1. art and entertainment
2. radio

Installing a
Broadcast
Studio
A short training manual
for low-power FM stations
Written / Compiled by
David A. Casement
Broadcast Technician
Galcom International
First Draft—August 2008
Table of Contents
Introduction
SECTION 1: POWER SOURCE
A. Proper voltage
B. Proper grounded Outlet
C. Wiring Procedures
D. Surge Protection
SECTION 2: ASSESSING THE NEED
A. Number of microphones
B. Number of machine inputs
C. Budget issues
SECTION 3: WIRING THE MIXER
A. Cables and Connectors
B. Balanced or Unbalanced
C. Inputs
D. Outputs
E. Testing the Signal through the System
SECTION 4: METERING
A. Inputs
B. Outputs
Section 5: Maintenance
Section 6: Studio Acoustics
Written by Ethan Winer
Used by permission
Section 7: Appendices
2
INTRODUCTION
n a radio station there is a long chain of equipment which the audio signals must pass through
in order to be prepared for transmission to the listener. This paper will focus on installing and
wiring an audio mixer for broadcasting and maintenance on the major pieces of studio equipment. The principals and information given should be a good reference for installing most
pieces of studio equipment which are not covered in this manual.
3
Section 1: Power Source
A. Introduction
One of the main things to establish before powering up a transmitter is that the power source is at the
proper voltage and is stable. Improper voltage, surges (spikes) and low voltage (brown outs) can
cause serious damage to a transmitter and other broadcast equipment. Also, improper wiring of the
outlet or the wiring in the building can cause damage to equipment. Before plugging in the transmitter
use a reliable voltmeter to check that the voltage at the outlet is correct and stable. Also verify that the
outlet is properly grounded. This is done by carefully placing the probes of the meter into the live and
neutral of the outlet and reading the voltage. Be sure that the voltmeter or multi-meter being used is in
good condition and operating properly. Check the test leads and probes for breaks or bare wires.
Make any repairs or replacements necessary before using the meter.
B. Testing the Outlet
It is acceptable if the above reading is within ten percent of its rating, five percent is preferable. If the
outlet is 220 volts then the reading should be between 200 and 240 volts and should not be any lower
or higher. This should not fluctuate very much while taking the reading but be relatively stable.
To verify that the outlet is grounded, place one lead in the live terminal and one in the ground/earth
and read the voltage. The reading should be the same as the one taken from live to neutral. If the
reading is zero or very low (only a few volts) then the outlet is not grounded and should not be used
until the ground terminal has been solidly connected to earth with at least a number 14 or 2mm diameter earth wire. The next reading, with the ground/earth terminal grounded, should be from earth to
neutral. The reading should be zero or just a few millivolts. The earth and neutral should both be connected to ground but the neutral should go back through the transformer which feeds power to the
building before it is connected to ground.
The connection to the transformer is done by the electricity supply company. If the outlet and electrical
system in the building are wired properly the reading between these two terminals should read zero or
in the millivolt range. If the reading is high, such as several volts or higher, the outlet is not wired correctly and the transmitter or other equipment connected to this outlet will not function properly and can
even be damaged. This condition will defeat and can even cause surge protectors to be ineffective or
even be permanently damaged! In extreme cases some protection devices can even EXPLODE!
Surge protection power bars have devices from live to earth and neutral to earth and depend on the
outlet being wired properly so the voltages on earth and neutral are in proper range. If either the earth
or neutral are “floating” too high then the limits of the protective devices can be greatly exceeded. If
any wiring beyond the outlet itself is required, call a qualified electrician to repair the wiring.
If the voltage in your location fluctuates or has frequent outages, a voltage stabilizer or an Uninterruptible Power Supply (UPS) will be needed for a studio or low power transmitter such as 30 to 200 watts.
A generator will be needed for higher power transmitters.
Make sure that the outlets used for your station equipment are wired properly and are supplying the
proper voltages BEFORE plugging in any equipment! Many times basic items such as lights and portable fans will appear to operate on power sources which are less than ideal. This does NOT mean
that the power source is correct. The above test must be done!
4
TYPE M
(used almost exclusively in South Africa, Swaziland and Lesotho)
Earth
Neutral
Live
This plug resembles the Indian type D plug, but its pins are much larger. Type M is rated at 15 amps.
Although type D is standard in India, Sri Lanka, Nepal and Namibia, type M is also used for larger appliances. Some sockets over there can take both type M and type D plugs.
C. Plugs and connectors
Electrical equipment often comes from the factory with molded ends on the power cords which plug
into the electrical outlets. Some manufacturers provide two or three cords with different ends or plugs
for various types of outlets which are used in different parts of the world. If the power cord(s) which
are provided with your transmitter are not the correct ones for the outlets in your country the correct
one must be installed properly.
Modifying electrical plugs to fit into outlets or removing earth pins is an unsafe practice and must not
be done. The earth pin is there for safety reasons. Removing it can cause severe electrical shock to
personel and can also cause problems with the functioning of the transmitter. The practice of not using
a plug on the end of an electrical cord and just pushing the bare ends of the wires into an outlet is also
very unsafe. This practice can lead to electrical shock and the connection will be unreliable at best.
Wires by themselves in the outlets will not make proper contact with the inside of each terminal which
can lead to arcing in the outlet and burn the wires off. It is imperative that the proper plug be used for
the outlet. If a different plug is needed on the power cord provided it is best to purchase a new power
cord at an electrical shop. The second option is to cut off the plug on the power cord and properly wire
on a new one. Before doing this make sure the colors for live, neutral and earths are known on both
the cord and the plug. (Normally in Africa on power cords, brown is live; blue is neutral and green/
yellow is earth.) If there is any uncertainty have an electrician wire the new plug on the cord.
5
D. Surge protection
This is a very important part of a radio station. Surge protectors are just what the name says; they protect the equipment from voltage surges or “spikes”. It is a good practice to use multiple outlet power
bars which have surge protection built in for studio use. With adequate power ratings these can be
used for low power transmitter sites as well. The power handling rating must be well above the total
power consumption of the transmitter plus any associated equipment. That is, the total of all the equipment plugged into the power bar.
Along with this kind of surge protection it is also very important that there
be a higher capacity surge protector in the system. There are specific units
which are designed to be wired into the circuit breaker panel. Other units
are plugged into an outlet and are in series with the equipment they protect
but are larger physically and will handle greater power surges than what is
built into power bars. Power bars tend to respond to surges faster than the
larger units. It is best to have both kinds of surge protection in the system.
All of these systems rely on the electrical wiring to be correct and a good
connection to earth! Without a proper connection to earth they will not provide adequate protection. Again, make sure the wire or cable used to connect to earth is a larger diameter than the other wires in the system.
Superior Electric makes the PT series of surge suppressors which wire
parallel to the load. It can be wired in the breaker panel or even on a plug
and
inserted into an outlet.
Tripp Lite is on manufacturer that makes different models of Power
Bars with surge suppression built in such as this one.
6
Section 2: Assessing the Need
A. Microphones
It is important to carefully consider how many microphone channels will be needed in the on-air mixer
(audio console). Thinking too big for the size of station you have may cost more money than your
budget can afford. If you do not have enough microphone channels it can severely limit the abilities of
the station and affect the on-air sound if some voices are not close enough to a microphone. Generally a basic on-air studio should have at least three microphones. One for the main presenter, one for
a co-presenter and an extra channel for a third presenter should it ever be needed. The third channel
will also serve as a back up if either of the other two channels should stop working.
Be sure to match the characteristic impedance of the microphones to the input of the mixer.
Generally the better the quality of microphone used the better the sound will be on-air. Cheap cone
shaped microphones used in sound reinforcement systems will work but will not give good clear
sound. Sony's SM 58 is extremely popular for sound reinforcement and very durable. It has become
very popular for radio stations on small budgets as well. Although it is popular affordable and gives
reasonable sound, it is not the best sounding microphone for a radio station. One to try is the Audio
OM2. It is said to have far superior fidelity and barrel noise reduction, and it's less expensive.
If your microphones are of any value at all, they will be balanced (XLR jacks--three leads). Most microphones are called dynamic; these do not require a power source. Some microphones are called condenser and require a power source fed through two of the leads from mixers that offer "phantom
power". In general condenser microphones are great for recording subtle or distant sounds, neither of
which are desired in a radio studio.
.
RE 20
PL 20
ATM 25
Shure SM7b
7
AT 2020
The other thing about radio studios is that we more interested in eliminating noise and sounding good
than in high sound fidelity. You sound good when far off sounds from fan vibrations and table tapping
are not audible are not picked up. This is accomplished by using a low-sensitivity mic; and using
shock mounts. This will also help reduce popping and sibilance. Proper mic placement helps with this
as well. Large diaphragm mics tend to produce the full audio spectrum.
Now on to some other popular favourites. The Electro-Voice RE20 microphone is widely popular in
larger budget stations. It is a large diaphragm dynamic microphone with built in noise canceling. The
RE20 retailed for over $800 in Canada, it is discontinued now and has been replaced by the RE30
and RE40. You can get one slightly used on eBay for US$325-$400. A better deal is its lesser-known
cousin the PL20; its guts and specs are identical but it was marketed to the US musician market rather
than radio and it has a slightly different finish: US$275-325 on eBay. In the same quality league is the
Shure SM7 (replaced by the SM7a and later the SM7b with superior hum-bucking in the presence of
fluorescent lights and computer screens). It's about the same price as the PL20. Electro-Voice makes
another microphone for stations with bigger budgets: the RE27 has three frequency cut-off switches,
greater sensitivity than the RE20, and superior hum bucking and pop filtering. One theory is that many
good mics designed to be kick-drum mics would also serve well in capturing voices for radio; the relatively cheap Audio Technica ATM25 looks like a good candidate. At the time of writing the AT 2020
and AT2030 are also being investigated. Canadian prices about $125 and $200.
B. Machine inputs
The same principal regarding the number of mic inputs also applies to machine inputs. It is best to
count up all the sources of audio that will be used and add two or three more channels to that. It is
common for a new station to install an audio console for on-air use and it be filled or nearly filled when
it is installed. This is not a good situation as there is no room for expansion and no spare channels if
C. Budget issues
The budget for starting a station is always a major factor in selecting equipment. Next to tower, transmitter and studio to transmitter link system one of the most expensive items is the audio mixer or console which will be used on-air. There are two types of audio mixers which can be considered. One is
the public address type which is used in churches and public buildings. The other is the actual broadcast console type of mixer. The public address type mixer has more controls for changing the sound
and routing sound to different outputs. This type of mixer varies greatly in quality depending on manufacturer and price.
Audio mixers which are specifically built for broadcasting tend to have fewer tone controls and outputs
than the public address units but are built and labeled specifically for broadcasting. The components
used are generally of much higher quality and they are built much sturdier and will last many years.
However they also cost a great deal more money than public address mixers. Some brands cost US
$6000 for just six channels. With this in mind there are two radically divergent philosophies regarding
mixers. Some insists that the big names in broadcaster mixers; Ward Beck, McCurdy, WheatStone
etc. consoles can't be beat, and that the small PA (public address) style mixers allow announcers to
fiddle with too many controls and they just don't look like a proper radio studio mixer.
8
Increasingly, new radio engineers are favouring Mackie-style PA mixers. They are inexpensive (very
much less expensive than a big broadcast board), light, small, and they generate little heat. They are
quiet in two respects: their audio specs are amazing and they don't require cooling fans.
Behringer makes slightly less expensive alternatives to the Mackie boards. However Galcom's experience with Behringer mixers is that they need very good surge protection ahead of the power supply.
They tend to be susceptible to surges damaging the power supplies and the mixer will then stop operating. Obtaining a new supply from the company is not always easy.
Another good brand of public address mixer is Soundcraft. They cost a bit more than the Mackie or
Behringer but are better built than the Behringer and do have some features which are desirable for a
radio station. Both Mackie and Soundcraft come with built in filtering against radio signals and other
electrical noise which may be introduced on the power or mic lines.
When considering the purchase of a mixer for a radio station it may be possible to purchase a used
one and save some money. Caution must be exercised when purchasing used equipment. Used public address type mixers should only be considered if they are in nearly new condition or if the mixer is
one that is know to someone at the radio station and is known to be in excellent condition.
If a used broadcast type console is available have it checked to make sure that it is fully functional and
nothing functional is damaged. These units can often be refurbished by a qualified technician and give
several years of good use. If it is not possible to try a used mixer for a time or have a reliable technician check it out then it is better to buy a new one.
Soundcraft M8
Arrakis 12005S
Audio Arts Air 1
9
Section 3: Wiring the Mixer
A. Cables and Connectors
Without them you would just have a pile of electronic equipment doing nothing! A connector can also
be called a ‘plug’. The receptacle that a plug plugs into is called a ‘jack’ or ‘socket’. If a connector is
regularly plugged and unplugged, eliminate problems in advance by choosing a sturdy one. Good
ones cost more, but by not compromising on cables and connectors, you will eliminate a critical variable when it comes to troubleshooting problems.
1) 1/4 inch Phone Connector - There are two types of
1/4" phone plugs; TS and TRS.
Two-conductor phone plug - has a tip and a sleeve,
thereby also called a TS plug. The tip carries the actual
signal and the sleeve acts as a ground, or return. This
is an unbalanced connection and is mainly used for
short-distance runs between high-impedance equipment. The major disadvantage of an unbalanced line is
that it’s more susceptible to picking up noise.
Three-Conductor phone plug - has a tip, a ring, and a
sleeve, thereby also called a TRS plug. There are three
common ways to use this type of connector; First, as a
balanced connection which uses two of the conductors
(tip and ring) carrying the same signal, while the sleeve
acts as a shield. This provides far better noise rejection
than an unbalanced connector.
Mini plugs and jacks are the same configuration as 1/4 inch just smaller in size.
2) RCA/Phono Connector -is the standard for consumer stereo components such as CD players, cassette decks and more. Their advantage is
that they’re extremely inexpensive and can be grouped together in small
areas. The disadvantages of RCA/Phono plugs are that they’re an unbalanced connection, and tend to be very fragile. You will usually find
they are needed for Tape In and Tape Out sockets because of their
common use in consumer grade CD and cassette players.
3)XLR Connectors - (sometimes called Cannon Connectors) are barrel-shaped plugs most often associated with microphones. Most XLR connectors are balanced and have three wires connected to three
separate pins. These connectors are almost always used
for low-impedance, balanced connections such as microphones. The advantage is that they are extremely sturdy
and reliable. However, they are more expensive than other
types. It will be wired so that Pin 1 is ground, Pin 2 is
‘hot’ (positive), and Pin 3 is ‘cold’ (negative). Occasionally
you will encounter audio equipment that require a different
configuration of hot/cold/ground. Make sure to consult the
owner’s manual if the standard XLR configuration doesn’t
work.
10
Note both male and female connectors in the back of this amplifier. When installing XLR connectors
make sure of which gender is needed for the equipment involved.
Standard wiring of an XLR type connector.
4) Adapters - are available for just about every possible combination of connectors including;
- XLR to TRS Used in three wire connections between equipment. Some equipment uses TRS for
balanced lines as well as XLR.
- RCA to TS TS ¼ inch are used in both professional and consumer grade equipment.
• mini-phone to 1/4" TRS phone Mini plugs are generally used on consumer grade equipment
or on mini disc players and Walkmans.
5) Audio cable
i. coaxial - This type of cable looks similar for both unbalanced microphone and radio signal. However, their electrical characteristics
are very different. They both consist of a center conductor (wire)
surrounded by insulating material, then an outer conductor which is
of screen-like construction. This is all covered by a protective
jacket. The size of conductors and the insulative material between
them determines the characteristics of these cables making them
suitable for either audio or small signal radio frequencies.
ii. Two-conductor shielded - For balanced microphones and many other
connections between audio equipment, two-conductor shielded cable is
used. There are different sizes and qualities. These are not coaxial
type cables. Better quality cables have a foil surrounding the two inner
wires, then a strip of bare wire laying on the foil to form 100% shield.
This again is contained inside a protective jacket. The two inner wires
are insulated from each other and the shield and t he insulation on
them is colour coded.
11
6) Soldering
Extreme caution should be used when soldering audio cables and connectors in every part of the radio station. A high quality low wattage (25-40 watt) soldering iron should be used. An iron that is too
hot will melt insulation on wires causing short circuits. It can also melt insulation inside connectors
causing them to be unusable. A good quality, high density sponge, slightly moist, should be used for
cleaning the tip of the iron. Never use acid-core or automotive type solder in audio or radio station
equipment. The best solder to use is 60/30/10 tin/lead/silver resin core. If this is not available, 60/40 is
second choice.
Note: Be sure BOTH ends of the cable are disconnected from equipment before preparing for soldering or soldering. The act of stripping insulation to get ready to solder can produce short circuits which
could result in damage to the equipment. Soldering a cable that is connected at the other end can
cause heat damage or a 'ground loop' through the soldering iron, resulting to damage to the equipment.
B. Balanced or Unbalanced
There are two basic types of connections for audio signals to audio and radio equipment (other than
optical); balanced and unbalanced. Unbalanced is generally intended to be used for shorter cables, 6
metres or less, where it is not in close proximity to strong interfering signals such as high power radio,
wireless communication, or high power electrical motors. Balanced connections are intended for situations where a much greater rejection of interference is crucial. They also allow much longer lines to be
run between pieces of equipment without degrading signals.
Unbalanced cables and connectors only use two
conductors. One is the so-called 'hot' (live) wire. The
other is called 'return' (ground). Typically, the cable
running between the pieces of equipment is a coaxial style. The center conductor is the hot lead,
which is surrounded by insulating material, which is
then surrounded by a second conductor which is the
return. The return conductor is most often formed
into a screen-like fashion in order to shield outside
electrical signals away from the hot lead.
Balanced cables and connectors use three conductors. Two of these are hot and the third is return. The
two hot wires carry the desired signal between
pieces of equipment. The third conductor does not
carry any of the desired signal but only acts as a
shield to keep out external interfering signals. Any
interfering signal which does get past the shield will
then be present on both of the hot wires at the same
time. Because these wires are connected to a transformer which cancels out anything that is not desired, interfering signals are not heard.
12
Step 8: If the music is not heard, turn the power off and recheck all connections between the mixer
and audio sources. Check that the wiring on the outputs from the mixer are correct. Begin at step 1
again.
Step 9: Before going on-air, if at all possible, send a 1000 Hz tone through the mixer. Set the output to
make the cue bus meter read 0. The manual for the mixer should tell what the level is in dbm when
the meter reads 0.
Step 10: Make sure the transmitter is switched off. Then switch to program and do the same as step 8.
Step 11: Adjust the master output control on the audio mixer until the correct nominal input level is
given to the transmitter.
Caution: Before switching on any radio transmitter, make sure the antenna and feed line are not damaged and are connected properly. (See How to Wire a Transmitter Manual compiled by Galcom International in partnership with Africa by Radio.)
Section 4: Metering
Broadcast mixers often have both input and output metering. Whereas sound reinforcement mixers
rarely have input metering and often have a series of LEDs used as an output level meter.
A. Inputs
Broadcast mixers which have analogue or meters with moving needles will also have a red light to indicate clipping (distortion – signal coming in is too strong). These meters work well but do need annual
calibration. Sound reinforcement mixers, while not having input meters, do have trim controls
(sensitivity control) plus a red LED which does indicate clipping on each channel. The trim control
should be adjusted for the audio source on that channel so that the average sound input gives 0vu
with the red LED never coming on. Broadcast mixers have switch settings or jumpers which determine
the level of input signal for each channel.
B. Outputs
Both types of audio mixer have level meters, as previously mentioned. Both the analogue type of meter and the row of LEDs work well to give an indication of the signal leaving the mixer. The LED style
meters tend to respond more quickly and do not require maintenance.
1. Program Bus Meter – Measures the audio level leaving the mixer going to the transmitter (or
to the transmitter of the Studio-to-Transmitter Link system.) When this meter reading is averaging 0vu,
the system should be set up to be giving the transmitter the correct input level. For example, +4dbm.
2. Audition Bus Meter – This bus is used when there is a need for more constant monitoring
(such as listening to a receiver on the frequency of the station). It can also be used for recording programs at the same time as a broadcast is going on-air. Hence, the Audition Bus Meter is indicating the
level of this material and not what is going out over the air. Its calibration, however, is the same as the
Program Bus Meter.
3. Cue Bus Meter – The function of Cue Bus is described in Section D-9. The meter indicates
only what is played through the Cue Bus. Again, its levels and calibration are the same as the Program and Audition meters.
One disadvantage of sound reinforcement mixers is that they do not have the flexibility in metering
that a broadcast mixer has. It is quite easy to set up the functions of program, audition, and cue bus,
but in most cases the only actual metering that is built-in is the output meter.
13
C. Inputs
1) Types – microphone or machine/line
On an analog broadcast type mixer there are two types of inputs for sound sources. Microphone inputs which tend to be mono and low impedance and high sensitivity. Line level inputs tend to be stereo and higher input impedance. Line level inputs generally can not be used for microphones as the
sensitivity is too low. When ordering these type of mixers balanced or unbalanced inputs have to be
specified for the line level inputs.
On analog sound reinforcement mixers tend to have mono channels which can be used for microphone inputs or mono line level inputs. The “trim” adjustments are used to set the sensitivity for the
given channel.
2) Impedance high or low/- levels
Broadcast mixers tend to have jumper wires or switch settings to select the input impedance for a
given channel as well as the input level. On line level channels the input signal level has to be set for
the type of equipment being connected. Consumer equipment has an nominal output of -10dbm, professional equipment has a nominal output of +4dbm.
3) Types of connection
On broadcast consoles the connections tend to be terminal strips which the cables from the audio
sources are wired in rather than ¼ inch or XLR connectors. Some also have special connectors which
the cables are wired into that are then plugged into the mixer.
Sound reinforcement mixers tend to use XLR and ¼ inch connectors for inputs.
14
D. Outputs
1) balanced or unbalanced
Outputs from both broadcast and sound reinforcement mixers tend to have both balanced and unbalanced outputs. The main outputs tend to be balanced and the sub-channels or monitor channels tend
to be unbalanced. The main channels in a radio station should be fed to the transmitter and the monitor channels should be used for cue and monitor speakers.
Professional transmitters will require balanced lines.
2) Impedance
Generally both types of mixers will have low impedance outputs. Most often the output is nominally
600 ohms. It is best to check with manual for the given transmitter to see if it has 600 ohm input. Most
often the transmitter will have a higher impedance than 600 ohm but is compatible with a 600 ohm
output from the mixer. It is best to check with the manual of the mixer to see if it requires a 600 ohm
termination. Some transmitters will have this as an option. Having this termination set properly will
give the correct audio levels into the transmitter.
3) Levels
Again both types of mixers tend to have a setting or a switch for the output level to select between 10dbm and +4dbm. Most professional transmitters require a nominal +4dbm input level. Some sound
reinforcement mixers do not have this option and are set at -10dbm.
4) Types of connection
Most broadcast mixers use wire terminal connections for the output lines the same way as the input
lines are wired. The balanced lines will have three terminals for the hot, return and shield. Some
newer smaller lower priced broadcast mixers are using three pin 1/4inch connectors. These tend to be
cumbersome for permanent connections and can be prone to having wires break when they are left
hanging from the 1/4inch connectors.
Most sound reinforcement mixers will generally use XLR connectors for the main outputs. Some will
use 1/4inch.
5) Number
In general broadcast mixers will have their outputs labeled for the function they perform. Some even
have small amplifiers and speakers built in which can be used for cue and monitoring. They will also
have one or two of these secondary outputs available for control room monitor speakers or even recording room monitoring speakers.
6) multi-purpose
Sound reinforcement mixers tend to have more outputs than broadcast mixers. They have monitor
outputs and effects outputs. These are all line level and can be used for what they are intended but
are really multi purpose and can be used for the cue bus and monitoring just the same as a broadcast
console. It is even possible to use one of these output “buses” to feed into a voice activated relay
(such as manufactured by Radio Design Labs) which can bring on an on-air light. This would be set on
only the microphone channels. This copies the “tally” relay built into broadcast mixers which costs
more money.
7) On-air
A convenient feature of the broadcast mixer is the on/off button on each channel. This allows the operator to leave the fader set to a given level and toggle the audio on and off if it is needed; such as the
presenter's microphone being set to optimum level but needing to be switched off while someone else
is speaking into another microphone or while music is playing. Sound reinforcement mixers have a
mute button on each channel which essentially performs the same function; a feature which is not
lacking in the less expensive.
15
8) Recording
The multiple outputs of both the broadcast and sound reinforcement mixers mentioned above can be
useful for recording programming while simultaneously broadcasting when this is desired. Broadcast
consoles with fewer actual output lines (buses) may in some cases need to feed the recording equipment through a distribution amplifier. This would be necessary when the output is already being sent
to other equipment, such as monitor speakers. Without a distribution amplifier it is possible that this
signal may not be strong enough to give quality sound to more than one system at the same time.
Sound reinforcement, as already mentioned, tend to have more output lines. This makes it very convenient to record programs at the same time they are being broadcast. Any one of the line level outputs, such as Aux send, Monitor send, or Effects send, can be dedicated for this task. In general, a
distribution amplifier would not be needed in this case.
9) Cue bus
Cue bus is used to prepare the next item for broadcast, whether music or speaking program, while
something else is on the air. The cue channel on the mixer does not affect the on-air channel but
sends the audio to a separate output which can be headphones or speakers installed only in the control room. The operator can then cue the next media (CD, cassette, computer) to the beginning of the
song or program which is to go on the air next. Any material played on the 'cue bus' does not go out
over the air. Once material is cued, the switch for monitoring would then be put back to the on-air signal.
Most broadcast mixers have a built-in cue bus and the monitor switch has a position labeled for it.
Some have a small amplifier and speakers built in. Other smaller mixers require separate amplifier
and speakers or powered speakers to hear what is played on cue bus.
Sound reinforcement mixers do not have a label 'cue bus' or a built-in amplifier and speakers. But, as
mentioned above, any one of the line level outputs labeled 'aux1, aux2, effects send, monitor send'
can be used for this function. Whichever is selected can be relabeled as 'cue bus'.
. Testing the Signal through the System
Step 1: When all of the inputs and outputs to the mixer have been wired correctly and it is time to play
audio through the system, go through the following basic steps to ensure the system is working properly:
Step 2: Set channel slider (level control) to half volume. Never turn a level control to Full for an initial
test or for troubleshooting. This could lead to sudden bursts of very high volume sound which can
damage microphones and speakers and hurt the human ear.
Step 3: Turn headphone or control room monitor speakers to approximately ¼ volume.
Step 4: Route the signal from the channel being used to the cue bus.
Step 5: Give the mixer audio input via a computer or CD player. It is preferable to use music for a
more constant signal for this test.
Step 6: The music should now be heard through the speakers connected to the cue bus at a low to
comfortable volume. If the sound is not distorted and is constant, the system is working properly. Adjust speaker volume for comfortable listening.
Step 7: Adjust channel slider until a proper reading is seen on the cue bus meter (average -6vu to
0vu) with no red light blinking showing that the signal is peaking (distorting) and going above +6vu..
16
Analog Peak Reading Meter
Analogue Volume Units Meter—Classic Style
17
Section 5: Maintenance
18
Section 6: Studio Acoustics
Build a Better Bass Trap
Tame the sonic gremlins that lurk in your mixing room
with acoustic treatment you can build yourself.
By Ethan Winer (Used with permission)
(This article first appeared in the June 1995 issue of Electronic Musician magazine.)
NOTE: I still gladly answer questions about these trap plans and offer advice, however I prefer
to answer in my Acoustics forum at EQ Magazine rather than by email. That way the effort I put
into answering can help others, and you can benefit from the answers of others too. Be sure to
also see the Acoustics FAQ which explains bass traps and acoustic treatment in much more
detail than this brief magazine article.
Please understand that the bass traps described in this article are intended mainly for larger
rooms. For rooms smaller than 25 by 15 feet you'll usually do better with broadband bass traps
made from thick rigid fiberglass placed in the room corners. This is described in my Acoustics
FAQ linked above. Once all of the corners are treated, then it makes sense to consider adding
bass traps like these flat on the walls.
When I built my first recording studio in the late 1960s there was no such thing as semi-pro audio
equipment. In those days, if you wanted a four-track tape recorder you parted with big bucks for an
Ampex or a Scully. A decent limiter cost nearly $1,000 and a good spring reverb unit would set you
back about $2,000. We can all be grateful that very high quality audio gear is now available for even
the most modest home studio budget. But there is still one important feature that sets apart professional studios from most basement or garage operations: proper acoustic treatment. When handled
correctly, acoustic treatment will make the difference between a mixdown room that sounds muddy
and one that sounds clear and tight. If the playback sounds muddy, of course, you have to work much
harder to create a good mix.
This article will explain what acoustic treatment is all about, why you need it, and perhaps most important - especially if you have more ambition than cash - how to construct it yourself. Treating a
mixdown room properly is not a trivial project, but it doesn't have to be expensive either.
A. ACOUSTICS 101
There are two very different and unrelated aspects of acoustic treatment used in recording studios:
One is sound isolation, which attempts to minimize the leakage between rooms and also between a
room and the outside. The other is acoustic treatment within a room, to minimize reflections that cause
reverb, echoes, and standing waves. It is the treatment within audio mixing rooms that will be addressed here.
f you walk into an empty room and clap your hands, you'll hear a series of closely spaced echoes. Often these echoes also possess a discernable musical pitch, called ringing, especially if the room is
small. Echoes and ringing are caused by sound striking the walls, and then bouncing back and forth
between the opposite walls. Besides the obvious intrusion of echoes in a room designed for playing
and mixing music, the ringing also causes certain frequencies to be emphasized. The time between
the echoes and which frequencies are emphasized depend on the room's shape and dimensions.
19
To avoid these problems, professional mixing rooms are designed to eliminate most reflections. Deadening the room helps you to hear any reverb and other effects being added to a mix, without being influenced by natural ambience within the room. It also kills the ringing along with the echoes, thereby
minimizing the need for 1/3-octave equalizers. (See the sidebar Fine Tuning the Control Room.) But
proper acoustic treatment involves more than just eliminating the audible echoes and ringing, which
impact only the midrange and upper frequencies. Unless your recording is limited to voice-overs and
narration, it is just as important to eliminate the reflections that occur at low frequencies.
Many home-studio owners install commercial acoustic foam on their control room walls, mistakenly
believing that is sufficient. After all, if you clap your hands in a room treated with foam (or fiberglass or
heavy blankets), you won't hear any echoes or ringing. But these products do nothing to control low
frequency reflections, and hand claps won't reveal that. Basement studios with walls made of brick or
concrete are especially prone to this problem - the more rigid the walls, the more they reflect low frequency energy. Indeed, simply building a new sheet rock wall a few inches inside an outer cement
wall can help to reduce low frequency reflections. The wall vibrates, thus absorbing some of the sound
energy instead of reflecting it all back into the room. But this alone is inadequate for a serious mixing
room, and you'll get much better results using resonating boxes designed specifically to absorb low
frequency energy. These boxes are called bass traps, and they absorb the lowest frequencies where
fiberglass and foam stop working. The bass traps I have found most effective are built from plywood
panels, and designed to vibrate over a broad range of bass frequencies. Fiberglass is mounted behind
the panels to damp the vibration, thus absorbing the bass energy from the room.
When bass frequencies bounce around in a room they generate standing waves. Standing waves are
pressure nodes created when a sound wave reflected from a wall collides with the direct sound emanating from the loudspeaker. At some frequencies the reflections reinforce the direct sound, creating
an increase in level at that location in the room. And at other frequencies the reflections tend to cancel
the direct sound, lowering the volume or in some cases eliminating it altogether. (Standing waves can
be reduced with non-parallel walls and an angled ceiling, but such construction is too costly for most
home studios.) The variation in bass response caused by standing waves is perhaps the single biggest obstacle to mixdown satisfaction for home-studio owners. You create what you think is a terrific
sounding mix in your studio, only to get complaints that it sounds either boomy or thin everywhere
else.
Standing waves can also occur at midrange frequencies, but they are less intrusive there because
most musical material does not contain sustained single notes as much as in the bass region. Further,
midrange wavelengths are short enough that moving your head even a few inches will bring back a
canceled tone. However, it is possible for a sustained note on a flute, French horn, or clarinet to create
a standing wave. For this reason, sine waves are never used when measuring the frequency response
of monitor speakers in a mixing room. Instead, pink noise is played through the loudspeakers because
no single frequency is present in pink noise long enough for a standing wave to develop.
20
B. THE COST OF PERFECTION
Although acoustic foam products are useful for absorbing midrange and high frequencies, they are
relatively expensive: Sculpted foam two inches thick costs about five times more than type #703 oneinch rigid fiberglass board which is just as effective. (Rigid fiberglass is similar to the fluffy type used
for home insulation, but it is much denser. A sheet of #703 one inch thick is equal in sound absorption
to a much thicker batt of regular fiberglass.) Likewise, pre-built commercial bass traps are readily
available, but they too cost many times more than the raw materials needed to build your own.
The plans provided here use sheets of rigid fiberglass one inch thick, covered with fabric for a better
appearance, to absorb the mid and high frequencies. These are complemented with two types of bass
traps made from plywood and one-inch rigid fiberglass to handle the low frequencies. One type of trap
uses 1/4-inch plywood to absorb the deepest bass frequencies; the other is built with 1/8-inch plywood
and handles the upper bass range. The rigid fiberglass is made by Owens-Corning, and can be purchased in boxes of two by four foot panels from a commercial insulation supplier. Nearly any porous
fabric can be used on the mid-high frequency absorbers to cover the fiberglass and make it more attractive. I chose an off-white dyed burlap because it is inexpensive and acoustically transparent, yet it
also looks good.
21
There are many different types of bass trap designs, and this one is as effective as any. It is particularly well-suited for small rooms because the traps protrude very little - the thickest unit is only four
inches deep. A two-foot width was chosen because it yields two traps from each sheet of plywood with
no waste. These traps are equally useful in large recording areas, perhaps mounted along only one
wall in a room that is to remain at least partially live sounding. You could also line one wall of a large
recording room with bass-only traps, to tame the bass but not affect the ambience in the upper frequencies at all.
The bass traps and fiberglass-only absorbers are built to a size of two by eight feet, and they are
mounted vertically on the walls of your control room in alternating order. That is, first is a deep bass
trap, then a mid/high fiberglass-only absorber, then a high bass trap, a mid/high absorber, a deep
bass trap, and so forth around the room. Because bass tends to build up in the corners of a room, try
to arrange the order of the traps such that the bass units are in the corners and the mid/high units then
alternate with the remaining traps along each wall. It is perfectly acceptable to have two bass traps
adjacent in a corner, but it's best if one is a high bass absorber and the other a deep bass unit. You
may also change the size of a mid-high absorber if necessary, to accommodate the room dimensions.
The center ceiling absorber shown in the photos was made slightly narrower than the others because
the full two feet was not available.
If your walls are less than eight feet high, you can shorten the traps accordingly. And in a room with
high ceilings you can raise the traps slightly so they are more centered vertically on the walls. It is not
necessary to line all four room walls, but you should cover at least three of them, optionally omitting
the front wall behind the monitor speakers.
Although this article focuses on adding bass traps to mixing rooms, you will notice in the photos of my
studio that they're in the mixdown area of a single room. Like many home-studio owners, I record
acoustic instruments as well as electric guitars and synthesizers. Since I mostly record myself, the recording and mixing areas are combined in one large room. This way I don't have to keep going between rooms to record and then hear the result. But in this case there is no rear wall in the mixing
area, so additional traps are mounted on the angled ceiling to maximize the absorption. There are also
four bass-only traps in the rear corners of the room that are not visible in the photos.
C. CONSTRUCTION DETAILS
Each of the three trap designs is based on a wooden frame which is first built and then mounted to the
wall using toggle bolts. The accompanying plans and parts lists provide most of the detail you will
need to build these traps, but there are some additional points to be aware of.
Select good quality #2 pine boards that are not warped, avoiding any with excessive or open knot
holes. Inexpensive fir is adequate for the 2x2s on the deep bass traps and mid/high absorbers since
these boards don't show and also because they are not part of the sealed system. I recommend "AC"
plywood for the bass trap fronts, which means that one side (the "A" side) is smooth and will look good
when painted. The good side is of course placed facing the outside. Any decent lumberyard will gladly
cut the plywood sheets in half lengthwise for you, though some may charge a small fee for the service.
The deep bass traps and mid/high absorbers use 1x4 and 1x3 frames, to which you attach the inner
2x2 boards that support the fiberglass. Be sure to use screws instead of nails throughout, for maximum strength and stability. The high bass traps use a single frame made of 2x2s only, which should
be joined in the corners using screws.
Note (added January 29, 2006): Finished wood is a little smaller than the stated size. For example, a
1x4 is closer to 5/8 by 3-5/8 inch. So the deep-bass traps will really be 3-5/8 inch deep, not 4 inches.
22
Attach the rigid fiberglass panels using one-inch diameter fender washers to keep the screw heads
from pulling through the fiberglass. For the high bass traps the fiberglass can be glued directly to the
wall if you prefer. Note that you will have to trim the fiberglass slightly to fit within the frames.
For the mid/high absorbers the burlap is stretched only slightly and attached to a second, removable
frame using short (two to three inch) strips of adhesive-backed Velcro, which are also stapled into the
wood for extra strength. The removable frames have a cross-bar in the middle to prevent them from
collapsing when the burlap is pulled snug. It is not necessary to pull the burlap too tightly, and if it ever
loosens you can mist it lightly with water in an old Windex squeeze bottle to make it shrink.
Finally, if the traps are being installed before the floor covering, be sure to leave a sufficient gap at the
bottom of the boxes before attaching them to the wall. I used a scrap strip of sheet rock to rest each
frame on when drilling the mounting holes and attaching the toggle bolts.
AL FINE
As you can see from the parts lists, the raw materials for these traps is relatively inexpensive. With the
help of a friend it took about a week to build the 21 traps and absorbers in my studio. Most home studios will probably need fewer traps. Also, it took us a little longer than it might have otherwise because
seven of the traps are mounted on the ceiling.
I have built three commercial studios using this trap design, and I can attest that it works very well. In
each case these traps tamed a room that was unusable for serious mixing into a room that sounds
good and is a pleasure to work in.
23
My personal philosophy is to avoid 1/3 octave equalizers if possible, because they can introduce as
many peaks and valleys in the response as they remove. However, I am not totally opposed to corrective equalization. The measured response of my JBL 4430 speakers was acceptably flat, except for a
4 dB. rise centered at 400 Hz. For a single bump like this, a parametric equalizer set to the appropriate frequency and bandwidth is ideal. But I was able to avoid adding yet another device to the signal
path by altering the existing line-out EQ circuit on my Rane MP24 mixer. By simply adding two capacitors, I changed the midrange center frequency from its original 1200 Hz. to the 400 Hz. needed to correct my speakers.
24
The plan drawing views are what you'd see looking down into each box, so they don't show the end
caps which seal and complete the boxes. However, the accompanying photos show how the finished
traps should look. Note that the length of the end caps in the parts list is slightly overstated, and the
true length depends on the actual thickness of the various boards. The correct length is two feet minus
twice the thickness of each 1x3 or 1x4 or 2x2, and I recommend that you follow the old carpenter's
adage: "Measure twice, cut once."
The assembled frames are mounted to the wall using five toggle bolts per side, evenly spaced. While
one person holds the frame against the wall, another drills 1/4-inch holes through the 2x2s and into
the wall. Then remove the frame and enlarge the 1/4-inch holes in the wall to 1/2 inch or 3/4 inch, or
whatever is needed to accommodate the toggle bolts. But please be careful not to drill into any existing electrical wiring! Once the final holes have been drilled you can bolt the frame to the wall and continue with the next frame.
It is imperative that the bass traps be sealed using a liberal coating of caulk at all joints and where the
wood frame joins to the wall. Caulk or construction glue is also needed between the frame and the plywood front panel. Apply the caulk that seals the frame to the wall before attaching the fiberglass, and
also work it in thoroughly with your fingers. Any seams in the wallboard behind the traps should also
be caulked. Apply the caulk to the front edges of the bass traps just before attaching the plywood to
the frame. I used one-inch finishing nails spaced about two inches apart to attach the plywood, but
you can use fewer nails if the plywood is glued instead of caulked. Be sure to use a high-quality silicon
25
Note (added October 7, 1999): Although it is not stated in the article, the low-bass trap is effective at
frequencies between about 80 Hz. and 160 Hz., and the high-bass trap absorbs frequencies between
150 Hz. and 300 Hz.
Note (added April 15, 2001): I uploaded a small (1.3 MB) Wave file of pink noise that you can use to
measure a room's frequency response. The file is only 15 seconds in length, but it will play seamlessly
in loop mode in Sound Forge or another audio editor. Right-click here and then select Save Target As
to download the file.
Note (added September 23, 2001): Ben Fury was kind enough to write this update with current parts
prices and a few tips.
SIDEBAR: Fine Tuning the Control Room
Note (added January 1, 2005): In the ten years since this article was written I have learned that 1/3
octave EQ is not appropriate for correcting room problems. See THIS updated article on the RealTraps web site for a more modern approach to room tuning.
Many recording studios use 1/3 octave graphic equalizers to flatten the frequency response of the
monitor speakers. These equalizers are inserted into the signal path between the mixer and power
amplifiers, to counteract the inevitable bumps and dips in any loudspeaker/room combination. Adjusting the frequency response of a room with equalizers is called "tuning the room."
When a manufacturer publishes response curves for a speaker, the measurements were made in an
anechoic chamber - a room that is completely dead at every audible frequency. Eliminating all reflections ensures the measured response is accurate. But like a car maker's inflated mileage claims,
measuring a loudspeaker's response in an anechoic chamber does not reflect reality. What really matters is the frequency response in your room.
There are different philosophies about the best way to tune a control room, and no one method is correct. What you do - if you do anything at all - depends on your mindset, the size of your wallet, and
perhaps the kind of music you produce. Many people are satisfied to adjust the tweeter level on the
speakers if one is provided, and accept the results. If the speakers are biamped, the relative level between low and high frequencies can be further adjusted with the controls on the electronic crossover.
Acoustic treatment as described in this article goes a long way toward eliminating response-skewing
reflections, and with a properly treated room equalization may not be worth the effort and expense.
Further, all monitor speakers have a "sound," and it's not wrong to pick a speaker that sounds the way
you like and simply leave it at that.
Another philosophy is to aim for a perfectly flat response at any cost. Once the speakers have been
made as flat as possible by adjusting the tweeter controls and crossover, 1/3 octave equalizers are
added to the signal path. If, for example, a 3 dB. dip is measured at 1 KHz., the equalizers are set to
boost that frequency by 3 dB. to compensate. In practice the left and right equalizers are usually set
independently, since each speaker and its location in the room may require a different correction.
To properly measure the frequency response of speakers in a room requires a pink noise signal
source, a spectrum analyzer (or sweepable filter that passes only 1/3 octave band at a time), a calibrated microphone, and a voltmeter with a decibel readout. All-in-one units are available that combine
these components into a single package, making the measurements fairly easy to perform. There are
also software programs available that work with your PC's sound card to generate pink noise and
measure the room response. The microphone doesn't have to be perfectly flat, as long as you know
what its response really is and incorporate the deviation into your measurements.
26
My personal philosophy is to avoid 1/3 octave equalizers if possible, because they can introduce as
many peaks and valleys in the response as they remove. However, I am not totally opposed to corrective equalization. The measured response of my JBL 4430 speakers was acceptably flat, except for a
4 dB. rise centered at 400 Hz. For a single bump like this, a parametric equalizer set to the appropriate frequency and bandwidth is ideal. But I was able to avoid adding yet another device to the signal
path by altering the existing line-out EQ circuit on my Rane MP24 mixer. By simply adding two capacitors, I changed the midrange center frequency from its original 1200 Hz. to the 400 Hz. needed to correct my speakers.
Radio studios
Radio studios are very similar to recording studios, particularly in the case of production studios which
are not normally used on-air. This type of studio would normally have all of the same equipment that
any other audio recording studio would have, particularly if it is at a large station, or at a combined facility that houses a station group.
Broadcast studios also use many of the same principles such as sound isolation, with adaptations
suited to the live on-air nature of their use. Such equipment would commonly include a telephone hybrid for putting telephone calls on the air, a POTS codec for receiving remote broadcasts, a dead air
alarm for detecting unexpected silence, and a broadcast delay for dropping anything from coughs to
profanity. In the U.S., FCC-licensed stations also must have an Emergency Alert System decoder
(typically in the studio), and in the case of full-power stations, an encoder that can interrupt programming on all channels which a station transmits in order to broadcast urgent warnings.
Computers are also used for playing ads, jingles, bumpers, soundbites, phone calls, sound effects,
traffic and weather reports, and now full broadcast automation when nobody is around. For talk
shows, a producer and/or assistant in a control room runs the show, including screening calls and entering the callers' names and subject into a queue, which the show's host can see and make a proper
introduction with. Radio contest winners can also be edited on the fly and put on the air within a minute or two after they have been recorded accepting their prize.
Additionally, digital mixing consoles can be interconnected via audio over Ethernet, or split into two
parts, with inputs and outputs wired to a rackmount audio engine, and one or more control surfaces
(mixing boards) and/or computers connected via serial port, allowing the producer or the talent to control the show from either point. With Ethernet and audio over IP (live) or FTP (recorded), this also allows remote access, so that DJs can do shows from a home studio via ISDN or Internet. Additional
outside audio connections are required for the studio/transmitter link for over-the-air stations, satellite
dishes for sending and receiving shows, and for webcasting or podcasting.
This page was last modified on 25 August 2008, at 22:48.
All text is available under the terms of the GNU Free Documentation License. (See Copyrights for
details.)
27
Acoustic Treatment and Design for Recording Studios & Listening Rooms
INTRODUCTION
by Ethan Winer
I've been very pleased to see the current growing interest in acoustic treatment. Even as recently as
five years ago, it was rare to read a magazine article or newsgroup posting
TABLE OF CONTENTS
about acoustics, bass traps, diffusors, room modes, and so forth. Today
such discussions are common. And well they should be - the acoustics of a
Introduction
recording or listening room are arguably more important than almost anyPart 1 - Acoustic Treatment
thing else!
Diffusors and Absorbers
These days, all gear is acceptably flat over the most important parts of the
audio range. Distortion, aside from loudspeakers and microphones, is low
enough to be inconsequential. And noise - a big problem with analog tape
recorders - is now pretty much irrelevant with modern digital recording. Indeed, given the current high quality of even semi-pro audio gear, the real
issue these days is your skill as a recording engineer and the quality of the
rooms in which you record and make mixing decisions. Top
What's the point in buying a microphone preamp that is ruler flat from DC to
microwaves when the acoustics in your control room create peaks and dips
as large as 20 dB throughout the entire bass range? How important really
are jitter artifacts 110 dB below the music when standing waves in your studio cause a huge hole at 80 Hz exactly where you placed a mike for the
acoustic bass? Clearly, frequency response errors of this magnitude are an
enormous problem, yet most studios and control rooms suffer from this defect. Worse, many studio owners have no idea their rooms have such a
skewed response! Without knowing what your music really sounds like, it is
difficult to produce a quality product, and even more difficult to create mixes
that sound the same outside your control room.
Midrange and High Frequency
Absorbers
Rigid Fiberglass
Bass Traps Overview
Fiberglass Bass Traps
Optimizing the Air Gap
Better Bass Traps
Part 2 - Room Design and
Layout
Room Sizes and Shapes
Room Symmetry
Live or Dead?
Noise Control
More Resources
Part 3 - Sidebar articles
Sidebar - Standing Waves
Sidebar - Fine Tuning the
Control Room
Sidebar - Measuring Absorption
Sidebar - The Numbers Game
Sidebar - Big Waves, Small
Rooms
Sidebar - Hard Floor, Soft
Ceiling
Sidebar - Room Modes and
ModeCalc
Sidebar - Creating
Revisions
This article explains the basic principles of acoustic treatment. Some of the
material is taken from my bass traps plans published in Electronic Musician
magazine, some is from my company's web site, and some is from my postings in audio newsgroups. However the vast majority is new content that
does not appear anywhere else. I have consolidated this information here to
provide a single comprehensive source that is free of commercial references. My goal is to offer advice that is complete and accurate, yet easy to
understand using common sense explanations instead of math and formulas. Although many books
about recording studio and listening room acoustics are available, most of the better ones are too
technical for the average audio enthusiast to understand without effort. And you'd need to purchase
and read many books to learn a few relevant items from each.
Please understand that I am not a degreed acoustician or studio architect. However, I am very experienced in acoustic treatment, and particularly in bass traps. The recommendations described here are
the result of my own personal experience and should not be taken as the final word. Although this article is intended mainly for audio engineers and home recordists, all of the information applies equally
to home theaters, small churches and auditoriums, and other rooms where high quality reproduction of
audio and music is required.
This text will surely expand as I learn more. And as people ask me questions or request elaboration, I
will incorporate the answers and additions here. Also, there is a growing list of newer acoustics articles on the Articles page of my company's web site you can browse through. If you have questions or
comments about anything related to acoustics, please do not send me email. Rather, I prefer that you
post publicly in my EQ Magazine Acoustics forum. This way the effort I put into answering can help
others, and you can benefit from the answers of others too.
28
PART 1: ACOUSTIC TREATMENT
There are four primary goals of acoustic treatment: 1) To prevent standing waves and acoustic interference from affecting the frequency response of recording studios and listening rooms; 2) to reduce
modal ringing in small rooms and lower the reverb time in larger studios, churches, and auditoriums;
3) to absorb or diffuse sound in the room to avoid ringing and flutter echoes, and improve stereo imaging; and 4) to keep sound from leaking into or out of a room. That is, to prevent your music from disturbing the neighbors, and to keep the sound of passing trucks from getting into your microphones.
Please understand that acoustic treatment as described here is designed to control the sound quality
within a room. It is not intended to prevent sound propagation between rooms. Sound transmission
and leakage are reduced via construction - using thick massive walls, and isolating the building structures - generally by floating the walls and floors, and hanging the ceilings with shock mounts. Sound
isolation issues are beyond the scope of this article. For learning more about isolation and the types of
construction needed I recommend Home Recording Studio: Build it Like the Pros by Rod Gervais.
Proper acoustic treatment can transform a muddy sounding room, having poor midrange definition
and erratic bass response, into one that sounds clear and tight, and is a pleasure to work and listen in.
Without effective acoustic treatment, it is difficult to hear what you're doing, making you work much
harder to create a good mix. In a home theater, poor acoustics can make the sound less clear, harder
to localize, and with an uneven frequency response. Even if you spent many thousands of dollars on
the most accurate loudspeakers and other equipment available, the frequency response you actually
realize in an untreated room is likely to vary by 30 dB or even more.
There are two basic types of acoustic treatment - absorbers and diffusors. There are also two types of
absorbers. One type controls midrange and high frequency reflections; the other, a bass trap, is
mainly for low frequencies. All three types of treatment are usually required before a room is suitable
for making mixing decisions and for serious listening.
Many studio owners and audiophiles install acoustic foam all over their walls, mistakenly believing that
is sufficient. After all, if you clap your hands in a room treated with foam (or fiberglass, blankets, or
egg crates), you won't hear any reverb or echoes. But thin treatments do nothing to control low frequency reverb or reflections, and hand claps won't reveal that. Basement studios and living rooms
having walls made of brick or concrete are especially prone to this problem - the more rigid the walls,
the more reflective they are at low frequencies. Indeed, simply building a new sheet rock wall a few
inches inside an outer cement wall helps to reduce reflections at the lowest frequencies because a
sheet rock wall that flexes also absorbs a little.
You may ask why you need acoustic treatment at all, since few people listening to your music will be
in a room that is acoustically treated. The reason is simple: All rooms sound differently, both in their
amount of liveness and their frequency response. If you create a mix that sounds good in your room,
which has its own particular frequency response, it is likely to sound very different in other rooms. For
example, if your room has a severe lack of deep bass, your mixes will probably contain too much bass
as you incorrectly compensate based on what you are hearing. And if someone else plays your music
in a room that has too much deep bass, the error will be exaggerated, and they will hear way too
much deep bass. Therefore, the only practical solution is to make your room as accurate as possible
so any variation others experience is due solely to the response of their room.
29
DIFFUSORS AND ABSORBERS
Diffusors are used to reduce or eliminate repetitive echoes that occur in rooms having parallel walls and a flat
ceiling. Although there are different philosophies about
how much natural reverberation recording studios and
listening rooms should have, all professional studio designers agree that periodic reflections caused by parallel
walls are best avoided. Therefore, diffusion is often used
in addition to absorption to tame these reflections. Such
treatment is universally accepted as better than making
the room completely dead by covering all of the walls
with absorbent material. For me, the ideal listening room
has a mix of reflective and absorptive surfaces, with no
one large area all live or all dead sounding. Understand
that "live" and "dead" as described here concern only the
mid and upper frequencies. Low frequency treatment is
another matter entirely, and will be described separately.
The simplest type of diffusor is one or more sheets of
plywood attached to a wall at a slight angle, to prevent
sound from bouncing repeatedly between the same two
walls. Alternatively, the plywood can be bent into a
curved shape, though that is more difficult to install. In
truth, this is really a deflector, not a diffuser, as described
Real diffusor designs use an irregular surface having a complex pattern
to scatter the sound waves even more thoroughly. Yet another type,
shown below, uses chambers having different depths. Note that for diffusion to be effective, you need to treat more than just a few small areas. When walls are parallel, adding diffusion to only a small percentage of the surface area will not reduce objectionable echoes nearly as
well as treating one or both walls more completely.
Again, the angled and curved walls described earlier are deflectors, not
diffusors. A true diffusor scatters sound waves in different directions
based on their frequency, rather than merely redirecting all waves in
the same direction. This is an important distinction because a flat surface that is angled or curved still fosters the boxy sounding response
peaks and dips known as comb filtering. A real diffusor avoids direct
reflections altogether, and thus has a much more open, transparent,
and natural sound than a simple flat or curved surface. Besides sounding less colored than an angled or curved wall in a control room, diffusors serve another useful purpose in recording rooms: they can reduce
leakage between instruments being recorded at the same time. Where
an angled wall simply deflects a sound - possibly toward a microphone
meant to pick up another instrument - a diffusor scatters the sound
over a much wider range. So whatever arrives at the wrong microphone is greatly reduced in level because only a small part of the original sound arrived there. The rest was scattered to other parts of the
room.
30
Unfortunately the better commercial diffusors are not cheap. So what are some alternatives for the
rest of us? Aside from the skyline type diffusors, which are sometimes made of plastic and sold for not
too much money, you can make a wall either totally dead or partially dead. For someone with a very
small budget, making the rear wall of a control room totally dead may be the only solution. At least that
gets rid of flutter echoes between the front and rear wall, though at the expense of sounding stuffy and
unnatural. But it's better than the hollow boxy sound you get from a plain flat reflective surface. Another option is to make the rear wall of a control room partially reflective and partially absorbent. You
can do this by making the wall totally dead, and then covering it with thin vertical strips of wood to reflect some of the sound back into the room. If you vary the spacing from strip to strip a little, you'll reduce the coherence of the reflections a little which further improves the sound.
Fast repetitive echoes - also called flutter echoes - can color the sound in the room and cause an emphasis at frequencies whose wavelengths correspond to the distance between the walls, and between
the floor and ceiling. Flutter echoes are often identified as a "boing" sound that has a specific pitch. If
you clap your hands in a live room or an empty stairwell or tunnel, you can easily hear the tone. If the
room is large, you'll probably notice more of a rapid echo rat-a-tat-tat effect - the "flutter." Smaller
rooms resonate at higher frequencies, so there you are more likely to hear a specific tone that continues even after the original sound has stopped. This effect is called ringing. Besides the obvious ill effects caused by the echoes, ringing creates an unpleasant sonic signature that can permeate recordings made in that room and negatively affect the sound of everything played through loudspeakers in that room.
Note that echo, flutter echo, and ringing are intimately related, so the delay time and pitch always depends on the distances between opposing surfaces. With small spacings the flutter echo's pitch is directly related to the distance. I have a long stairwell in my home with a spacing of 36.5 inches between walls. When I clap my hands loudly I hear a distinct tone at the F# whose pitch is about 186 Hz,
and the half wavelength for 186 Hz is 36.5 inches. But with larger distances you may hear a higher
frequency than the spacing would indicate, depending on what sound source excites the echoes. For
example, when you clap your hands or otherwise excite a room with only midrange frequencies, the
only resonances that can respond are also at mid/high frequencies. So if the distance between parallel
walls fosters a resonance at, say, 50 Hz, you might hear 200 Hz, or 350 Hz, when you clap your
hands.
Like diffusion, midrange and high frequency absorption helps minimize echoes and ringing. But unlike
diffusion, absorption also reduces a room's reverb time. This makes the sound clearer and lets you
hear better what is in the recording by minimizing the room's contribution. For example, if you make
mixing decisions in a room that is too reverberant, you will probably add too little reverb electronically
because what you hear includes the room's inherent reverb. Likewise, if the room is overly bright
sounding due to insufficient absorption, your mixes will tend to sound muffled when played on other
systems because the treble adjustments you make will be incorrect. Therefore, diffusion is used to
avoid flutter echo, ringing, and comb filtering, but without reducing the room's natural ambience.
Low frequency absorbers - bass traps - can be used to reduce the low frequency reverb time in a
large space, but they are more commonly used in recording studios and listening rooms to reduce modal ringing and flatten the frequency response in the bass range. This is especially true in smaller
rooms where a poor low frequency response is the main problem. In fact, small rooms don't really
have reverb at all at low frequencies. Rather, ringing at the room's individual mode frequencies dominates. But in large recording studios, churches, and auditoriums, reducing low frequency reverb is an
important reason for adding bass traps.
31
MIDRANGE AND HIGH FREQUENCY ABSORBERS
Without question, the most effective absorber for midrange and high frequencies is rigid fiberglass.
Owens-Corning 703 and 705, or equivalents from other manufacturers, are the standard absorbing
materials used by professional studio designers. Besides being extremely absorbent they are also fireproof and, when applied to a wall, can even retard the spread of heat. Rigid fiberglass is available in
panels 2 by 4 feet and in thicknesses ranging from 1 to 4 inches. Larger sizes are available, but 2 by 4
is more convenient for most studio applications, and can be shipped more economically. As with all
absorbent materials, the thicker it is, the lower in frequency it will absorb to. That is, 703 fiberglass one
inch thick absorbs reasonably well down to 500 Hz. When two inches thick, the same material is
equally absorbent down to 250 Hz. See the sidebar Measuring Absorption for more information about
how these measurements are made.
For a given thickness, 703 is about twice as absorbent as acoustic foam at the lower frequencies, and
it generally costs much less. Even better for low frequencies is 705-FRK, which is much more absorbent than 703 at 125 Hz and below. FRK stands for Foil Reinforced Kraft paper. This is similar to the
paper that grocery bags are made of, but with a thin layer of metal foil bonded to one side. The FRK
paper was not intended for acoustic purposes, but to serve as a vapor barrier in homes. It just happens to be good acoustically too. Be aware that the paper reflects mid and high frequencies when installed with that side facing the room; this may or may not be desirable for a given application. 705 is
also available without a paper backing.
Although 703 and 705 fiberglass panels are more effective than foam of the same thickness, they are
usually covered with fabric for appearance, and to prevent the glass fibers from escaping into the air.
This adds to the expense and difficulty of building and installing them. (In practice, fiberglass particles
are not likely to escape into the air unless the material is disturbed.) A comparison of 703, 705-FRK
with the reflective paper exposed, and typical foam is shown in Table 1 below. Note that foam panels
sold as acoustic treatment are often sculpted for appearance, and to better absorb sound arriving at
an angle. Removing some of the material reduces foam's effectiveness at low frequencies. If rigid fiberglass was compared to solid foam panels of the same thickness, the disparity in low frequency performance would likely be less. However, not having a sculpted surface would then reduce foam's absorption at higher frequencies.
It's not difficult to understand why 705 fiberglass is so much more absorbent than typical sculpted
foam at low frequencies. Besides the fact that sculpted foam has about half the mass of solid foam
due to material being removed to create the irregular surface, another consideration is density. According to test data published by several manufacturers of rigid fiberglass and rock wool, the denser
types absorb more at low frequencies. The data published by Johns-Manville for their line of rigid fiberglass shown below is one example. Acoustic foam has a density of less than 2 pounds per cubic
foot (pcf) compared to 705 fiberglass which has a density of 6 pcf.
My own tests in a certified acoustics lab confirm this, showing denser types of rigid fiberglass absorb
as much as 40 percent more than less dense types at 125 Hz and below. More recently I performed
THIS series of measurements in my company's test lab, which shows the relationship between density
and low frequency performance even more conclusively. Regardless of the reason, there is no disputing that for a given panel size and thickness, 705-FRK is substantially more effective at low frequencies than the same thickness of typical acoustic foam. However, it is important to understand that a
material's density is but one contributor to its effectiveness as an absorber. Obviously, if the density is
made too high the material will reflect more than it absorbs, so it's a mistake to conclude that higher
densities are always better. For this reason, test data must be the final arbiter of a product's effectiveness.
32
As the data above clearly shows, 6 pcf rigid fiberglass panels absorb substantially more at low frequencies than less dense 3 pcf material.
One important way to improve the low frequency performance of any absorbent material - besides
making it thicker - is to space it away from the wall or ceiling. For a given material thickness, increasing the depth of the air gap lowers the frequency range it absorbs. For example, 703 that is two inches
thick and mounted directly against a wall has an absorption coefficient of 0.17 at 125 Hz. Spacing the
same material 16 inches away from the wall increases that to 0.40 - a nearly three-fold improvement.
Of course, few people are willing to give up that much space in their rooms! And even very thick (four
inch) 705-FRK with a one-foot gap will not absorb the lowest frequencies as well as a purpose-built
bass trap which is optimized for that purpose. Bass traps, absorption coefficients, and spacing of absorbent material will be described separately and in more detail later.
IS "RIGID FIBERGLASS" AN OXYMORON?
There is some confusion about the term "rigid fiberglass" because it is
not really rigid like a piece of wood or hard plastic. Rather, the term
rigid is used to differentiate products such as 703 from the fluffy fiberglass commonly used for home insulation. Rigid fiberglass is made of
the same material as regular fiberglass, but it is woven and compressed to reduce its size and increase its density. Rigid fiberglass that
is one inch thick contains about the same amount of raw material as 3
to 6 inches of regular fiberglass. The photo below shows a piece of 703
one inch thick folded slightly. As you can see, it is rigid enough that it
doesn't flop over when not supported (right side of photo), but not so
rigid that it can't be bent or squeezed.
33
Now that you know what rigid fiberglass is, where the heck do you buy it? You probably won't find it at
your local hardware store or lumber yard, but many insulation suppliers stock it or can order it. Start by
looking in your telephone directory under Insulation and also Heating / Air Conditioning Suppliers. You
can find the name of an Owens-Corning dealer near you by calling 800-GET-PINK (800-438-7465) or
from the Locator page on the Owens-Corning web site. Other companies, such as Knauf, Armstrong,
and Delta, make similar products, and they often cost less than fiberglass from Owens-Corning. You
can contact them directly to find a distributor near you. In the interest of completeness, here are some
other manufacturers that make similar products: Johns-Manville, CertainTeed, Roxul, Ottawa Fibre,
and Fibrex.
When assessing rigid fiberglass, it is important to know its density so you can compare equivalent
products. Owens-Corning 703 has a density of about three pounds per cubic foot (45 kilograms per
cubic meter), and 705 is about six pounds per cubic foot (90 kilograms per cubic meter). Therefore,
products from other companies that have a similar density will have similar absorption characteristics
at the same frequencies. Note that some companies call their products mineral wool, mineral fiber, or
rock wool, but acoustically they are equivalent to fiberglass.
Rigid fiberglass is great stuff, and you can cut it fairly easily with a razor knife, but it's not very pleasant to work with because the fibers can make your skin itch. While handling it you should wear work
gloves, and you won't be too cautious if you also wear a dust mask. The usual way to mount rigid fiberglass to a wall is with sheet rock screws and large diameter washers with a small hole, often called
fender washers. These washers are needed to prevent the screw heads from pulling through the fiberglass. Fender washers are available at Home Depot and other hardware stores. If your wall is made of
cement or brick, you can instead use construction glue like Liquid Nails to attach small strips of wood
to the wall, and then screw the fiberglass to the strips. Since fiberglass works better when spaced
away from a wall or ceiling, wood strips make sense even when you are able to screw directly into the
wall.
Once the fiberglass is attached to the wall, you can build a wooden frame covered with fabric and
place the frame over the fiberglass for appearance. If that's too much work, you can cut pieces of fabric and staple them to the edges of the wood strips. Nearly any porous fabric is appropriate, and one
popular brand is Guilford type FR701. Unfortunately, it's very expensive. One key feature of FR701 is
that it's made of polyester so it won't shrink or loosen with changes in humidity when stretched on a
frame. But polyester is a common material available in many styles and patterns at any local fabric
store. Another feature of FR701 is that it's one of the few commercial fabrics rated to be acoustically
transparent. But since you're not using it as speaker grill cloth to place in front of a tweeter, that feature too is not necessary.
Shiny fabrics having a tight weave should be avoided because they reflect higher frequencies. The
standard test for acoustic fabric is to hold it to your mouth and try to blow air through it. If you can blow
through it easily, it will pass sound into the fiberglass. Burlap and Muslin are two inexpensive options,
but nearly any soft fabric will work and also keep the glass fibers safely in place.
BASS TRAPS - OVERVIEW
The most common application of bass traps in recording studios and control rooms is to minimize
standing waves and acoustic interference which skew the room's low frequency response. (See the
sidebar, Why They're Called Standing Waves.) As you can see in Figure 1 below, acoustic interference occurs inside a room when sound waves bounce off the floor, walls, and ceiling, and collide with
each other and with waves still coming from the loudspeaker or other sound source. Left untreated,
this creates severe peaks and dips in the frequency response that change as you move around in the
room. At the listening position, there might be near-total cancellation centered at, say, 100 Hz, while in
the back of the room, 100 Hz is boosted by 2 dB but 70 Hz is partially canceled.
34
Figure 1: Acoustic interference causes direct and reflected waves to combine in the air, creating peaks and dips in the frequency response.
Here, a positive wave front from the loudspeaker (left) is reflected off the rear wall on the right, and the
reflection collides with other waves that continue to emanate from the loudspeaker. Depending on the
room dimensions and the wavelength (frequency) of the tones, the air pressure of the reflected waves
either adds to or subtracts from the pressure of the waves still coming from the speaker. Worse, different locations in the room respond differently, with a boost at some frequencies and a reduction at others. When waves combine in phase and reinforce each other, the increase in level can be as much as
6 dB. But when they combine destructively, the dip in response can be much more severe. Level reductions of 25 dB or more are typical in untreated rooms, and near-total cancellation at some frequencies and locations is not uncommon. Further, most rooms have many peaks and dips throughout the
entire bass range, not just at one or two frequencies. Figure 2 below shows the frequency response of
the 10- by 16-foot untreated control room at a friend's studio. Note the large number of ripples, and
their magnitude, all within just one octave!
Figure 2: Your worst nightmare? Yes, this response really is typical for an untreated small room!
35
Here, a positive wave front from the loudspeaker (left) is reflected off the rear wall on the right, and the
reflection collides with other waves that continue to emanate from the loudspeaker. Depending on the
room dimensions and the wavelength (frequency) of the tones, the air pressure of the reflected waves
either adds to or subtracts from the pressure of the waves still coming from the speaker. Worse, different locations in the room respond differently, with a boost at some frequencies and a reduction at others. When waves combine in phase and reinforce each other, the increase in level can be as much as
6 dB. But when they combine destructively, the dip in response can be much more severe. Level reductions of 25 dB or more are typical in untreated rooms, and near-total cancellation at some frequencies and locations is not uncommon. Further, most rooms have many peaks and dips throughout the
entire bass range, not just at one or two frequencies. Figure 2 below shows the frequency response of
the 10- by 16-foot untreated control room at a friend's studio. Note the large number of ripples, and
their magnitude, all within just one octave!
Figure 2: Your worst nightmare? Yes, this response really is typical for an untreated small room!
The action of sound waves colliding and combining in the air is called acoustic interference, and this
occurs in all rooms at all low frequencies - not just those related to the room's dimensions. The only
thing that changes with frequency is where in the room the peaks and nulls occur. The principle is
identical to how phaser and flanger effects work, except the comb filtering happens acoustically in the
air.
The only way to get rid of these peaks and dips is to avoid, or at least reduce, the reflections that
cause them. This is done by applying treatment that absorbs low frequencies to the corners, walls,
and other surfaces so the surfaces do not reflect the waves back into the room. A device that absorbs
low frequencies is called a bass trap. Although it may seem counter-intuitive, adding bass traps to a
room usually increases the amount of bass produced by loudspeakers and musical instruments. When
the cancellations caused by reflections are reduced, the most noticeable effect is increasing the bass
level and making the low frequency response more uniform. As with listening rooms, bass traps are
also useful in studio recording rooms for the same reasons - to flatten the response of instruments
captured by microphones and, with large studios, to improve the acoustics by reducing the low frequency reverb decay time which makes the music sound more clear.
36
For recording engineers, problems caused by standing waves and acoustic interference are often first
noticed when you realize your mixes are not "portable," or do not "translate" well. That is, songs you
have equalized and balanced to sound good in your control room do not sound the same in other
rooms. Of course, variations from different loudspeakers are a factor too. But bass frequencies are the
most difficult to judge when mixing because acoustic interference affects them more than higher frequencies. Another problem is that the level and tone quality of bass instruments vary as you walk
around the room. The sound is thin here, too bassy over there, yet not accurate anywhere. Indeed,
even if you own all the latest and most expensive recording gear, your mixes will still suffer if you can't
hear what's really happening in the low end. Aside from portability concerns, it's very difficult to get the
bass instrument and kick drum balance right when acoustic interference and modal ringing combine to
reduce clarity. And when every location in the room has a different low-end response, there's no way
to know how the music really sounds.
Many people wrongly believe that using near-field monitor speakers avoids the need for acoustic treatment. In truth, even with small loudspeakers playing softly, acoustic interference still causes standing
waves - the imperfect frequency balance is exactly the same but at a lower level. Although higher frequency reflections and echoes are proportionately reduced as you get closer to the loudspeaker, the
skewed frequency response caused by low frequency reflections remains. Likewise, adding a subwoofer will not fix problems that are due to poor room acoustics. While a subwoofer can be useful to
compensate for inadequate loudspeakers, it will not solve the problem of an irregular response caused
by acoustic interference. In fact, a subwoofer often makes matters worse by compounding and hiding
the real problem.
Another common misconception is that equalization can be used to counter the effects of acoustic
problems. But since every location in the room responds differently, no single EQ curve can give a flat
response everywhere. Over a physical span of just a few inches the frequency response can vary significantly. Even if you aim to correct the response only where you sit, there's a bigger problem: It's impossible to counter very large cancellations. If acoustic interference causes a 25 dB dip at 60 Hz, adding that much boost with an equalizer to compensate will reduce the available volume (headroom) by
the same amount. Such an extreme boost will increase low frequency distortion in the loudspeakers
too. And at other room locations where 60 Hz is already too loud, applying EQ boost will make the
problem much worse. Even if EQ could successfully raise a null, the large high-Q boost needed will
create electrical ringing at that frequency. Likewise, EQ cut to reduce a peak will not reduce the peak's
acoustic ringing. EQ cannot always help at higher frequencies either. If a room has ringing tones that
continue after the sound source stops, EQ might make the ringing a little softer but it will still be present. However, equalization can help a little to tame low frequency peaks (only) caused by natural
room resonance, as opposed to peaks and nulls due to acoustic interference, if used in moderation.
Yet another common misconception is that small rooms cannot reproduce very low frequencies, so
they're not worth treating at all. A popular (but incorrect) theory is that very low frequencies require a
certain minimum room dimension to "develop," and so can never be present at all in smaller rooms.
The truth is that any room can reproduce very low frequencies, as long as the reflections that cause
acoustic cancellations are avoided. When you add bass trapping, you are making the walls less reflective at low frequencies, so sound that hits a wall or ceiling will be absorbed instead of reflected. The
net result is exactly the same as if the wall was not there at all - or as if the wall was very far away whatever does come back is greatly attenuated due to distance and, therefore, not loud enough to
cause as much cancellation. See the sidebar Big Waves, Small Rooms for more elaboration on this
topic.
37
Some people mix using headphones in an attempt to avoid the effects of their room. The problem with
headphones is that everything sounds too clear and present, making it difficult to find the ideal volume
for some tracks. When listening through headphones, a lead vocal or solo instrument can be heard
very clearly, even if it is quiet, so you'll tend to make it lower in the mix than it should be. Likewise, it's
difficult to assess the amount of reverb and echo being added electronically when using headphones.
Note that standing waves and acoustic interference also occur at higher frequencies, such as sustained clarinet or flute tones. You can hear the effect and identify the problem frequencies and locations fairly easily by playing sine waves (not too loudly!) through your loudspeakers. This is also a
good way to assess how important bass traps are for your particular studio and control rooms. If you
have SoundForge, WaveLab, or a similar audio editor program, it's simple to create sine wave files at
different low frequencies for testing. Special CDs that contain various tones and pink noise suitable for
room testing and analysis are also commonly available. To determine the severity of low frequency
problems, play different sine waves one at a time through your monitors, and then slowly walk around
the room. It will be very obvious at which frequencies the peaks and valleys occur, and where they
cause the most harm. There's no point in playing frequencies below what your speakers can produce
cleanly - I suggest 60 Hz, 80 Hz, 100 Hz, and so forth through maybe 200-300 Hz. If you have a computer connected to your loudspeakers, you can download the NTI Minirator program, which generates
a variety of useful audio test signals.
Besides helping to flatten the low frequency response, bass traps serve another purpose that is
equally important: They reduce the modal ringing that causes some bass notes to sustain longer than
others, which harms clarity. The ETF 3D "waterfall" graph below shows the modal ringing in my 16'2"
by 11'6" by 8' test lab. Both graphs show not only the low frequency response (the "back wall" of the
graph), but also the bandwidth of each room mode and its decay time. As you can see, adding bass
traps lowers the Q of modal peaks (widens their bandwidth) and also reduces their decay time. When
the bandwidth of the modes is widened individual bass notes stick out less than other, adjacent notes.
This solves the problem commonly known as "one note bass."
The other change is the large reduction in ringing time (the "mountains" come forward over time).
Without traps, some bass notes ring out for as long as 1/3 of a second, so they muddy other subsequent bass notes. After adding bass traps the ringing time is cut in half or even less, except at the lowest mode which in this room is about 35 Hz. But even at 35 Hz there's a noticeable, if slight, improvement in bandwidth and decay time.
38
This ETF graph shows how bass traps reduce ringing by making it decay faster and lowering the Q of the resonances.
39
Generally speaking, most rooms need as many bass traps as you can fit and afford. Although it is definitely possible to make a room too dead at midrange and high frequencies, you probably cannot have
too much low frequency absorption. The effectiveness of bass traps is directly related to how much of
the room's total surface area you treat, which includes the walls, floor, and ceiling. That is, covering
thirty percent of the surface with bass traps reduces low frequency reflections far more than covering
only five percent. It would be great to invent a magical acoustic vacuum cleaner that could suck the
waves out from the air. But, alas, the laws of physics do not work that way. At the minimum I recommend placing bass traps in all of the corners. For even better results, put additional traps on the walls
and optionally on the ceiling.
FIBERGLASS BASS TRAPS
There are a number of ways to create a bass trap. The simplest and least expensive is to install a
large amount of thick rigid fiberglass, spacing it well away from the wall or ceiling. As noted earlier,
705-FRK that is four inches thick and spaced 16 inches away from the wall can be quite effective to
frequencies below 125 Hz. But many rooms have severe problems far below 125 Hz and losing twenty
inches all around the room for thick fiberglass and a large air space is unacceptable to most studio
owners and audiophiles. Fortunately, more efficient bass trap designs are available that are much
smaller. However, studios on a tight budget can apply rigid fiberglass in the room corners as shown in
Figure 3a and lose only the small amount of space in the corners. Since bass builds up the most in the
corners of a room, this is an ideal location for any bass trap.
Figure 3a shows the corner viewed from above,
looking down from the ceiling. When the rigid fiberglass is mounted in a corner like this, the large air
gap helps it absorb to fairly low frequencies. For this
application 705-FRK is better than 703 because the
goal is to absorb as effectively as possible at low
frequencies. However, you can either absorb or deflect the higher frequencies by facing the paper
backing one way or the other, to better control liveness in the room. Using 705 fiberglass that is two
inches thick does a good job, but using four inches
works even better. Note that two adjacent two-inch
panels absorb the same as one piece four inches
thick, so you can double them up if needed. However, if you are using the FRK type you should remove the paper from one of the pieces so only one
outside surface has paper.
Figure 3a: A thick piece of 705 mounted across a corner is effective to fairly low frequencies.
Besides the corners where two walls meet as in
Figure 3a, it is equally effective to place fiberglass in the corners at the top of a wall where it joins the
ceiling. With either type of corner, you can attach the fiberglass by screwing it to 1x2-inch wood strips
that are glued or screwed to the wall as described previously. The 1x2 ends of these strips are shown
as small black rectangles in Figure 3a above. One very nice feature of this simple trap design is that
the air gap behind the fiberglass varies continuously, so at least some amount of fiberglass is spaced
appropriately to cover a range of frequencies.
When mounting 705-FRK directly to a wall - not across a corner - you'll achieve more low frequency
absorption if the paper covered side is facing into the room. However, that will reflect mid and high frequencies somewhat. One good solution is to alternate the panels so every other panel has the paper
facing toward the room to avoid making the room too dead. Panels attached with the backing toward
the wall should be mounted on thin (1/4-inch) strips of wood to leave a small gap so the backing is
free to vibrate. For fiberglass across a corner as shown in Figure 3a, the backing should face into the
room to absorb more at low frequencies.
40
For a typical unfinished basement ceiling you can take advantage of the gap between the support
beams and the floor above by placing rigid fiberglass between the beams. Short nails or screws can
support the fiberglass, making it easy to slide each piece of fiberglass into place. Then cover the fiberglass with fabric as shown below in Figure 3b. You can optionally pack the entire cavity with fluffy fiberglass one foot thick and you'll probably get similar results.
Figure 3b: 705 between support beams, covered with fabric.
Treating a "dropped" grid ceiling is even easier: Simply lay fluffy fiberglass batts on top of the grid,
above the ceiling tiles. The thicker the fiberglass, the better. One foot thick R38 is perfect for this if you
have the space. If you don't want to bother covering the entire ceiling that way, at least put fiberglass
batts around the perimeter to treat the important wall-ceiling corners. And since the fiberglass is not
exposed to the room and doesn't show, you don't need to cover it with fabric.
Another great and inexpensive way to make a bass trap - if you have a lot of room - is to place bales
of rolled up fluffy fiberglass in the room corners. These bales are not expensive, and they can be
stacked to fill very large spaces. Better still, they are commonly available and you don't even have to
unpack them! Just leave the bales rolled up in their original plastic wrappers, and stuff them in and
near the room corners wherever they'll fit. Stack them all the way up to the ceiling for the most absorption.
OPTIMIZING THE AIR GAP
While increasing the depth of the air gap does
indeed lower the frequency range absorbed,
for thinner panels it can also reduce the absorption at some higher bass frequencies. The
maximum amount of absorption for a given
frequency occurs when the air gap depth
equals 1/4 the wavelength for that frequency.
Figure 4 below shows the velocity of a sound
wave, which is greatest at its positive and
negative peaks. Because the velocity is greatest at the peaks, more energy is present to
force the waves through the absorbent material.
Figure 4: A sound wave reaches its maximum velocity at 1/4 of its length.
But at half the length the velocity is at a minimum. Then it rises again at
3/4 length. This pattern repeats indefinitely.
41
The reason an absorbent material like fiberglass works better when spaced away from a surface is
that sound waves passing through it have a greater velocity there. As a wave approaches a boundary,
such as a wall, the velocity is reduced, and when it finally hits the boundary, the velocity is zero. Imagine a cue ball as it approaches the side rail on a pool table. The ball could be travelling 100 miles per
hour, but at the exact point where it hits the rail it is not moving at all.
Likewise, fiberglass placed exactly at a rigid boundary does nothing because the air particles are not
moving there. And since there's no velocity, the fiberglass has very little effect. As fiberglass is spaced
further from the wall, the air particles passing through it have greater velocity. They are slowed down
as they pass through the fiberglass, which converts the sound energy into heat therefore absorbing
some of the sound.
As you can see in Figure 5 above, borrowed from Alton Everest's Master Handbook of Acoustics, absorption for a given
gap depth is maximum at 1/4 wavelength
multiples - in this case starting at around
250 Hz. It then falls off at a higher frequency where the gap depth equals 1/2
wavelength. It rises again when the gap
matches 3/4 of the length of the next
higher frequency, and so forth. This irregular absorption is most severe with
thin absorbing materials, and gradually
diminishes as the material is made
thicker. You can avoid the reduction in
absorption either by using thicker rigid
fiberglass, or by filling the entire gap with
material instead of using only a thin piece
spaced away from the wall or ceiling.
When the entire depth is filled, material is
available to absorb all of the frequencies
whose 1/4 wavelengths fall within that
depth.
Although I promised not to use any math,
I promise that the following simple formula is the only exception. To determine
the best gap depth for a given frequency,
you first need to determine the equivalent
wavelength:
Wave Length in Feet = 1130 / Frequency
Then simply divide the result by 4 to get
Figure 5: Absorbent material is most effective when mounted with an air gap
the optimum depth. So for 100 Hz the
equal to 1/4 the wavelength of a particular frequency. But a gap that is ideal for
wavelength is 1130/100 = 11.3 feet, and one frequency is not ideal for all of the higher frequencies.
1/4 of that is about 2.8 feet. The number
1130 is the approximate speed in feet per second of sound waves travelling through air at normal
room temperature and humidity. For a given thickness of absorbent material, the ideal air gap is equal
to that thickness because it avoids a hole in the range of frequencies absorbed. For example, if you
install fiberglass that is four inches thick with a four-inch gap, higher frequencies whose 1/4 wavelength falls within the four-inch material thickness are absorbed regardless of the gap. And for those
frequencies whose 1/4 wavelength is between four and eight inches, the fiberglass is also at the
proper distance from the wall or ceiling. This is shown below in Figure 6.
42
In practice, you don't necessarily have to measure
wavelengths and calculate air gaps, and the first few
inches of space yield the most benefit. Most people are
not willing to give up two or more feet all around the
room anyway, so just make the gap as large as you
can justify. If you can afford to fill the gap entirely with
material, all the better. And even though the velocity is
indeed highest at 1/4 wavelength, there's still plenty at
1/8th of the wavelength too. Note that the angle at
which sound waves strike a fiberglass panel can make
the panel and its air gap appear thicker than they really
are. Further, low frequency waves that strike an absorbing panel at an angle may be absorbed less than
when they strike it at 90 degrees, due to a "grazing"
effect. The explanations in this section are a simplification and are correct only for a 90 degree angle of incidence, which is not always the case.
I should mention another popular type of absorber, the
tube trap, which is available commercially and also as
do-it-yourself plans on various web sites. Although
these are often referred to as "bass traps," even the
largest tube models are not very effective below about
100 Hz, and the smaller ones become ineffective much
higher than that. Marketing hype aside, the real absorption mechanism in a tube trap is simply the rigid
fiberglass inside. The reason a 20-inch tube trap works
at all down to 100 Hz is that the tube's diameter serves
to space some of the fiberglass away from the nearest
boundary, which helps extend its absorption to a lower
frequency. But a tube design is no more effective than
using plain rigid fiberglass spaced similarly.
Figure 6: The higher frequencies (top) are absorbed well because their velocity peaks fall within the material thickness. The
lower frequency at the bottom does not achieve as much velocity so it's absorbed less.
BETTER BASS TRAPS
Yet another type of bass trap is the Helmholtz resonator. Unlike foam, fiberglass, and tubes fitted with
fiberglass, a Helmholtz resonator can be designed to absorb very low frequencies. This type of trap
works on the principle of a tuned cavity and is often very efficient over a narrow range of frequencies.
Think of a glass soda bottle that resonates when you blow across its opening, and you have the general idea. Although a Helmholtz design can be very efficient, the downside is that it works over a fairly
narrow range and needs to be rather large to absorb very low frequencies. The range can be widened
by filling the cavity with fiberglass, or by creating several openings having different sizes. One common design uses a box filled with fiberglass with its front opening partially covered by a series of thin
wood boards separated by air spaces. This is called a slat resonator. Another also uses a box filled
with fiberglass but has a cover made of pegboard containing many small holes. Although there is no
denying that a Helmholtz trap can be very effective, the fact that it works over a narrow range of frequencies limits its usefulness. While it can be sized to absorb the dominant resonant frequencies in a
particular room, it cannot absorb all the other low frequencies. And broadband absorption is needed to
prevent acoustic interference that skews the frequency response throughout the entire bass range.
43
One of my favorite types of bass trap is the membrane absorber, also called a panel trap because it's
made with a wood front panel. One huge advantage of membrane traps is that they do not have to be
very thick to absorb very low frequencies. Because the bass range spans about four octaves, most
panel traps are designed to work over only part of the bass range. Therefore, you will need an equal
mix of trap types, with one intended to absorb the lower bass frequencies and the other for the higher
bass range. Besides absorbing low frequencies very well, the wood front on a panel trap is reflective
at higher frequencies. So installing enough of them to treat a room properly for low frequency problems will not make the room too dead sounding at mid and high frequencies.
The photo below shows eight panel traps I built for my home studio. Besides the panels, which are
painted white, there are also many 703 fiberglass absorbers covered with tan fabric. Not shown are
four more panel traps in the rear corners, plus another four on the side walls farther back in the room.
The photo shows both types of panel traps (low-bass and high-bass), with the thinner units absorbing
the higher bass range. Since this is a fairly large room for a home studio (18 by 34 feet), many traps
are needed to cover a significant amount of the room's surfaces. A smaller room would need fewer
traps to cover the same percentage of surface area.
This room has an even mix of low-bass and high-bass panel traps, and an equal number of fiberglass absorbers to handle the
midrange and high frequencies. More bass traps are in the rear of the room.
Figure 7 below shows a cut-away view of a typical wood panel membrane trap. When a wave within
the effective range of frequencies reaches the front panel, the panel vibrates in sympathy. Since it
takes energy to physically move the panel, that energy is absorbed rather than returned into the room.
The fiberglass then damps the plywood panel so it doesn't continue to vibrate. Were the panel allowed
to vibrate freely on its own, less energy would be needed to keep it moving, so it would absorb less.
Further, a panel that continues to vibrate on its own after the source sound stops actually generates
sound similar to reverb and the ringing effect described earlier, and obviously that is not desirable!
44
Similar to an acoustic suspension loudspeaker, panel
absorbers like this one are sealed air-tight, and the
fiberglass converts the acoustic energy into heat. Note
how the fiberglass is spaced away from the back
panel, which is more effective than simply attaching it
directly against the rear surface. The closer the fiberglass is to the plywood panel, the more effectively it
damps the panel's vibration. But it is important that the
fiberglass not touch the panel because that would restrict its movement. For a panel trap to absorb as efficiently as possible, the panel must be free to vibrate
with no restriction other than the damping action of the
nearby fiberglass.
There are a few reasons for sealing panel traps. If
there's a place for air to escape - let's say at the seam
between the front panel and the side of the box - then
pressure from the diaphragm as it pushes into the box
will send the waves out the leak rather than push them
into the fiberglass. Another, more relevant, reason is
that a leak will let the internal pressure escape, reflecting the waves back into the room instead of absorbing
them. Think of a panel trap as equivalent to an open
window. If you cut a hole in an outside wall and cover
the hole with a piece of heavy cardboard, the cardFigure 7: Sound striking the plywood front panel causes
board will reflect mid and high frequencies but let lower it to vibrate. The fiberglass then damps that vibration.
frequencies pass through. Those frequencies end up on
the outside of the wall and so are not reflected back into
the room. A sealed membrane bass trap is similar in that sound passing through the panel goes into
the box and does not come out. But the most important reason a panel trap must be sealed is because the air inside acts as a spring, and an air leak reduces that effect.
Although mounting fiberglass across the corner of a room is best for treating bass frequencies, panel
traps work on a different principle where the gap does not help. So with panel traps it's better to put
two in each corner, flat against the wall, because that provides twice the surface area of just one trap
mounted across the corner. Bass traps built from porous materials like fiberglass and acoustic foam
work by absorbing the sound waves as they pass through the material. This type of trap is called a
velocity absorber because it is velocity (speed) that drives the sound wave into the absorbing material.
A wood panel trap works on an opposite principle, wave pressure, and is considered a pressure absorber because the wave pressure is greatest at the room boundaries. You can think of a wood panel
bass trap as being a "shock absorber" for sound waves. As a wave approaches a wall it has plenty of
velocity (the speed of sound) but no pressure. And when it hits the wall there's no longer any velocity
but now there's plenty of pressure. This is similar to driving a car into a tree. You can be going 60
miles per hour toward the tree - lots of velocity - and the instant you hit the tree there's no velocity but
plenty of pressure!
45
AS GOOD AS IT GETS?
Okay, so how much improvement can you really expect after installing bass traps? Unless you cover
nearly all of the wall and ceiling surfaces with material that is 100% absorbent at all problem frequencies - which is pretty much impossible - you will still have some deviation from a perfectly flat response. But even with a more practical amount of treatment, you can achieve far less severity in the
ripples and also increase their bandwidth so the peaks and dips are broader, making them less damaging and less likely to affect single bass notes. You'll probably still have peaks and dips that you can
measure and identify with sine wave tests, but the music will sound much better, and the bass levels
will vary much less around the room. The difference in frequency response of a room with and without
membrane panel bass traps is shown in Figure 8 below.
Figure 8: That's more like it! This response plot is for the same control room shown in Figure 2 but after treatment with wood panel bass traps.
PART 2: ROOM DESIGN AND LAYOUT
One of the most important properties of a room is its modes, or natural resonant frequencies which
are related to its length, width, and height. More often than not the room you use for a studio or home
theater has already been built, so knowing the modes and other permanent properties of the room is
academic at best. After all, what's the point in calculating the modes if you can't do anything about
them? And since all listening rooms need treatment at all low frequencies, knowing the modes doesn't
even help you determine what type of bass traps you need. Perhaps you are lucky enough to have the
luxury of designing an audio room and controlling its size and shape before it is built. In that case you
can make a meaningful difference in the room's acoustic qualities by carefully choosing proper dimensions. If not, there's still plenty you can do to make an existing room as good as it can be. Top
ROOM SIZES AND SHAPES
The size and shape of a room determines its natural resonances - often called room modes. Every
rectangular room has three sets of primary modes, with one each for the length, width, and height. If
you have an irregular room shape or angled walls, you can average the dimensions to get a rough
idea of the mode frequencies. That is, if the length wall is angled, making the width 10 feet at one end
and 12 feet at the other, you can use 11 as the average for the width dimension. Rooms with irregular
shapes, such as an alcove, have more than three sets of modes and are more difficult to calculate.
46
Generally speaking, larger rooms are better acoustically than smaller rooms because the modes are
spaced more closely, yielding an overall flatter response. Acoustics experts recommend a minimum
volume of at least 2500 cubic feet for any room in which high quality music reproduction is intended.
Figure 9 below shows the modes for just one dimension - let's say the length - of two different rooms.
Here, the larger room (top) has a
length of 28 feet, so the fundamental mode frequency, which occurs at
half the wavelength, is 20 Hz. Subsequent modes, similar to harmonics of a note played by a musical
instrument, occur at 20 Hz intervals.
Even though this creates many little
resonant peaks in the response, the
peaks are close together, so the
average response is fairly flat. And
as one peak is falling, the adjacent
peak is rising to help compensate
and fill in the void. (Please note that
Figure 9: In a large room (top), the resonant peaks caused by modes are
Figures 9 and 10 are approximatogether than those in a small room (bottom). The closer spacing
tions as drawn in a graphics program, so the closer
yields an overall flatter response.
shapes of the peaks and dips are not truly
accurate.)
Now consider the length modes for the smaller room, shown at the bottom of Figure 9. Here the first
peak is at 60 Hz, which corresponds to a half wavelength of about 9-1/2 feet. Therefore, subsequent
modes occur at 60 Hz intervals making the overall response less uniform because a wider range of
frequencies is attenuated, and more deeply, between each set of peaks.
Another important factor in
the design of studios and
listening rooms is the ratio
between the length, width,
and height. The worst
shape is a cube having all
three dimensions the same.
A cube has the fewest number of peaks, and therefore
the greatest distance between peaks, because all
three dimensions resonate
at the same frequencies. In
an ideal room, each dimension will contribute peaks at
different frequencies, thus creating
more peaks having a smaller distance
between them. This is shown in Figure
10.
Figure 10: The modes for a room with ideal ratios (top) give a more even response
overall than a room with poor ratios (bottom). When the room proportions are less
than ideal, some of the natural resonances are spaced far apart while others are
clustered very close together.
47
Besides making the overall response less uniform, uneven mode spacing can make one note on a
bass instrument louder than adjacent notes. This is much worse than having a gentler curve created
from many in-between peaks that, even if not flat, affects a wider range of notes. The principle is similar to using EQ to boost midrange presence in a recording - a broad boost always sounds more natural than a narrow one. Narrow peaks tend to impart a nasal quality that sounds like a wah-wah pedal
set to a fixed point near the middle of its range. Note that besides creating peaks in the frequency response, the modes also determine at which frequencies the room's natural reverb is most pronounced. It is better for a room's reverb to be even across the spectrum rather than comprise a few
dominant frequencies, which colors the sound unnaturally. So for all of these reasons a room should
have different and non-related dimensions for the length, width, and height. When all three dimensions
are the same - the worst case - you get widely spaced resonant peaks at the fundamental frequency
and its harmonics only. With different dimensions you have more peaks at more in-between frequencies, which taken together gives an overall flatter response.
There are a few "ideal" ratios of room height, width, and length that professional studio designers
agree should be used if possible. Three of these ratios, developed by L.W. Sepmeyer, are shown in
Table 2.
Height
1.00
1.00
1.00
1.60
2.33
Table 2: The ideal room has a ratio of height, width, and length similar to
one of these.
There are other good ratios, but those shown above are the ones I see referenced most often. Note
that when a room has a suspended tile ceiling the real height, as far as low frequencies are concerned, is to the solid surface above the tiles. Likewise, in a basement with exposed joists the true
height is to the bottom of the floor above, not the bottom of the joists.
That said, I believe the importance of room modes is often overstated. You don't want the width to be
the same as the depth or an even multiple such as 10 feet by 20 feet. But the modes just describe
where the resonances will be worst. Regardless of the room's size and shape, standing waves and
acoustic interference happen at all low frequencies. So you still need bass traps that handle the entire
range, not just the frequencies determined by the room modes. As far as acoustic interference is concerned, the only thing that changes with different room dimensions is where in the room the peaks and
dips at each low frequency occur.
There are many freeware and web-based room mode calculators, but all the ones I've seen just list a
table of the modes, so you still have to plot them by hand on semi-log graph paper to get a sense of
how close they are to each other. Here is a link to ModeCalc (only 57 KB to download), a room mode
calculator I wrote that runs in DOS and Windows. It plots the first ten primary room modes graphically
so you can see how the modes are distributed and how they relate to one another. The modes for
each dimension are displayed in a different color, and when two or more modes occur near the same
frequency, the duplicates are shown on a separate line so one does not obscure the other. The program is easy to use, and pressing F1 displays complete instructions and explains how to interpret the
results. Since the instruction manual contains additional explanations about room modes, it is reprinted below in the sidebar Room Modes and ModeCalc.
48
ROOM SYMMETRY
Unless you plan to record and mix in mono only, the symmetry of your room and loudspeaker placement are very important. If both loudspeakers are not situated symmetrically in a room they will have a
different frequency response, and your stereo imaging will not be balanced. In a room that is longer
than it is wide, it's better to place the speakers near the shorter wall so they fire the long way into the
room as shown on the left in Figure 11 below. This puts you farther from the rear wall where the low
frequency peaks and nulls are most severe.
Figure 11: Symmetry matters! In a typical stereo mixing room, the loudspeakers are spaced equally from the walls and corners, and form an
equilateral triangle at the mix position. The arrangement shown on the left above is better than the one on the right because it's more symmetrical within the room. The layout on the right also suffers from a focusing effect caused by the wall-wall junction behind the listener.
Besides positioning the loudspeakers symmetrically, you should also place your console and chair so
your ears are the same distance from each speaker. Likewise, acoustic treatment - whether absorption or diffusion - should be applied equally on both sides. In many home studios it is not possible to
create a completely symmetrical arrangement, but you should aim for as close to this ideal as possible. Especially in the critical front part of the room where the first reflections to reach your ears are
those from the side walls, and from the floor and ceiling if they're not treated with absorbent material.
What happens in the rear of the room is probably less important.
Although the sample rooms shown above in Figure 11 are rectangular, I prefer angled walls and an
angled ceiling because that provides deflection which reduces flutter echoes and ringing. Some people argue that parallel walls are preferred because you can better predict the room modes, and then
treat the inevitable flutter echoes with absorption. But as I explained earlier, simply knowing the
modes is not always that valuable, and with angled walls you can make the average dimensions comply with the ideal ratios. Further, if a room has parallel walls that must be treated with absorptive material to avoid echoes and ringing, you may not be able to make the room as live as you'd like. See the
sidebar Creating a Reflection Free Zone for more related information.
49
A peaked ceiling is better than a flat ceiling because it avoids the echoes and ringing that occur when
the ceiling is parallel to the floor. But a peak creates a focusing effect, much like a parabolic dish,
which is less than ideal. For this reason it's a good idea to place absorption or diffusion under the
peaked portion, as shown in the photo below.
These MiniTraps (commercial acoustic
panels) were installed under the peaked
ceiling in the author's home recording
studio to avoid focusing sound in the
room to the area under the peak. Top
One somewhat controversial aspect of control room design is soffit mounting the main loudspeakers.
Most home studio owners simply put their speakers on stands, or sit them on the mixing desk, and
leave it at that. But many pro studios prefer to install the speakers into the wall so the front surface of
the speaker cabinet is flush with the wall. There are sound scientific reasons to use soffit mounting,
yet some engineers say it's not necessary or that it gives poorer results. Those in favor of soffit mounting point out that it reduces reflections called Speaker Boundary Interference, or SBIR, that cause
peaks and dips in the low frequency response. If a loudspeaker is out in the room away from the wall,
low frequencies from the rear of the cabinet will bounce off the wall behind it and eventually collide
with the direct sound coming from the front of the speaker. (Even though it may not seem obvious,
very low frequencies do in fact leave a speaker cabinet in all directions.) Proponents also claim that
soffit mounting improves stereo imaging by reducing mid and high frequency reflections.
I happen to side with those in favor of soffit mounting, yet I also respect the opinions of those who disagree. One thing nobody will dispute is that soffit mounting requires a lot more effort! If you do use
soffit mounting, please understand that the speakers must be built into the real wall. You can't just apply a lightweight facade around the front of the speaker cabinet and expect the same results.
50
LIVE OR DEAD - WHICH IS BEST AND WHERE?
If you've ever seen photos of high-end recording studios in magazines, you probably noticed that the
studio room floors almost always use a reflective material like wood or linoleum. A hard floor gives a
nice ambience when miking drums, guitar amps, and acoustic instruments. Likewise, auditorium
stages and school band rooms always have a reflective floor surface too. As mentioned earlier, "live"
in this context refers only to mid and high frequencies. I cannot emphasize enough the importance of
a reflective floor for achieving a natural sound when recording acoustic instruments. If you record in
your living room and your spouse refuses to let you remove the carpet, get a 4- by 8-foot sheet of 1/4inch plywood to put over the carpet when recording. You can cut it in half for easier storage and put
the halves next to each other on the floor when needed.
Control room floors are sometimes carpeted, sometimes wood, and often a combination of the two.
Ceilings in these types of rooms also vary between fully reflective, fully absorptive, or a mix of surface
types. There is no one correct way to treat every room because different engineers prefer a different
amount of liveness. However, you should never make a room completely dead because that produces
a creepy and unnatural sound. The only time you might consider making a room entirely dead is when
treating a small vocal booth or a very small studio or control room - smaller than, say, ten by ten feet.
When a room is very small the reflections are too short to be useful and just make the room boxy
sounding. In that case the best solution is to cover all of the surfaces entirely with absorbent material
and, for a studio room, add any ambience electronically later.
In a more typical room I recommend a mix of hard and soft surfaces for the walls, with no one large
area all hard or all soft. I suggest applying absorbent material to the walls using stripes or a checkerboard pattern to alternate between hard and soft surfaces every two feet or so. This makes the room
uniformly neutral everywhere. You can make the spacing between absorbent stripes or squares larger
or smaller to control the overall amount of liveness. If you are using 705-FRK rigid fiberglass or an
equivalent product, you can cover more of the wall and still control the liveness by alternating the direction of the paper backing. That is, one piece of fiberglass will have the paper facing the wall to expose the more absorbent fiberglass, and the next piece will have the paper facing out to reflect the
mid and high frequencies. In fact, when the paper is facing into the room the lower frequencies are
absorbed even better than when it is faces the wall.
Alternating hard and soft surfaces is also advisable with wood panel bass traps - simply place a fiberglass absorber between each trap. You can see this arrangement in the photo of my studio (above
Figure 7), where each type of bass trap alternates with the other type and with fiberglass panels. That
is, first is a low-bass trap, then a fiberglass panel, then a high-bass trap, then fiberglass, then lowbass, and so forth. I'll also mention that wood panel bass traps can be mounted horizontally when
book shelves and other obstacles prevent placing them vertically. Since the corner formed by a wall
and the ceiling, or a wall and the floor, is just as valid as any other corner, mounting a panel trap sideways near the top or bottom of a wall is equally effective.
Of course, many studios do have large live areas, and there's nothing wrong with that! If the room is
big enough to avoid short echoes between closely spaced walls, having an entire wall reflective can
yield a very big sound. And even in smaller rooms a hard floor with one or more bare walls can be
useful. My cello teacher, who is a total audio neophyte, blew me away with the quality of a recording
she made in her small Manhattan apartment. She recorded while playing with her back against the
corner, facing into the room, using an inexpensive stereo mike placed a few feet in front of her cello.
The key to a realistic and present sound, especially for acoustic instruments, is capturing some
amount of ambience - even when the reverberation of a large space is not appropriate.
51
Although it is often desirable to alternate hard and soft surfaces on the walls, I often recommend covering the entire ceiling with absorbent material, especially if the ceiling is low. Besides eliminating floor
to ceiling flutter echoes, full absorption can make the ceiling appear acoustically to be much higher.
Most home studio owners cringe at the thought of making their ceilings even lower than they already
are, but it really can help the sound. If you cover the entire ceiling with 2- to 4-inch thick 705, suspended with strings or wires to leave an air gap, the room will sound as if the ceiling were much
higher. There's no difference between reflections that are reduced by the greater distance of a high
ceiling and reflections from a low ceiling that are reduced by absorption. Using thick, dense fiberglass
extends the simulated increase in height to lower frequencies. Where thin fiberglass makes the ceiling
appear higher at midrange and high frequencies, using thicker and denser fiberglass with an air gap
raises the apparent height at lower frequencies as well.
Another advantage of full absorption on a low ceiling is that it avoids the comb filtering that occurs
when miking drums and other instruments from above. Placing microphones high over a drum set or
string section puts the mikes very close to the ceiling. If the ceiling is reflective, sound will arrive at the
mikes via two paths - the direct sound from the instrument and the same sound after being reflected
off the nearby ceiling. When the difference in distance is very small, let's say one foot, the reflections
cause many peaks and dips in the response, which are very audible and can sound like a flanger effect. (When reflections cause a series of peaks and dips, the effect is often called comb filtering because the frequency response plot resembles a hair comb.) Again, reducing strong reflections from a
nearby ceiling via total absorption is acoustically identical to having a ceiling that's infinitely high.
NOISE CONTROL
Reducing noise and sound leakage is beyond the scope of this article, but I will share a few tips studio
owners may find useful. If your studio has forced air ventilation, be sure to place the microphones
away from the vents while recording. If the vents have adjustable deflectors, set them to direct the air
away from where you normally place your microphones. Better, allow the room to get to the desired
temperature before you start recording so you can turn off the blower. You can turn it on again between takes if needed. Likewise, radiators often make creaking sounds due to expansion and contraction as they warm up and cool down, so use them before you start recording.
Another troublesome noise source in many studios is the fan noise from a computer. You can buy a
low noise replacement power supply from PC Power and Cooling and other companies. Easier, buy a
computer from one of the better manufacturers because they often have much less fan noise than the
cheaper brands. My last three computers were Dells, and they have all been very quiet. The small
premium you pay for a better brand is easily gained back by not replacing the power supply or having
to build or buy a sound proof enclosure.
I also attached 703 fiberglass wrapped with fabric to the rear and underside of my desk, as shown in
the photo below (left), to absorb the fan noise rather than reflect it into the room. Between the Dell's
quiet power supply and the fiberglass, I can record myself playing the cello or acoustic guitar while
sitting in front of the computer, with the mikes pointed right at me and the computer, and still pick up
very little noise. A second piece of 703 (right) can be placed in front of the computer to reduce the
noise even further while recording.
52
One easy way to reduce noise from a computer is to line the surrounding surfaces with absorbent material.
If you've done all you can to reduce ambient noise and it's still too loud in a recording, consider using
digital noise reduction. Many programs are available that do a remarkable job of removing any type of
steady noise - not just hiss, but hum and air conditioning rumble too - after the fact. I use Sonic Foundry's Noise Reduction plug-in, but other affordable programs are available that also do an excellent
job.
MORE RESOURCES
I have tried to make this article as complete as possible, but it is impossible to cover every aspect of
acoustics. Many books have been written about acoustics and studio design, and my goal here has
been to cover only the issues that are most important to recording engineers and audiophiles. Further,
acoustics is as much an art as a science, and surely mine are not the only valid opinions. Fortunately,
the Internet offers many resources for more information including my own Acoustics forum at EQ
Magazine, John Sayer's Studio Design forum, the SAE web site, the Acoustics newsgroup, and Angelo Campanella's Acoustics FAQ. Perhaps the most valuable resource of all is Google, where you
can find web pages that cover nearly any topic.
SIDEBAR: WHY THEY'RE CALLED STANDING WAVES
If you've ever used an ultrasonic cleaner to clean jewelry or small electronic components, you've
probably seen standing waves in action. When you drop a pebble into a pond, a series of waves is
created that extends outward from the point of impact. Since a pond is large, the waves dissipate before they can reach the shore and be reflected back to the place of origin. But in a contained area like
the tub of an ultrasonic cleaner, the waves bounce off the surrounding walls and create a pressure
front that makes them literally "stand still" within the cleaning solution. The exact same thing happens
in your control room when your loudspeakers play a sustained bass tone. Static nodes develop at different places in the room depending on the loudspeaker position, the room's dimensions, and the frequency of the tone.
53
SIDEBAR: FINE TUNING THE CONTROL ROOM
Some recording studios use 1/3 octave graphic equalizers to flatten the frequency response of their
monitor speakers. These equalizers are inserted into the signal path between the mixer and power
amplifiers, to counteract the inevitable bumps and dips in any loudspeaker/room combination. Adjusting the frequency response of a room with equalizers is called "tuning the room."
When a manufacturer publishes response curves for a speaker, the measurements were made in an
anechoic chamber - a room that is completely absorbent at all but the lowest frequencies. Eliminating
reflections ensures that the measured response is accurate and not skewed by room reflections. But
like a car maker's inflated mileage claims, measuring a loudspeaker's response in an anechoic chamber does not reflect reality. What really matters is the frequency response in your room.
There are different philosophies about the best way to tune a control room, and no one method is correct. What you do - if you do anything at all - depends on your mindset, the size of your wallet, and
perhaps the kind of music you produce. Many people are satisfied to adjust the tweeter level on their
speakers if one is provided and accept the results. If the speakers are bi-amped, the relative level between low and high frequencies can be further adjusted with the controls on the electronic crossover.
Acoustic treatment as described in this article goes a long way toward eliminating response-skewing
reflections, and with a properly treated room, equalization may not be worth the effort and expense.
Further, all loudspeakers have their own unique "sound," and it's not wrong to pick speakers that
sound the way you like and simply leave it at that.
Another philosophy is to aim for a perfectly flat response at any cost. Once the speakers have been
made as flat as possible by adjusting the built-in controls and crossover, equalizers are added to the
monitor signal path. If, for example, a 3 dB dip is measured at 1 KHz, the equalizers are set to boost
that frequency by 3 dB to compensate. In practice the left and right equalizers are usually adjusted
independently, since each speaker and its location in the room may require a different correction.
My personal philosophy is to avoid room equalizers because they can introduce as many resonant
peaks and valleys in the response as they remove. And as explained in the main text, room equalization that improves the response in one part of the room almost always makes it worse in other places.
Years ago I added a small amount of EQ cut at 400 Hz to the monitor system of my home recording
studio, to correct a measured boost at that frequency. But after a year or so I removed the EQ because it seemed to make my mixes worse. I have heard similar reports from other studio owners - after going to great effort and expense to equalize their loudspeakers to make them perfectly flat, the
result was generally poor and their mixes sounded worse in other rooms instead of better. But if you
want to measure your own room just to know its response, following are a few ways to do that.
The old fashioned way to measure the frequency response of speakers in a room uses a pink noise
signal source, a sweepable filter that passes 1/3 octave bands one at a time, a high-quality smalldiaphragm omnidirectional condenser microphone, and a voltmeter with a decibel readout. All-in-one
spectrum analyzers are available that combine these components into a single package, making the
measurements fairly easy to perform. There are also software programs available that use your PC's
sound card to play pink noise while recording from the microphone, and then display the room response.
54
Place the microphone at ear level where you sit while mixing. Then play the pink noise in mono
through both of your loudspeakers, loudly enough to drown out all ambient noise by at least 30 dB.
That is, the difference shown on the record level meter should be at least 30 dB between playing the
noise and not playing the noise. Record about ten seconds of the noise from the mike onto a high
quality medium like a DAT or a computer - 16 bits at 44.1 KHz is fine. Now load the file into SoundForge or another audio editor that offers a spectrum analyzer, and view the results at the highest resolution offered. You can also measure each speaker separately, if you'd like, or place the microphone in
various locations in the room to measure the response in those places. The microphone does not
have to be perfectly flat, as long as you know what its response really is and incorporate the deviation
into your measurements. If you have an expensive mike, it probably came with a custom printed frequency response curve, and you can add/subtract that curve from your measurements to remove the
mike's contribution to the results.
Another test uses plain sine waves, which more closely reflects real music, such as a bass player sustaining a note on a slow ballad. The sine wave test is similar to the test that uses pink noise, except
you'll play a series of low frequencies one at a time and read the results on a meter. Start at the lowest
frequency your loudspeakers are rated to reproduce, and continue at 1 Hz intervals up to about 300
Hz or so. I also suggest that you play a few individual tones and walk around the room listening for
places where the tones get louder and softer. In most rooms if you play a single frequency, the difference between loud and soft will be 15 dB or more, and complete cancellations are common.
There are also more modern and sophisticated analysis tools such as TEF and the ETF and Smaart
programs. These systems measure much more than just the raw frequency response. They also take
into account the delay time and frequency content of the reflections, the room's reverb time at different
frequencies, and they offer more sophisticated ways to view and interpret the results. I use the ETF
program and find it most worthwhile, especially considering its very reasonable cost. Top
SIDEBAR: MEASURING ABSORPTION
The standard way to specify the effectiveness of absorbent materials is with an absorption coefficient.
This number ranges from zero to 1.0, with zero doing nothing and 1.0 being 100 percent absorbent.
Since all materials absorb more at some frequencies than at others, the absorption coefficient values
are also accompanied by a frequency. This frequency is really an average of all frequencies within the
stated third-octave band.
Material that has an absorption coefficient of 0.5 at a given frequency absorbs half of the sound and
either reflects or passes the rest. For example, 703 rigid fiberglass that is two inches thick has an absorption coefficient of 0.17 at 125 Hz. Because this frequency is at the low end of the fiberglass's useful range, the other 83 percent of the sound passes through it. On the other hand, 705-FRK fiberglass
becomes more reflective at higher frequencies because of the metalized paper facing, so its absorption coefficient of 0.34 at 4 KHz means that the other 66 percent is reflected off the surface back into
the room. Out-of-band frequencies for other materials are also either reflected or passed. A wood
panel bass trap that absorbs well between 50 Hz and 200 Hz reflects most of the higher frequencies
because of the hardness of the wood front panel, while lower frequencies instead pass through the
trap to the wall behind. Top
Note that sound usually passes through absorbent material twice - once on its way toward the wall,
and once again after being reflected off the wall back toward the room. Unless, of course, it's a very
low frequency that first passes through the material, and then passes through the wall too. Even when
a material like foam or fiberglass passes a low frequency instead of absorbing it, the wall will likely reflect the sound, so the net result is reflection.
55
Some vendors specify absorption using sabins instead of absorption coefficient, perhaps because it
obscures the results and gives a larger number that is more impressive looking! After all, who wouldn't
prefer a product that offers 9.0 sabins of absorption compared to one having an absorption coefficient
of only 0.12? However, specifying absorption in sabins is sometimes justified, such as for tests of nonstandard materials or when using unconventional mounting where the standard methods do not apply.
You can convert sabins to an equivalent absorption coefficient by simply dividing the sabins by the
square feet of front surface area.
Besides giving the absorption values at different frequencies, many product specifications also include
the Noise Reduction Coefficient, or NRC. This is an average of just the midrange bands (250 Hz
through 2.0 KHz) and is not useful when comparing materials for recording studios and music rooms.
For example, one material may absorb mainly low frequencies while another works best at higher frequencies, yet both can have similar NRC values. Table 1 in the main text shows the NRC for 2-inch
foam as 0.80 while the same thickness of 705-FRK is only 0.60. Yet 705-FRK is nearly six times more
absorbent than foam at 125 Hz!
Absorption is typically measured in a special room that is very reverberant at all frequencies. The reverb decay time is measured at each frequency of interest with the room empty, and then again with
the test material present. By comparing the difference in reverb times with the room empty and with
the test material in place, the amount of absorption can be computed. Some minimum amount of material is required for testing so that the difference in decay times is large enough to ensure accurate
measurements. In the tests I observed at IBM's acoustic labs, at least 64 square feet of material is required in order for the test results to be certified.
The standard way to measure reverb time is to play an impulse sound, such as a burst of pink noise
through loudspeakers, and then measure how long it takes for the sound to fall by 60 dB. This type of
test is called RT60, where RT stands for Reverb Time and 60 indicates the time it takes to fade by that
many decibels. But it's difficult to measure levels that low because of ambient noise in the measuring
room. So more often these tests measure the time it takes the reverb to decay by 30 or even 15 dB,
and the time it would have taken to fall the full 60 dB is calculated from that.
At IBM's lab broadband pink noise bursts are played through loudspeakers, and a high quality microphone records each burst and its decay. A single test takes about 40 minutes to run because one hundred separate noise bursts are played. Since pink noise is random in nature the results of all the tests
are averaged, separated into 1/3 octave bands. While the noise is playing the microphone that records
the sound is constantly moving around the room. Instead of just placing a mike in one location, a special motorized boom stand rotates slowly in all three dimensions. That is, it swings around the room in
a circle and also goes up and down from a few feet off the floor to seven or eight feet high. This way
the reverberation at many places in the room is averaged into the measurements.
Another type of absorption test places the material being tested in a device called an impedance tube.
This is a long narrow sealed chamber made of concrete or brick in which standing waves along the
length of the tube are measured with and without the test material present. When using either test
method, the ambient temperature and humidity must be constant and known precisely, as these affect
the absorption of air and thus must be factored into the measurements.
Labs that perform acoustic tests are certified by NVLAP, a department of the US Government's National Institute of Standards and Technology. Testing of acoustic materials is defined by ASTM, a US
organization that establishes standards and practices used by acousticians and their companies. By
ensuring that its members follow exactly the same rules and guidelines, materials tested to ASTM
standards in different facilities can be compared with confidence.
56
Most absorption measurements are taken with the test material mounted directly to a wall. But since
spacing absorbent material away from a wall improves its low frequency performance, absorption figures that include spacing are often included in published specifications. In this case the specs indicate
the type of mounting and also the spacing. The "A" mounting method means the material is flat
against a wall, and E-### means the material was spaced, where the number (###) indicates the size
of the air gap in millimeters. E-400 is common, which is about 16 inches. When "E" mounting measurements are made according to ASTM standards, a reflective skirt is applied around the edges of the
material that extends to the mounting surface. This prevents reflections from bouncing off the mounting surface at an angle and entering the material from behind, which would wrongly increase the absorption measurements.
You may notice that absorption coefficients sometimes have a value greater than 1.0. Although it is
impossible for any material to absorb more than 100 percent of the sound, measurements can yield a
value greater than 1.0. The main reason this occurs is that all material has a finite thickness, and the
edges - which are not included in the stated surface area - absorb some of the sound. So for a piece
of 703 fiberglass that is two by four feet and four inches thick, the real surface area includes the fourinch thick edge around the material. If included in the measurements, this would add four square feet
to the stated surface area of eight square feet - a 50 percent increase! (See the sidebar The Numbers
Game for a more detailed explanation.) But even when the edges are included in the total surface
area, values greater than 1.0 are still possible due to diffraction effects at the material's corners. When
the corners are rounded, this effect is reduced.
As you might imagine, the fee to use a lab that performs certified acoustical testing is high because it's
very expensive to build a reverberation test room. Such a room must have a very low ambient noise
level, which requires isolated structures, special sound proof doors, and a low air flow ventilation system. Building a room large enough for testing very low frequencies is even more expensive. For this
reason it is rare to see absorption specifications that extend below 100 Hz - it simply costs too much
to build a room that can measure absorption accurately below that frequency. Further, most industrial
manufacturers - the main customers of testing labs - do not need measurements at very low frequencies. However, even when a room is certified down to only 100 Hz, it is still possible to assess relative
absorption. That is, you can test different materials at, say, 50 Hz and see which are more absorbent
even if the absolute measurements are not guaranteed accurate.
SIDEBAR: THE NUMBERS GAME
Acoustic products are commonly specified by their absorption coefficient. This number ranges from
zero (no absorption) to 1.0, meaning 100 percent of the sound is absorbed. For example, an absorption coefficient of 0.5 means that half the sound is absorbed and the other half either passes through
the material or is reflected. Since no material absorbs all frequencies by the same amount, absorption
coefficients are usually given for different frequency ranges.
Although 1.0 is the largest legitimate value possible, you may have seen higher numbers claimed for
some products. Needless to say, this causes confusion, and makes it difficult to compare published
data. Once you understanding how absorption is measured, and how data can be manipulated - both
fairly and unfairly - you'll be able to assess room treatment products and materials more wisely.
57
Acoustic absorbers are tested using methods defined by the ASTM, a US organization that establishes standards and practices used by acousticians and testing labs. By requiring its members to follow the same rules, materials tested to ASTM standards in different facilities can be compared with
confidence. However, a flaw in the test method does not take into account the edges of the material.
Although the edges are exposed when the material is tested, the calculation for absorption coefficient
considers only the size of the front surface, and ignores the edges completely. For a panel that is two
by four feet and four inches thick, the edges add 50 percent to the absorbing surface during testing,
yet they are ignored in the coefficient calculation. This is further complicated because there is no standard sample size. Since a small sample has proportionally more edge than a larger one, a sample
that's 8 by 8 feet will measure better than one that's 10 by 12 feet, even if they're the same thickness
and made of the exact same material.
In practice, multiple panels are placed adjacent to each other during testing, to minimize the contribution of the edges. So when 2 by 4 foot panels are tested, typically eight of them are arranged into a
larger square. But even when placed to form a single surface area of 8 by 8 feet, four-inch thick edges
still inflate the measurements by more than 16 percent.
This panel is 2 by 4 feet and 4 inches thick.
During testing the four edges add 50 percent to
the total surface area, yet they're excluded from
the absorption calculations. And when many
panels are mounted adjacent on a wall, the
edges are not absorbing even though they contributed to the published specs.
Further, most acoustic panels are meant to be installed adjacent on the wall in a cluster. In this case
the edges are not available to absorb even though
they were when the material was tested. When an entire wall is covered with four-inch thick panels none of
the edges are exposed, so the real absorption is only
2/3 what the published numbers indicate - and those
numbers were already inflated!
The same thing occurs with corner absorbers, as
shown in the figure below. Unless the vendor describes how these triangle shaped samples were
grouped during testing, there is no way to determine
how much of the stated absorption is due to
Foam blocks like this are meant to be mounted in a corner, stacked one
the edge effect and how much is due to its
above the other from floor to ceiling. When measured for absorption four of
effectiveness as an absorber.
the five surfaces are exposed, but when installed as intended only the front
surface absorbs. So in practice, a two-foot corner wedge like this provides
only 65 percent of the absorption claimed. The shorter the wedge, the larger
the disparity between the published and actual absorption.
58
For some products, like a tube trap, it is not practical to specify an absorption coefficient because
there is no front surface. In that case the correct way to specify absorption is in sabins, named for
acoustics pioneer W.C. Sabine (1868-1919). The sabin is an absolute measure of absorption, independent of surface area, and it can be used to compare any two absorbing devices directly and on
equal terms.
SIDEBAR: BIG WAVES, SMALL ROOMS
There is a common myth that small rooms cannot reproduce low frequencies because they are not
large enough for the waves to "develop" properly. While it is true that low frequencies have very long
wavelengths - for example, a 30 Hz wave is nearly 38 feet long - there is no physical reason such long
waves cannot exist within a room that is much smaller than that. What defines the dimensions of a
room are the wall spacing and floor-to-ceiling height. Sound waves generated within a room either
pass through the room boundaries, bounce off them, or are absorbed. In fact, all three of these often
apply. That is, when a sound wave strikes a wall some of its energy may be reflected, some may be
absorbed, and some may pass through to the outside.
When low frequencies are attenuated in a room, the cause is always canceling reflections. All that is
needed to allow low frequency waves to sound properly and with a uniform frequency response is to
remove or at least reduce the reflections. A popular argument is that low frequencies need the presence of a room mode that's low enough to "support" a given frequency. However, modes are not necessary for a wave to exist. As proof, any low frequency can be produced outdoors - and of course
there are no room modes outdoors!
Here's a good way to look at the issue: Imagine you set up a high quality loudspeaker outdoors, play
some low frequency tones, and then measure the frequency response five feet in front of the speaker.
In this case the measured frequency response outdoors will be exactly as flat as the loudspeaker.
Now wall in a small area, say 10x10x10 feet, using very thin paper, and measure the response again.
The low frequencies are still present in this "room" because the thin paper is transparent at low frequencies and they pass right through. Now, make the walls progressively heavier using thick paper,
then thin wood, then thicker wood, then sheet rock, and finally brick or cement. With each increase in
wall density, reflections will cause cancellations within the room at ever-lower frequencies as the walls
become massive enough to reflect the waves.
Therefore, it is reflections that cause acoustic interference, standing waves, and resonances, and
those are what reduce the level of low frequencies that are produced in a room. When the reflections
are reduced by applying bass traps, the frequency response within the room improves. And if all reflections were able to be removed, the response would be exactly as flat as if the walls did not exist at
all.
SIDEBAR: HARD FLOOR, SOFT CEILING
The following is from an exchange that took place in the rec.audio.pro newsgroup in May, 2003:
Bill Ruys asked: Why it is recommended to have bare (un-carpeted) floors in the studio? One web site
I visited mentioned that a bare floor was a prerequisite for the room design with diffusors and absorbers on the ceiling, but didn't say why. I'm trying to understand the principal, rather than following
blindly.
Paul Stamler: Carpet typically absorbs high frequencies and some midrange, but does nothing for
bass and lower midrange. Using carpet as an acoustic treatment, in most rooms, results in a room that
is dull and boomy. Most of the time you need a thicker absorber such as 4-inch or, better, 6-inch fiberglass, or acoustic tile, and you can't walk around on either of those. Hence the general recommendation that you avoid carpet on the floor and use broadband absorbers elsewhere.
59
Lee Liebner: the human ear is accustomed to determining spatial references from reflections off of
side walls and floor, and a low ceiling would only confuse the brain with more early reflections it doesn't need. Everywhere you go, the floor is always the same distance away from you, so it's a reference
that your brain can always relate to.
John Noll: Reasons for having wood floors: they look good, equipment can be rolled easily, spills can
be cleaned up easily, provide a bright sound if needed, sound can be deadened with area rugs.
Ethan Winer: In a studio room, versus a control room, a reflective floor is a great way to get a nice
sense of ambience when recording acoustic instruments. Notice I said reflective, not wood, since linoleum and other materials are less expensive than wood yet sound the same. When you record an
acoustic guitar or clarinet or whatever, slight reflections off the floor give the illusion of "being right
there in the room" on the recording. It's more difficult to use a ceiling for ambience - especially in a
typical home studio with low ceilings - because the mikes are too close to the ceiling when miking from
above. And that proximity creates comb filtering which can yield a hollow sound. So with a hard floor
surface you can get ambience, and with full absorption on the ceiling you can put the mike above the
instrument, very close to the ceiling, without getting comb filtering.
Dave Wallingford: I've always preferred wood floors for a few reasons: 1) It's easier to move stuff
around, 2) You can always get area rugs if you need them, And the main reason: 3) Pianos sound like
crap on carpet.
SIDEBAR: ROOM MODES AND MODECALC
60
Calculator (1.3 MB) that has a greatly improved interface and offers more information than the original
DOS version below.There's nothing to install - just Unzip the files into any folder, then run the modecalc.exe program file.
If you have a slow dial-up connection you can click HERE to download the original DOS version of
ModeCalc (only 56 KB).
ModeCalc runs on all Windows computers, and displays the axial modes for any rectangular room using dimensions you enter as either feet or meters. ModeCalc can help you design a new room that
sounds as good as possible, or predict the low frequency behavior of an existing room. The tutorial
below explains the basics of room modes, and tells how to use ModeCalc and interpret its results.
ModeCalc Tutorial
ModeCalc calculates and displays the first 16 axial modes, up to 500 Hz, for any rectangular room using dimensions you enter as either feet and inches or meters. It can help you design a new room that
sounds as good as possible, or predict the low frequency behavior of an existing room.
This tutorial explains the basics of room modes, and tells how to use ModeCalc and interpret its results. Please understand that ModeCalc is not meant to help you determine low frequency treatment
for an existing room. Regardless of what is predicted (or measured using test equipment) the solution
is always the same - as much broadband bass trapping as you can manage. Whether your current
room happens to have favorable dimensions or not is irrelevant, unless you're willing to move the
walls!
Room Modes
Room modes are natural resonances that occur in every enclosed space, and the frequency of each
resonance is directly related to the room's dimensions. For example, a room 16 feet long has a mode
at 35 Hz because walls that far apart provide a natural resonance at 35 Hz. Additional modes occur at
multiples of 35 Hz because those frequencies also resonate in the same space. Wall spacing that accommodates one cycle of a 35 Hz wave also fits two cycles of 70 Hz, three cycles of 105 Hz, and so
forth.
When you play a musical note having the same pitch as a natural resonance of the room, that note will
sound louder and have a longer decay time than other notes. Of course, this is undesirable because
some notes are emphasized more than others, and the longer decay times reduce clarity. Therefore,
room modes are important because they directly affect the character of a room. Although room resonances can be reduced by adding bass traps, they cannot be eliminated entirely. Top
For this reason, rooms for recording and playing music are designed to have many resonances that
are distributed evenly, rather than just a few resonances at the same or nearby frequencies. Playing
music in a room with poor mode distribution is like listening through a 5-band graphic equalizer with
one or two bands turned up all the way. A room with good modes is more like having a 31-band equalizer with all the bands turned up. The frequency response still isn't perfect, but all the small peaks
combine to yield an overall response that's reasonably flat. Therefore, the frequency response of a
room with many modes close together is flatter overall than a room that has fewer modes farther
apart.
Small rooms have modes that are spaced farther apart than large rooms because the first mode in a
small room starts at a higher frequency. For example, when the longest dimension of a room is only
10 feet, the modes for that dimension start at 56.5 Hz and are 56.5 Hz apart. In larger rooms the first
mode is at a lower frequency so subsequent modes are closer together. Therefore, a large room has a
flatter low frequency response because it has more modes spaced more closely.
61
The formula used by ModeCalc is extremely simple. For dimensions given in feet the first mode occurs at 1130 divided by twice the dimension. (1130 is the speed of sound in feet per second.) All subsequent modes are multiples of that result. When using meters the formula is 344 divided by twice the
dimension. Twice the dimension is used because a room 10 feet long really has a total distance of 20
feet - the wave travels from one end to the other and back to complete one cycle. So for a room 10
feet long the first mode occurs at 56.5 Hz:
The second mode for that dimension is two times 56.5 or 113 Hz, the third is three times 56.5 or 169.5
Hz, and so forth to the tenth mode at 565 Hz.
Room Ratios
The worst type of room shape is a perfect cube - say, ten feet long, ten feet wide, and ten feet high because all three dimensions are the same and all three dimensions resonate at the same frequency.
A 10 foot cube shaped room will have a strong inherent resonance near 55 Hz, which is the open A
string on a bass. So when that low A is played it will sound much louder than other notes. Such a
room also has a longer natural decay time at that pitch, so A notes will sustain longer and conflict with
other bass notes that follow.
A room whose dimensions are multiples of each other - for example, 10 feet by 20 feet - is nearly as
bad because many of the same frequencies are emphasized. Therefore, the goal is to have a room
shape that spreads the modes evenly throughout the low frequency range. This is done by designing
the room with dimensions whose ratios of length, width, and height are as unrelated as possible. And
here is where ModeCalc is useful because it tells you at what frequencies the modes occur and how
close together they are. ModeCalc also shows you the ratios of the dimensions you entered, and lets
you compare them to ratios commonly recommended by acoustic engineers and studio designers.
Using ModeCalc
Instructions at the bottom of the screen explain how to use the program. Simply enter the Length,
Width, and Height using the Tab and Shift-Tab keys to go between fields, then hit Enter to see the result. Up to 16 modes for each room dimension are displayed graphically so you can see where they
occur and how they relate to one another. Each set of modes is shown in a different color, and when
two or more modes occur near the same frequency the mode lines are drawn taller. Note that the
graphic display portion of ModeCalc uses logarithmic spacing. This is how octaves and musical intervals are arranged, and is also how mode spacing should be viewed.
You can also enter the room dimensions as meters and/or centimeters instead of feet by using m or M
or cm or CM and so forth after the dimension values. Values can be entered with or without spaces,
so '10m' and '10 m' both specify 10 meters. Likewise for feet and inches.
Many rooms are not rectangular, and in fact having angled walls or a vaulted ceiling is desirable. Unfortunately, with angles there is no direct way to determine the room modes exactly. Modes still exist they're just more difficult to predict. If the angles are not too severe you can average the dimensions.
For example, if the ceiling varies from 8 feet to 10 feet high, you can use 9 feet when entering the
height.
When viewing the results look for an even spacing of the modes regardless of their color (good), and
also look for multiple modes that occur at or near the same frequencies (bad). Also compare the ratios
of the dimensions you entered with the recommended ratios, and compare your room's volume with
the recommended minimum of approximately 2500 cubic feet or 70 cubic meters.
62
To make it easier for you to identify modes that are close together ModeCalc draws those mode lines
taller, which simulates the larger response peak that occurs. The normal line height is marked with a
thin gray horizontal line. When two modes are adjacent, or at least close, both lines are drawn taller.
The closer the modes are to each other, the taller the lines appear. This also lets you identify modes
that fall on identical frequencies. ModeCalc draws all the lines for one mode, then the next, and the
next. So if several modes are at identical frequencies one line will hide the other. If you notice that an
isolated line is higher than usual, that means there are at least two modes at that same frequency.
You can then use the Frequency Table display to see them all. Also note that modes naturally become
closer at higher frequencies. Therefore, having taller lines toward the right side of the graph is normal,
and does not mean your room will really have that rising frequency response.
Finally, if you are using ModeCalc to check an existing room, please don't be discouraged by poor results. All rooms need bass trapping anyway, and poor modes can be improved enormously by adding
a few more traps. You can also enter dimensions for a large room having one of the recommended
ratios, such as 23.3 by 16 by 10 feet. Then you'll see how even with the recommended ratios the
modes are still somewhat uneven, and two modes still sometimes occur at the same frequency. So
unless you are willing to move the walls, just accept what you have, and maybe install a few more
bass traps than you had planned for originally. Then relax and enjoy the music!
A Few Words About Mode Types
There are two basic types of room modes - axial and non-axial. Axial modes occur between two parallel surfaces, where non-axial modes take a more circuitous route travelling much like a cue ball going
around a pool table in a diamond pattern. The two non-axial mode types are called Tangential (the
reflected waves touch four room surfaces) and Oblique (the waves touch all six surfaces).
Every rectangular room has three axial modes, with one between the floor and ceiling, another between the left and right walls, and another between the front and rear walls. Axial modes are the most
important type simply because they are the strongest. They contribute more to peaks and nulls, and
modal ringing (extended decays), than non-axial modes. This is because the boundaries are parallel
and so the distance between reflections is shorter too. Therefore, axial modes are the most damaging
to accurate music reproduction, and the most important type to consider.
SIDEBAR: CREATING A REFLECTION FREE ZONE
WHAT: A useful goal for any room where music plays through loudspeakers is to create a Reflection
Free Zone (RFZ) at the listening position. The concept is very simple - to prevent "early reflections"
from obscuring the stereo image. This occurs when sound from the loudspeakers arrives at your ears
through two different paths - one direct and the other delayed after reflecting off a nearby wall.
Just as damaging is when sound from the left loudspeaker bounces off the right wall and arrives at the
right ear, and vice versa. Similarly, early reflections off the ceiling and floor can also harm clarity and
imaging. In all cases the reflections obscures fine detail and make it difficult to localize the source of
the sound or musical instrument.
This drawing, viewed from above, shows the three main paths by which sound from a loudspeaker
arrives at your ears. The direct sound is shown as black lines. The early reflections - a single bounce
off a nearby surface - are the red lines, and late echoes and ambience arrive as shown in blue. In
truth, the blue lines are much more complex and dense than the single path shown here, but this is
sufficient to explain the concept.
The general goal of a Reflection Free Zone is to eliminate the red early reflection paths by placing absorbing panels on the side walls in key locations. Not shown, but just as important to avoid, are early
reflections off the ceiling, floor, and mixing desk if present.
63
WHY: When a direct sound is accompanied by an echo that arrives
within 20 milliseconds or less, the ear is unable to distinguish the
echo as a separate sound source. So instead of sounding like an
echo or general room ambience, the sounds coming from different
directions combine, which obscures clarity and confuses the stereo
image. You can still tell when an instrument is panned all the way to
the left or right, but the in-between positions are not as well defined.
Put another way, listening to music in a Reflection Free Zone is similar to listening with headphones - musical instruments sound clearer,
and their placement in the stereo field is much better defined.
Another important reason to control early reflections with absorption
is to reduce comb filtering. This is a very specific type of frequency
response error that's caused when a source and its reflections combine in the air. Depending on the difference in arrival times, some
frequencies are boosted and others are reduced. The graph at left
shows the comb filter frequency response measured with and without MicroTraps at the first reflection points.
When the budget allows for dedicated construction, early reflections
can be avoided by angling the side walls and sloping the ceiling upward. The studio below was designed by noted studio designer Wes
Lachot, and offers a beautiful example of such construction. Given a large enough angle the reflections are directed behind the listening position without having to apply absorbing materials to the walls
or ceiling. This lets you better control the overall ambience in the room because you don't need additional absorption just to get rid of the reflections. But most people do not have the luxury of building
new walls, so the only option is to apply absorption at key locations.
64
HOW: The easiest way to tell where to place absorption to avoid early reflections is with a mirror. This
is the thin green object to the right of the listener against the wall in the drawing. While you sit in the
listening position, have a friend place a mirror flat against the side walls and move it around. Any location in which you can see either loudspeaker in the mirror should be covered with absorption. It's a
good idea to treat a larger area of the wall than you identify with the mirror, so you'll be free to move
around a little without leaving the Reflection Free Zone. Once the side wall locations are identified do
the same on the ceiling. Although it's more difficult to slide a mirror around on the ceiling, one way is
to attach a hand mirror to a broomstick May 4, 2004: Minor edits to the RFZ sidebar to make it clearer
that early reflections from either side wall are a problem, not just the left wall with the right speaker
and vice versa.
65
REVISIONS
March 5, 2003: Initial release.
March 16, 2003: The paragraph just above Figure 5 was clarified by Wes Lachot. Thanks Wes!
March 20, 2003: Added Figure 3b and accompanying text showing how to install rigid fiberglass in an unfinished basement ceiling. Also
added a new paragraph under Table 1 explaining the importance of density in absorbing materials.
March 22, 2003: Added a new last sentence to the second paragraph after Figure 7 offering a better reason panel bass traps must be sealed
air tight.
March 25, 2003: Added the sidebar The Numbers Game to expand on the role of edges when testing absorbing materials.
March 28, 2003: Changed the formula for calculating the ideal 1/4 wavelength air gap (below Figure 5) from 1100 to 1130 feet per second,
which is more accurate.
April 5, 2003: Added the sidebar Big Waves, Small Rooms to better debunk the myth that low frequencies require a minimum room size.
April 28, 2003: Added text (just below Figure 3b) describing using large bales of fluffy fiberglass as a bass trap.
May 14, 2003: Added the sidebar Hard Floor, Soft Ceiling to explain in more depth why that combination is better than a carpeted floor with a
standard drywall ceiling.
June 10, 2003: Added a link to my Graphical Room Mode calculator program in the text below Table 2.
June 23, 2003: Added three sentences to the end of the paragraph under Figure 6 explaining how the angle of incidence can affect the absorption of a fiberglass panel. Thanks to Eric Desart for bringing this to my attention.
June 26, 2003: Added, clarified, and edited throughout to remove what some have perceived as my bias against foam treatment.
July 12, 2003: Added the online manual for ModeCalc (see June 10 above) as a sidebar, so you don't have to download and run the program just to read the explanation of modes.
August 28, 2003: Added the sidebar Creating a Reflection Free Zone above.
November 23, 2003: Added two paragraphs about diffusors in the section Diffusors and Absorbers, directly under the photo of the wood
diffusor made by ACID acoustics.
January 14, 2004: Removed the qualification that the front wall angle must be "at least 35 degrees" in the RFZ sidebar because Wes Lachot
says the exact minimum angle depends on several factors. Also made a few other minor edits to clarify the text.
April 21, 2004: Added a few more manufacturers of rigid fiberglass, and explained that FRK stands for Foil Reinforced Kraft paper.
April 26, 2004: Clarified the difference between a deflector and a true diffusor. Also, clarified the importance of density in porous absorber
materials like rigid fiberglass.
May 4, 2004: Minor edits to the RFZ sidebar to make it clearer that early reflections from either side wall are a problem, not just the left wall
with the right speaker and vice versa.
May 12, 2004: Added text under the Room Symmetry heading explaining the importance of having loudspeakers fire the long way into the
room. It was a slow day so I also added several other minor enhancements and clarifications.
May 27, 2004: Replaced the graph I had made showing absorption versus air gap with a more compelling graph from Everest's Master
Handbook of Acoustics. Also added test data from fiberglass manufacturer Johns-Manville showing how increasing the density of rigid fiberglass enhances its performance at low frequencies.
December 30, 2004: Enhanced ModeCalc to allow entering feet and inches (12'6") rather than having to use decimal feet (12.5).
February 11, 2005: Added a link to my Density Report in the section about fiberglass density. Added a new section under Bass Traps Overview to explain modal ringing. Added a paragraph to enhance the explanation of flutter echo.
February 21, 2005: Numerous minor edits, additions, and clarifications.
August 19, 2005: Added a paragraph to explain that two rigid fiberglass panels 2-inch thick can be combined to work as well as one panel 4
inches thick. Also added a warning not to have FRK paper between the layers, or on both the front and rear.
September 21, 2005: Clarified the difference between fiberglass bass traps that act on velocity versus wood panel traps that absorb via
damping wave pressure.
February 3, 2006: Clarified that density is only one factor that affects a material's usefulness as an absorber, and that at some point the density can be too high.
March 9, 2006: Added text and a graph to the sidebar Creating an RFZ showing how comb filtering is an equally important problem that is
solved by placing absorption at the first reflection points.
66
June 18, 2006: Added text near the end of the section Fiberglass Bass Traps explaining how to treat the cavity above a hung ceiling.
October 12, 2006: Updated the ModeCalc sidebar to link to the new Windows version, and updated the text to match.
August 10, 2007: Updated Figure 4 with an improved version submitted by Markus Mehlau.
Ethan Winer is a reformed rock 'n' roll guitarist who sold his successful software business in 1992 at
the tender age of 43 to take up the cello. Ethan has, at various times, earned a living as a studio musician, computer programmer, audio engineer, composer/arranger, technical writer, and college instructor. He has had more than 70 feature articles published in computer and audio magazines including
Mix, PC Magazine, EQ, Electronic Musician, Audio Media, Computer Language, Microsoft Systems
Journal, IBM Exchange, Strings, Keyboard, Programmers Journal, The Strad, Pro Sound News,
prorec.com, Recording, and Sound On Sound. He now heads up RealTraps, manufacturer of high
performance acoustic treatment, and also hosts the EQ Magazine Acoustics Forum at the MusicPlayer web site.
In addition to technical writing, Ethan has produced two popular Master Class videos featuring renowned cellist Bernard Greenhouse, as well as five CDs for Music Minus One including a recording of
his own cello concerto. Besides writing and recording many pop tunes, Ethan has composed three
pieces for full orchestra, all of which have been performed. He lives in New Milford, Connecticut, with
his wife Elli and cat Bear, and plays in both the Danbury Symphony and the Danbury Community Orchestra where he serves as principal cellist. You can read more about Ethan's musical exploits on his
Music and Articles pages. Top
Entire contents Copyright © 2003-2008 by Ethan Winer and RealTraps, LLC. All rights reserved.
Ethan Winer
34 Cedar Vale Drive
New Milford, CT 06776
860-350-8188
ethanw@ethanwiner.com
Used by permission for Galcom International, August 2008.
67