dB. The frequency response curve shows that the biggest
cut in power, called the center frequency is at 1,000 Hz, that the
filter removes half of the power (-3 dB) between 645 Hz and 1550 Hz,
the Q of the filter is 1550-645 Hz/1000 Hz (.905), and the maximum
depth is -12 dB. Fig. 6 shows the frequency response of two adjacent
sliders pulled down 12 dB. Notice that the center frequency of the
two sliders is at 885 Hz. The combined filter width is 1.49 octave
and the two filters add together to give a maximum depth of -19.3
dB. |
Fig.
6 |
Back to top
|
|
| |
|
Noise Gate/ Comb Filters
As we mentioned earlier, every microphone creates a potential source
of feedback, so it is advantageous to turn off microphones that are
not currently being used. Noise gates do this automatically by continuously
monitoring the program's loudness. If the loudness falls below a threshold
set by the user, the noise gate automatically turns off the microphone.
Once the loudness exceeds the threshold, the microphone channel automatically
turns back on.
Noise gates are useful for a number of important sound applications
besides feedback control. For example, if a person or instrument is
picked up by two microphones placed in different locations, the combined
mic signals will interfere with each other, causing a type of distortion
called comb filters. Comb filters add gain at certain frequencies
and thus increase the chance of feedback. At the same time, they cut
the gain at other frequencies, causing the program to sound thin and
over-equalized. Gating the unused microphones eliminates this source
of comb filtering. Noise gates are often employed in CD players to
eliminate noise between songs. They are similarly used in sound systems
to mute the hiss of noisy electronic components during quiet periods.
Most Sabine FBX Feedback Exterminators® feature user-programmable
noise gates. |
Back to top
|
|
| |
|
The Frequency Spectrum
People with excellent hearing can hear frequencies between 20 and
20,000 vibrations per second or Hertz. Fig. 10 shows an imaginary
120 key keyboard that would be big enough to play all the notes that
we can hear. The lowest key would play a 20 Hz "E" and the
highest key would play a 19,912 Hz "D#." Notice that doubling
the frequency raises the pitch one octave. We hear the same one-octave
musical interval between 40 and 80 Hz as we do between 10,000 and
20,000 Hertz. A graphic equalizer is superimposed that shows which
sliders affect the notes of several instruments. For example, the
chart shows that the 250 Hz slider affects most of the bottom 1/3
of a guitar's range.
The typical FBX filter below the EQ shows the relatively smaller size
and effect on sound of FBX filters and illustrates why they cause
less tonal change and gain loss. The nine FBX filters are not preset
on any particular frequencies like EQ filters. They are placed precisely
where feedback occurs. |
Click images
to view
Fig.
10 |
Back to top
|
|
| |
|
Constant-Q Filters
It is common to describe a filter's quality factor, or "Q,"
as the center frequency of the filter divided by the filter width
(in Hertz) measured at the -3 dB point. Filters that have the same
Q, or width, at the -3 dB point regardless of the filter's cut or
boost are called constant Q filters. Filters that get wider as the
filter gets deeper are called proportional Q filters. There seems
to be a new development in the audio industry. The definition of constant
Q is blurring. Many equalizer manufacturers claim their equalizers
have constant Q filters, when in fact they get substantially wider
as they get deeper. The only way to know for sure if the filters are
truly constant Q is to inspect their frequency response curves. |
Constant
Q filter
Proportional
Q filter |
Back to top
|
|
| |
|
Parametric Equalizer
A parametric equalizer allows the user to precisely specify three
critical values that determine an equalizer’s quality: the center
frequency of the EQ band that is boosted or cut (measured in Hertz),
the amount of boost or cut imposed at the center point (measured in
dB), and the width of the bell-curve shaped frequency band that is
affected (typically measured in octaves). |
Back to top
|
|
| |
|
FBX Filter
An FBX filter is essentially an automatically placed, narrowly attenuated
parametric
filter, with the center point of its narrow cut tuned to a precise
frequency that feeds back when a sound system amplifies one or more
microphones to a sufficient volume. For example, the Sabine Graphi-Q
will automatically place up to 12 FBX filters in the signal path,
corresponding to 12 distinct frequencies of feedback. |
Back to top
|
|
| |
|
Fixed FBX Filter
A fixed FBX filter will not change the frequency of the filter notch.
Once it sets itself, it remains at the same frequency. However, unless
it is LOCKED, a FIXED filter may move its notch deeper without changing
frequency. Fixed filters are typically set by turning up system gain
to the point of feedback prior to sound check or performance, and
will represent the “first layer” of feedback protection.
|
Back to top
|
|
| |
|
Dynamic FBX Filter
A Dynamic FBX Filter acts like a fixed filter, until all available
FBX filters (Fixed or Dynamic) are in use and a new frequency begins
to feedback. When this happens, whichever Dynamic filter was set earliest
in the performance will drop its original frequency and move to the
new one. Dynamic filters are especially useful with mobile or wireless
microphones (where feedback frequencies may change due to microphone
repositioning) and represent the “second layer” of feedback
protection. Note that both Fixed and Dynamic filters can be set while
music is playing, as one of the distinguishing properties of the Sabine
FBX algorithm is its ability to distinguish music (or speech, or other
sounds) from feedback. |
Back to top
|
|
| |
|
Locked FBX Filter
is a Fixed filter locked in place; i.e., it cannot get any deeper
or change its frequency. Locking filters prevents the placement of
unnecessary filters in the signal path. |
Back to top
|
|
| |
|
Filter Width
generally refers to the width (measured in octaves, or fractions thereof)
of an equalizer, including graphic EQ filters, parametric filters,
and FBX filters. More specifically, width is defined by determining
the outer frequencies (surrounding the filter center point) that are
altered ± 3 dB when the filter is imposed. This is shown in the
diagram below:
In this example, the filter width is defined as approximately one-half
octave, corresponding to the band of frequencies attenuated 3 dB or
more when the filter is pulled down. In this example, the width is
the same whether the filter depth is -9 dB or -19 dB. |
Back to top
|
|
| |
|
Setup Mode
Setup mode maximizes the Sabine FBX Feedback Exterminator®
clip level and sets all FBX parameters to a more sensitive-to-feedback
condition than normal operation. It also places a “moving limiter”
that tracks gain changes as feedback occurs, but allows the feedback
to occur at a quieter level. Feedback is created by slowly raising
the master mixer output with all the mics that might create a feedback
problem open and turned on. Please read Setup
Mode Cautions |
Back to top
|
|
| |
|
Auto Setup Mode
Automatic Setup mode is a form of Setup Mode that lets the Sabine
FBX Feedback Exterminator® automatically control the
gain and automatically “ring out” feedback frequencies.
Automatic Setup also imposes a limiter on the feedback volume, thus
allowing system setup with quiet feedback levels. Please read Setup
Mode Cautions |
Back to top
|
|
| |
|
Setup Mode Cautions
Setup Mode is designed to allow fast and quiet feedback elimination
during setup. Setup Mode should ONLY be used for pre-performance
setup. DO NOT USE SETUP MODE DURING A PERFORMANCE! This will produce
distorted audio and set filters on music or audio program. Setup Mode
also may not work well during setup in a very noisy environment. To
speed up feedback elimination, Setup relaxes its criteria for distinguishing
“good” audio from feedback and places filters more readily.
If the environment is noisy, there is a greater likelihood of placing
a filter on audio that is not feedback. Consult your Operating Guide
for details on how to operate Setup Mode on your Sabine FBX Feedback
Exterminator®. |
|
|
| |
|
| |
Back to top
|
|
| |
