ADDING A SUB
Whether you are adding a sub to an existing
2-channel "stereo" system, or building a "surround sound"
Home Theater system, let's examine every detail involved, including all
sorts of setup options and adjustments, subtle and not so subtle. Whether
you fancy yourself a pristine audiophile, a professional recording / mixing
engineer, a Sci-Fi movie buff, or a professional musician, you have your
own rainbow of views, and objectives, and they each have a "scale"
whether they are emotional, financial, audiophile, social, practical, experimental...
you have to "fit" your purchase(s) and adjustments into this range
of ideas you have.
Well, be prepared for a ride.
The ONLY correct way to add a sub to
system is to define everything ABOVE the sub's range as an entity; clearly
define the impulse, phase, and lastly frequency response of this entity;
and then make a new "2-way" system where the sub is one way and
everything above it is the 'other' way. The parts must be combined correctly
so that there are no cancellations and no smearing of time-related musical
This CANNOT be easily measured in the
frequency domain, because you could have (as an example) an 80 Hz signal
coming from both the mains and the sub, and if the sub is 12.5 msec late
the two sources will "seem" to be in phase but the sub really
will be 360 degrees late. It is the impulse smearing that this affects,
but people don't measure that because there is no simple "hand held"
phase or impulse meter as there is an SPL meter. The REASON this meter does
not and essentially can not exist is in order to measure impulse response
or phase response you need a starting REFERENCE point, ( in time) and in
a speaker system since the signal has to travel through circuitry, amplifiers,
crossovers inside the speaker box and then hit the driver; therefore the
first reference point MUST be acoustic. There ARE computer based impulse
response systems such as the TEF, (quick history HERE;
very quick technical blurb HERE;
full story HERE
) but they are involved, require real instrumentation, are expensive, have
a seriously steep learning curve, and they are absolutely not the kind of
thing most 'consumers' can be bothered with.
So the overall view of adding a sub
is this: In essence you are designing and assembling a new speaker
system which is "2-way": the sub is one way and everything
"else" above it in frequency is the 2nd way.
Simply connecting a sub to existing
mains speaker (or amp) terminals is the WORST POSSIBLE WAY to do this. EVERYTHING
scientific and acoustic about this method is wrong, from the additive delay
issues to the back EMF of the mains affecting the LF signal. However there
are plenty of people who simply do not understand correctly integrated bass,
and they will be reasonably happy simply sticking another box on to their
system without regard to timing, phase and frequency issues, and they will
think it sounds "ok" or "good" and for those people
it doesn't really matter.
Indeed the only thing that does
matter is an individual's happiness with their system, whether I or anyone
else thinks it's right or wrong.
But to get purely technical...
There are a separate set of issues for
2-channel stereo "audiophile" and Home Theater systems, which
we may call "Surround Sound", i.e., 5.1, 6.1, 7.1 etc. systems.
Later you may want to read my Surround Sound page here.
With Home Theater, there is relatively
little correlation between the LFE channel (the true Low Freq Effects channel)
contained on the DVD and all the frequencies ABOVE 80; all that is really
essential for the low frequency part of movie enjoyment is the best coupling
of the below 80 Hz effects to the room and indeed the listening position.
There is ALSO the rest of the bass; all the below 80 Hz information from
all 5 channels that is stripped off and summed together into mono and sent
out the SUBWOOFER OUT connector on every modern HT receiver. That plus the
LFE channel (if it in fact exists on that particular DVD, and it may not)
constitutes all the BASS MANAGED BASS. Therefore the MOST desirable scenario
in a HT situation is to best couple the sub(s) to the room FIRST, and THEN
timing and phase match the sub(s) to the rest of the system.
But let's back up one step. DO NOT CONFUSE
LFE with BASS. LFE is a separate channel in a movie theater
(called the 'boom' track in the industry) which is necessary because there
is not enough dynamic range (headroom, actually) in the existing film optical
sound tracks and their associated playback hardware for additional
"Low Frequency Effects". In a movie theater, (as will further
be explained below) you CAN have multiple low frequency sources because
there are essentially no standing waves of any consequence in that size
Impulse response is the holy grail of
all of audio. With more pristine 2-channel sound, (and when you are playing
music through your Home Theater system) as we approach, want, or expect
audiophile quality, the issue is to get the IMPULSE RESPONSE through the
crossover region (and therefore both the phase response AND frequency response,
which is contained under the mathematical umbrella of impulse response)
as smooth as possible, so that IF we were playing back a correctly recorded
IMPULSE, for example a well recorded kick drum, its fundamental (50-60 hz),
and its subharmonic, an octave lower (25-30 Hz) and its mostly odd order
harmonic structure (all the way up to 8 kHz and then some) are presented
correctly by the time they arrive at the acoustic summation point which
is your ears. This is the basis of "high fidelity".
We also have to assume and this
is a huge assumption that the manufacturers of our "mains"
speakers have ALREADY correctly addressed the issues of both impulse response
and frequency response. So for the purposes of this discussion (my entire
book isn't ready yet) we will assume that whatever your mains are, from
a 2-way bookshelf to an 8 foot tall floorstander monster, that within the
desired passband of the mains, the impulse response and frequency response
are already handled.
the subject of absolute polarity. This has no phase relationship
to anything other than ITSELF. Imagine you are standing in front of a nice,
large, beautifully tuned drum kit. The drummer obliges us and plays just
the KICK drum, perhaps loudly and once every second. So the pedal is a mechanical
impulse hammer device which hits the skin on the drummer's side; this pressurizes
the air in the drum, and the front skin moves forward.
That's an IMPULSE. It's actually the
leading edge of a square wave, with a little slope to it. A square wave
by definition has a fundamental and only odd harmonics. A sine wave has
only it's fundamental frequency, and a triangle wave is the fundamental
and only even-order harmonics. So the impulse of a kick drum is nearly a
square wave, with some sine wave fundamental and some even order harmonics,
but less than the odd order harmonics present in the square wave part.
The net human result, since you are
standing or sitting in front of the drum, is you feel and hear this positive
pressure wave, and your ears, body, intellect, social acuity, and previous
memories of such things all converge and you "hear" this phenomena
as a kick drum hit. You see it; you hear it, you recognize it, and it fits
your preconceived notions about what a kick drum should sound like. In theory,
this sound is then picked up by a microphone. Positive [air] pressure on
the diaphragm of the mic produces a positive-going (+) voltage at pin 2
(of the 3-pin connector); then this goes into a mic preamp, the rest of
the line amplification, and at some point in the control room of the studio,
out to a monitor amp and then a loudspeaker. If all goes well, we then stand
in front of that speaker, and listening to the monitor system, we are socially
convinced there is a drummer obliging us by playing a kick drum right in
front of our face. If the absolute polarity of that impulse is "backwards"
i.e. the polarity of anything in the circuit is changed, such as
the monitor speaker is wired out of polarity then the absolute polarity
is not the same as the original and we can hear that. This is one
instance where this phenomena is very easy to both set up as a test and
easy to discern. Clark Johnson has written a entire book about this called
The Wood Effect, available HERE.
Imagine we are playing back a well recorded
cello: we have the fundamentals of both the strings and the wood STARTING
in the subwoofer (that means below 80 Hz, and you may wish to refer to my
frequency-wavelength chart here: www.soundoctor.com/freq.htm
) and the harmonics extending smoothly up through the various drivers in
the rest of the system. Being a recording of actual "wood", (and
the strings!!!) these harmonics are mostly even order. If we can correctly
preserve the exact timing (and therefore phase) relationships of the ratios
of the harmonics of these signals, we will preserve the imaging "in
space" of this instrument. If we do not do this, then the focus is
lost. One part of this assumption is that the instrument is correctly recorded
in the first place, ideally with a stereo pair of microphones which therefore
ARE picking up the 3-dimensional phase and harmonic structure of the instrument
You CANNOT have multiple low frequency
sources of differing phase relationships in a living room-sized room. Let's
examine the acoustic "spaces" we might be dealing with. There
are 3 useful separate sizes of acoustic spaces in life:
1) The inside of a car,
where you are essentially living inside the speaker cabinet. (the pressure
2) A large movie theater,
amphitheater, or outdoor space where there either are no reflecting
walls or the walls are so far away in time that any reflections,
partially because of the Haas effect and frequency cancellation effects
are essentially of no importance; and...
3) The inside of a typical
living room / home theater room. In this size room you will ALWAYS have
standing wave issues somewhere in the bass passband from 20 -125 Hz. You
CANNOT NOT have these issues in a room this size unless
you have a REAL acoustically treated room with full size, perhaps 32' bass
traps in the walls and all the correct ratios of absorption vs diffusion
especially at low freqs. This does not mean a couple of pillows in the corners
or ineffectual 400 Hz absorbers on the side walls. IF you were to have a
room with REAL bass trapping then there would be no bass standing waves
because the LF signals hitting the walls would be absorbed before they had
a chance to bounce back. (what a concept!!!) Rooms like this are a revelation,
(not to mention extremely rare) because for the first time you are actually
able to hear the speaker, and not the speaker "in" the room.
But back to "most rooms"...
If you have 2 LF sources of differing
phase relationships and/or timing relationships they will cancel. Period.
And if they are "in phase", but 1, 2, 3 or more full cycles (that
means wavelengths) shifted, (that means 360 or 720 degrees out of phase)
then the overall frequency response will not seem bad but the impulse
response and clarity and focus will be smeared, and localization and imaging
will be lost. This is the main reason measuring in the frequency domain
especially in a home-sized room is such an incredible waste of time. Your
measurements "seem" pretty flat and yet you don't like the end
result - isn't as "clear" as you think it should be, and it isn't
as focused as you think it should be. The issue is ONLY timing.
We can call the red and green waves
signals from 2 separate "speakers", 2 separate subs, or a sub
and a "mains" speaker. Here are the diagrams that show this:
Fig 1. Obviously "in phase"
Fig 2. 90 degrees "out of phase" (the red
wave is lagging the green wave by 90 degrees)
Fig 3. 180 degrees out of phase (the net result is
Fig 4. ...an
example of group delay. This only shows one cycle of many.
It's entirely possible the signals are overlaid so they look
like they are "in phase" but they are actually 360
degrees (one wavelength or cycle), 720 degrees (two wavelengths or
cycles), or 1080 degrees etc. shifted in time out of phase.
Fig 5. Group
delay drawn another way. The GREEN
wave might be coming out of your "mains". The RED
wave is coming out of your sub. Notice how at first they "look"
as if they are "in phase" but the red wave (from the sub)
is actually a full wavelength LATE.
How did the sub get to be 360 or more
degrees late? It's the overall physics of how it's built. The only correct
way to implement a sub so the frequency response and phase response can
be controlled and have it socially acceptable in a living
room is to implement a sealed box design, and that means EQ circuitry. Also
most of the better brands of subs, JL Audio included, use massive drivers
which have a relatively large X-max (that means cone movement). The combination
of the air pressure in the sealed box, and the rest of the equalization
circuitry necessary equal a mechanical and electronic phenomena which equals
an overall time or group delay. Therefore IF the sub is already 8-10 msec
late, AND it is placed in the corners further away than the mains (just
for example) then relative to the mains it might be 12-16 msec late. YOU
CANNOT TAKE THIS DELAY AWAY.
You might enjoy referring to my handy
FREQUENCY-WAVELENGTH-PERIOD chart here: www.soundoctor.com/freq.htm
TYPES OF "MAIN" SPEAKERS
In addition to all the above, there
is the complex issue of the "main" speaker you are coupling to.
There are essentially 6 types of speakers that exist:
2) port in the front
3) port in the bottom
4) port in the back
5) a dipole, which is a flat panel such as an electrostatic (Sound
Lab, Magnepan, Quad, Beveridge, Martin Logan, etc.)
6) an true omnidirectional system such as the MBL
or the BEOLAB
Each of these speaker types couples
somewhat differently to the room, and certainly to a sub.
A port is ALWAYS nothing more than a
cheap way to attempt to get free bass out of an enclosure and /or driver
that's too small. It's a holdover from the 1930's when because of driver
inefficiencies (especially when compared to today's units) you had to do
everything possible to increase the useable output over the desired range
At one level, all the guyz want 9 foot
speakers in the living room (read "man-cave"). All spouses, of
whatever gender, want tiny 3" speaker cubes that disappear, but expect
9-foot results from them. Since they haven't repealed ohms law or any other
laws of physics while we were sleeping, the only way to get correct sound
is to move a correct amount of air.
Lets examine ported speakers. We'll
start with the worst case, the port in the front. At mid bass frequencies,
say 50-80 Hz, the LF driver moves IN the cabinet, the air in the cabinet
is elastic, and the port air moves out of the cabinet. Because of the frequency
at which the cone is moving, by the time the cone moves back out again,
the port air is now moving out, so in front of the cabinet the two air pressure
sources sum together and you get a fake bass "bump" or "boost".
As you go
lower and lower in frequency, at some low frequency the air pressure from
the LF driver and the air pressure from the port are exactly opposite each
other, so they cancel, and there is no more audio at that frequency: it
disappears. This defines the -3dB "cutoff" point of the cabinet
in question. When the manufacturer of a speaker cabinet defines the frequency
response (i.e., 37 Hz - 20kHz +/- 4dB) this is what is defined by the entire
arrangement of the port and the air in the cabinet and the driver.
You must understand that ANY driver
goes down to 0 Hz, or DC. If you put a battery across a speaker, the cone
moves out and stays there. If you were to have a DC coupled power amp feeding
a speaker - ANY speaker, from a 1" dome tweeter to an 18" rock
n roll stage bass driver - and you put 4 Hz into it, it would simply move
back and forth at 4 Hz. Of course in order to actually "hear"
the audio it would have to be in the generally accepted passband of 20-20,000
Hz and the cone diameter would have to be enough to actually move some air
in the room. So it is the overall combination of the driver size, the excursion,
the box size, (therefore the air back pressure) and many other factors that
determines the overall response of that "speaker" AS AN ENTITY.
That means IF you were to simply put
those same frequencies through the mains and the sub (that means with no
crossover, and this is the mistake that nearly everyone makes) you would
now have 3 sources of LF energy and differing phase: the 'main' LF driver,
the port, and the sub, all fighting with each other. A further corollary
is that since the air inside the [mains] cabinet is elastic, the phase relationship
of the port air to the driver air is also a sliding one; that means it's
"out of phase" and smearing over a wider range of
frequencies than you might think.
If the port is on the back, again a
cheap attempt to use the back wave bouncing off a wall to give 'additional'
bass, you have the ADDITIONAL issue of the transit time it takes for the
back port pressure (already delayed because of the elasticity) to leave
the cabinet, travel back, hit a wall, and bounce back around the front of
the cabinet again; therefore this LF wave MIGHT be "in phase"
with the front driver BUT BE 360 OR EVEN 720 DEGREES LATE; therefore it
sounds like the bass frequencies are ok in the frequency domain but the
IMPULSE RESPONSE is now muddied.
Also, in the case of back
ported or (type 5) dipole speakers, since the path length from
the back of the speaker to the wall and bouncing back around to
the front of the speaker is a fixed physical entity, at some frequencies
you are adding and at some frequencies you are canceling: you
have simply made a physical/mechanical frequency comb filter that
you can't do anything about. Sound Lab's answer to this (for use
with their flat panel electrostatic speakers, which are dipoles)
is they sell you a "Sallie",
which is an absorber to absorb the entire back wave output of
the electrostatic panel. Since now there is no comb filtering;
all you are therefore hearing is the front signal.
A ported sub for home use is even more
wrong than ported mains. Now you would be attempting to acoustically add
together in the room at least SIX low frequency sources with differing phase
and frequency slope conditions: the LF drivers in your two mains, their
ports, the sub driver, and its port. In addition, since it's a bandpass
it cannot go down low enough for serious Home Theater effects.
In some cases
such as a bandpass sub used in a club or on stage, you are most concerned
with efficiency and directionality and not with getting frequency response
"flat" down to 20 Hz; therefore a correctly set up bandpass box
that might roll off at 34 to 40 Hz is quite sufficient and also very efficient
for the defined purpose. And again, as a point of reference, "flat"
response in the frequency domain is FAR AND AWAY the LEAST important phenomena:
impulse response in the time domain is the most important, but it cannot
be measured with a handheld meter therefore almost everyone simply ignores
it. If you're interested in learning about the newest pro sound system /
stage methods of "steering" bass, Dave Rat has some very cool
part 1 www.youtube.com/watch?v=VwLH7zP6Lwo
part 2 www.youtube.com/watch?v=B-3pURYOwfw
part 3 www.youtube.com/watch?v=aSZK9Altvm8
But back to our Home / HI-FI / 2-channel
/ Audiophile / Surround Sound systems: There is ONLY ONE truly correct way
to "add a sub" to a system in an controlled listening room situation:
you must correctly cross over the 2 sealed cabinets; and their timing must
be correct. ANY other method will lessen the focus and clarity you have
tried so hard to preserve.
I have many clients and customers with
extremely exotic high-end 2-channel systems that are all chasing the holy
grail of 3D holographic sound imaging, and until they follow my distinct
guidelines they are never completely satisfied with the results.
A similar situation exists with home
theater setups where the customer THINKS that the front speakers are "full
range". Even so, the BEST approach is to seal the ports, operate the
5 channels as "small", crossover at 80 (or even a bit higher,
but NEVER lower) and correct the timing issues inherent in all modern
subs by setting (in the receiver or processor's setup menu) ALL the
distances THE SAME, and to a small number such as 7 feet; then set the sub
to 12 feet MORE (i.e. 19 feet) and THEN use the variable phase control on
the sub to fine tune the relationship at the 80 Hz crossover point. Some
better speaker companies that make "large" speakers (such as B&W)
are aware of this port issue and supply port plugs just for this purpose.
Kudos to them.
People who have fought with their systems
for weeks or years finally email and call me to tell me that for the first
time they are finally satisfied in fact thrilled with the
incredible integration of their JL Audio subs.
All of this discussion barely scratches
the surface of the true complexity involved in flawless integration, so
THE RECORDING PROCESS
On top of
all the previous variables we have all the issues and errors inherent in
the recording process. It is simply laughable (and pathetic) when I read
the magazine articles where the "soundstage" of a rock recording
is "palpable". Sorry, but every modern rock recording made in
the last 40 years is composed of a series of panned mono signals that have
absolutely no "depth". They are each separately sent to an echo/reverb
device, the returns of which are usually (but not always) panned full left
and right. The summation of all the L-R panning placement and the summation
of all the reverb returns fools you into thinking there is a "soundstage".
Sadly, precious few recordings are made with any regard to true stereo or
binaural sound in anything resembling a true form; even better classical
recordings of large orchestras have morphed into combinations of stereo
miking and "some" local more-nearfield mono miking added to the
mix to achieve whatever the producer determines is a suitable balance, perhaps
between a soloist and the rest of the orchestra. Yes, there are precious
few companies who do pay attention to this; AIX
records is one. But to think that any modern, commercial pop recording
mix has any true acoustic space is, for the most part, sadly mistaken.
Oh, and to touch upon "stereo bass"
for a moment... there almost is no such thing. Going back to vinyl,
every stereo record cut in the last 60 years has mono bass. It has to. If
the bass were 180 degrees out of phase L and R then there would be vertical
modulation and the stylus would jump out of the groove. Therefore every
cutting lathe on the planet has a "compatalyzer" circuit, that
dumps frequencies below 160 hz into mono (typically a single-order filter,
therefore 6dB/octave). You MAY have out of phase bass (i.e. "low
frequencies") on a CD, but precious few producers/engineers are savvy
enough (or care enough to even bother, since what's the point?) to make
use of those sort of tricks. There are some trance / psychedelia / electronica
dance music releases where there are bass tracks where there is stereo bass
in the form of something like 24 Hz in one channel and 24.2 Hz in the other
channel; therefore you get an air pressure differential which travels around
the room. Cool! In the above example, the "traveling wave"
would take 5 seconds to go back and forth around the room. If you're a really
bored or obsessive techweenie you can have a lot of fun with this - we played
with this phenomena at Moog Synthesizer as far back as 1969. As far as PLAYING
BACK signals like this goes, as mentioned above, in a large theater or outdoors
you CAN have multiple bass sources of differing phase because there are
no standing waves, and so your ears (and indeed your whole receptive system)
can process and differentiate all the phase issues. In a much smaller room
like a living room, it is more difficult but you might be able to pull it
off if your subs were more nearfield (the pressure zone). If you invent
something new, let me know. Bass is fun!
ABOUT USING TWO OR MORE
|Fig 6. A symmetrical
|Fig 7. An asymmetrical
So most people's reasons for multiple
subs in a room is "more even coverage". Let's examine the instances
of multiple subs and what they do. One interesting issue with using multiple
subs concerns arrival times. Here's a hypothetical situation. You are feeding
the same signal to 2 subs. So this begs the question what's your room like?
Is it symmetrical? L-shaped? A closed room? A Huge open space? Notice we're
back to acoustics?
Referring to Fig 6, the subs are also
equidistant from your body. So the subs each couple to the room however
they do. The whole setup is essentially symmetrical.
Now here's another example. Refer to
Fig 7. We are still putting essentially the same signal into both subs.
There might be 3 ways to do this:
1) from the L and R of a stereo preamp (MUCH more on this later)
2) using a "Y" cord from the BASS MANAGED output from a Home Theater
3) In the case of the JL Audio sub, it might be a Master/Slave setup.
The point is, the sound leaves
both subs at exactly the same time. Notice in Fig 7 the R sub
is closer to your face. Perhaps the L sub is 11 feet away and
the R sub is 4 feet away. That's a 7msec differential. So you
hear the leading edge of the bass wave from the Right sub, then
7 msec later the leading edge of the L sub... then the note dies
away from the R sub and then 7 msec later the note dies away from
the L sub. What have you accomplished? Here comes the magic:
YOU HAVE FATTENED UP THE LOUDNESS ENVELOPE IN TIME ! This
is the magic that humans love. This is why someone says, "OMG,
two subs are SO much better than one!" So you have a combination
of the arrival time differential, and to a certain extent you
have the separate room coupling issues such that each sub is its
own entity coupling into the room with slightly differing standing
So now we have 2 ways to view the multiple
sub issue: as a method of attempting to get better coverage over a larger
seating area of a multiple-seat Home Theater room, or as a method of fattening
up the bass presence for one or two listeners in a sweet spot.
In the case of better subs, that have
variable phase adjustments, my suggestion in setups like this is to use
either method (1) or (2) above, and then adjust the phase control knob on
each sub for most accurate transitioning at the crossover frequency.
It's slightly tricky, but you will keep the real phase relationship between
each sub and the mains, AND you will keep the arrival time differential
that you "love".
Some people think that "bass
is non-directional". That is a mis-statement. The reality
is that as you go lower and lower it becomes less localizable
by your mechanism of hearing; above about 100 Hz you can start
to localize it and the precision of the localization depends on
the rest of the frequencies playing (or not); and the standing
waves in the room at the frequency you are trying to determine.
Feeding two subs with the same sine wave from a test oscillator
or test cd and adjusting the phase knobs separately will show
you just how directional it can be. It can be steered with surprising
precision, and in my years of night club building we used to adjust
the steering of arrayed subs so that the bass was correct on the
dance floor and much less off the dance floor in the corners
of the club.
Understand that the largest percentage
of all audio issues is room acoustics. You cannot put a great speaker in
a marble shower stall and expect it to sound good. Room acoustics itself
is a complex set of interactions of physics and perception. Sadly, there
are many instances where manufacturers or individuals skew the relevant
terms and confuse people. For example, beware of (and be aware of) the dangerous
term "room tuning". You CANNOT tune a room using an "equalizer".
You are tuning THE SOUND SYSTEM with the equalizer - the room is still the
same. REAL room tuning means anything from sticking pillows in the corners
to rebuilding the room (perhaps correctly) from scratch, incorporating a
set of acoustic devices and parameters which sometimes seem nebulous but
get a desired result. Because of the nebulosity of all the acoustics
terminology (not to mention the international differences, which are substantial)
it is often difficult for an end user (and many audio professionals, for
that matter) to be able to mentally visualize just what a room without standing
waves will sound like, or a room which is so rolled off that the high frequencies
seem to "fall to the floor". To make matters even worse, a term
like "soundproofing" is essentially an audio non sequitur; you
would have to define how many dB, and at what frequencies... and what is
the ambient noise level of the area of interest? And so on. So real room
tuning is one entire entity, and then once the room is deemed to be as useable
as it's going to get, then we enter the realm of SYSTEM TUNING. The big
trick of course is getting the correct balance of all of these items in
a row, so you have an end result you like.
Some people say they are going to put
a sub in the corner because of "room gain". Another misnomer!
There is no gain; there is no amplifier attached to the room!
What is taking place is the corner of a room has the MOST EFFICIENT coupling
at the lowest frequencies because the 2 walls and the floor are acting like
3 sides of a horn at those large wavelengths. So it's not that the corner
has any gain; it's that everywhere else in the room has apparent
loss. The middle of each wall has the most apparent loss, because the
sound leaves the driver, goes in all directions, folds back on itself and
cancels out. If you put the sub in the middle of a wall left and right and
ALSO placed it in the middle of the wall floor-to-ceiling you would get
NO bass in the room.
So you have the 3 options for sub placement:
1) Wherever you can, or wherever
your spouse tells you to put it
2) Where it's mathematically correct
to couple to the most applicable part of the room
3) By doing the crawl-around test
and matching up the sub coupling into the room inversely the best to your
For some very enlightening
articles about bass, room modes/nodes, standing waves, and room
coupling, see Art Noxon's articles HERE .
And for an in-depth listing of acousticians, acoustic materials,
design/build companies, and so on, see my links list HERE.
So AFTER you have addressed the issue
of room acoustics to the best of your ability, and this means you have decided
if you have a 2-channel system, a Home Theater system, (perhaps both, even
perhaps separate!) what your seating priorities might be, and the rest of
your decor, you might have decided to make the sub placement a priority.
Or not. IF YOU ARE ABLE, here is generally the best method: THE CRAWL-AROUND
TEST. While it might seem funny or silly the end result compared to hours
or days of computer analysis is usually spectacular.
The methodology is outlined on my test
CD page, here: www.soundoctor.com/testcd.
The crawl-around test has nothing to do with the rest of your system.
What you are doing is coupling one or more subs back to your listening position
based on the physics of the room. AFTER you have finished the test, you
THEN match the subs with the rest of your system in the frequency crossover
mode, and in phase and absolute timing mode.
If you DON'T couple the sub(s) to your
listening position or area as well as they might be, you could be throwing
away "a few" dB in coupling efficiency. If you are "throwing
away" 3dB PER SUB you might as well not have bought the 2nd sub in
the first place. Remember 3dB is twice the power, and 6dB is four times
the power. Most people who are NOT used to audio tend to equate 10dB (10
times the power) as "twice as loud", while engineers who are all
too familiar with the financial issues of trying to make something louder
have learned that 6dB is, in fact, twice (or half) the loudness, or Sound
Pressure Level. Actually there IS NO SUCH THING as "twice as loud".
Your brain and senses operate on a 20 log scale, and you should learn how
that equates to real life. It's fun.
But back to reality: there is a place
in life for subs connected almost ANY way, where there's just another extra
bass boom which impresses some people. To someone who only has experienced
a cheap table radio or a the moral equivalent in any sort of surround system,
ANY sub, even one poorly set up will "seem" like a revelation.
AUDIOPHILES? COMPUTERIZED ROOM SETUP?
Do not fall into the trap of having
a home theater receiver / processor with a "computer" inside and
your JL Audio sub with it's ARO inside and think you are going to run these
two computers and your life is gwanna be great: you might be in for a rude
awakening. You will more than likely be like a person with 2 watches who
is never really sure exactly what time it is...
Until there is a real holographic computer
system which really can sample the room in a 3 dimensional grid (for example
in 36 or 48 places) the best we can do right now it to attempt to approximate
the net results in a room at a few (1, 2, 3, or 4) places. In SOME setups
like this the results can be great. But here is where it sometimes falls
apart: If the room is so bad that you really "need" a setup computer
in the first place, it can't necessarily determine what is real,
what is reflection, what is standing waves, and so on, and it simply won't
work as you expect. Imagine trying to adjust a sound system in the aforementioned
marble shower stall. You cannot fix or change the room reverberation or
standing waves no matter what you do with a computer. Someday there will
probably be computers powerful enough to do subtractive room decorrelation,
and they will probably work by scanning the room with laser interferometers
first, then build a 4 dimensional graphic of the room, (by then probably
in n-dimensional space, but I digress) then correlate all the standing waves
at all frequencies, calculate all the Rt60 times at all frequencies, then
adjust the output of all the amps to decorrelate all this... (hear that,
Darpa?) but don't hold your breath.
My suggestion is to follow the necessary
steps separately AND MANUALLY, and in the correct order; learn
the equipment, and then experiment with ONE "computer" at a time
(I would suggest the JL Audio ARO first) and determine if it helps you.
If not, try something else. The only way you can determine if something
works is to make one change at a time. Remember, the JL Audio ARO does NOT
correct issues in the time domain. It only attempts to correct one frequency
anomaly and flatten that out if it finds one. You should make every attempt
to correct the timing FIRST. And if you have two JL Audio subs (or more)
and have followed the rest of the procedures, then I suggest running each
ARO separately, because the test microphone will be listening to that sub
ONLY coupling into the room.
Some people incorrectly use
a Y cords to feed both inputs of a sub. This is or should be completely
unnecessary; all it does is the same thing as turning up the level
6dB. And if you happen to have TWO subs you should actually wind
up turning each one down 3dB, so you wind up with the correct
level in the room and you have gained 3 dB of HEADROOM
in each sub. If you were to leave each volume at its reference
level you might find that it's easier to turn DOWN the SUBWOOFER
LEVEL in the setup menu of your Home Theater receiver/processor.
ELF TRIM and BOUNDARY SETTINGS
On the JL audio subs, the
ELF trim is an equalizer operating in the 25 Hz region which can
compensate for the [apparent] bass buildup if you are placing
the sub in the corner. (See the paragraph on room acoustics, above)
Typically IF you placed the sub in the corner you might want to
turn the control down. If for some reason you place the sub at
the middle of a wall or in another less than desirable position,
you can add 3dB. Remember 3dB is twice the power!
Some receivers/processors have THX and
other settings for "boundary" effects, and these are similar to
the ELF trim on the JL Audio subs.
A further discussion includes crossovers,
whether passive, active, tube, solid state, analog, digital, balanced or
unbalanced; and proper methodology of both measuring and correcting the
inherent group delays in modern equipment to fine tune the impulse response.
We're getting to that !
ABOUT GROUP DELAY AND IMPULSE TIMING
So now let's examine the aforementioned
group delay. It takes time for a signal to go through a circuit. Inasmuch
as everyone thinks electricity travels at the speed of light, that's not
quite true. Electrons going through a wire, which we can call a transmission
line are slowed down by a certain amount. For some types of cables this
is called the velocity factor, and it's typically 66% of the speed of light.
It also takes a certain amount of time for the signals to get through each
piece of equipment, although relative to other human events, this is quite
fast: it might take 5-50 microseconds for the signal to go through a power
amp, because there are no mechanical devices in the way. Once we get a signal
into a mechanical device such as a speaker, whether it is passive or active,
we now have the sum total of all the electrical plus mechanical phenomena
to take into account. The typical group delay through a modern, sealed box
subwoofer, is perhaps 8 to 15 msec. That's milliseconds, not microseconds.
In the digital world, delay issues are
often called latency. Specifically this refers to some circuitry
where the signal starts out as analog, goes through an A:D converter (not
an A/D converter as incorrectly stated in much literature; it's all math
and it's a ratio, not a division... but I digress even further...) then
gets processed digitally in some fashion, then goes through a D:A converter,
and then we hear it as an analog signal. This is a HUGE issue with modern
recording studios and live "digital" mixing boards and everyone
is continually fighting against seemingly impossible odds...sometimes there
is so much latency when devices are used in series with each other that
the musicians hear themselves as an echo and this makes it nearly impossible
to play. The entire premise of the "convenience" and "power"
of "digital" is sometimes negated by these latency issues and
the difficulties in "fixing" them.
This is also an issue inside Home Theater
receivers/processors, where the purely digital HDMI signal is stripped apart
and reconverted back to analog. Collectively, this mess is responsible for
instances where the picture and sound are "out of sync" in modern
equipment. Since you CAN'T get rid of the delay, the only answer is to delay
something else so it all "matches up" in the end. In the
analog world it still takes time for a signal to go through a circuit, and
although the phenomena should probably be called transit time,
group delay is what has stuck; a holdover from the early telephony days,
when the concern was the delay of the audio frequencies, not the DC control
or bell ringing signals (all carried on the same lines), and the term meant
a "group" of frequencies we were concerned about.
Let's start with a 2-channel (stereo)
setup and look at this block diagram:
|Fig 8. The same signal
applied to both the main power amp and the sub are delayed going through
As shown, the delay of the sub would be 1 wavelength at 80 Hz, or 12.5
Fig 8. shows THE INCORRECT METHOD many
people use when connecting a sub. It pains me to even have to use this diagram.
NO crossover is shown. The full range signal goes through the power amp
and into the mains; and the full range signal goes into the sub, where the
sub's own LOW PASS / HIGH CUT filter is engaged.
Here's the clincher: since the sub is
always at least 8, 9, 10, 11 msec late, the phase relationship CAN
NEVER be correct. It can be corrected in one of 2 ways only: you
can use some electronic means to ADD the same amount of delay to the top
(mains); or you can move the sub(s) closer to your body the correct number
of msec. You CANNOT match the phase of the sub to the mains because you
CANNOT use the phase control on any sub to remove delay; you can ONLY ADD
Crossovers are always a slippery issue.
Many 'audiophile' dealers don't necessarily sell them because (go ahead:
squirm) they don't really understand them, and they require a lot
of handholding therefore they can't make any money on them... and most speaker
manufacturers won't admit or suggest that their speakers need a sub because
they don't (or may not) make a sub; therefore they port their speakers in
an attempt to get extra "free" bass and therefore the coupling
and delay timing issue is made ever so much more complicated. Many customers
that I talk to simply buy a sub (or two) parallel ("Y") the output
of their preamp into the main amp and the sub, and are then unhappy with
the results. They think that because their speakers go down
to 38 Hz or 32 Hz or 27 Hz that they ONLY want to use the sub between 20
and 32 Hz... it simply doesn't work like that, because of the incorrect
port, and the fact that the sub is simply not matched to the mains. The
results are muddy, indistinct bass, and users who incorrectly attempt this
setup often blame the sub.
One brief word about all the terms being
bandied about: yes, a LOW CUT and a HIGH PASS are the same
thing. It is MOST USEFUL to use the terminology so it fits
the use of the situation. In one example, we have a filter in a recording
studio Microphone Preamp. Of course WE KNOW THE AUDIO GOES "THROUGH"
the thing; what we want to know is what we are doing - what "change"
we are going to hear when we click the switch! We are CUTTING THE LOWS.
In this instance the correct terminology is LOW CUT FILTER. In the case
of "filtering" a signal that's going to our mains, yes, of course
we are "letting the highs through" and we are also "blocking
the lows". So the typical useage for this would be "HIGH PASS"
filter. Technically and mathematically, either is correct. But it's always
a good idea to use the term which will yield the least confusion, especially
where people are concerned who don't necessarily have audio as a first language.
Manufacturers, pay attention...
don't want to introduce yet another active "thing" in their precious
signal path, not realizing that adding the crossover is very much the lesser
of two evils.
Actually adding a crossover is really
a WIN-WIN situation:
WIN # 1) Since you are
now NOT putting in 20 Hz - 80 Hz into the mains you are not
using up the available LF cone movement with bass, so the LF cone in
your mains is able to play its higher freqs (up to IT'S crossover
point) much more cleanly. You get an apparent 6dB or more dynamic
range. You can play your system LOUDER, and also with less compression
distortion in the LF driver when you're having that Saturday night dance
party and you're playing urban bass technopop at 110 dB. Really.
WIN # 2) Since you are
not putting bass into that same driver you are not Doppler modulating everything
between 80 and 600, or whatever the next crossover point is. This means
cleaner mids. By far.
WIN #3) You are not sucking
current out of your main power amp at low frequencies, so there is more
current reserve to play those highs louder...
WIN # 4) Since the cones
aren't moving as far at the low freqs the driver itself is not generating
as much back EMF therefore the damping factor and all of its issues are
greatly negated. And you don't need to run silver plated cold water pipes
to your mains as speaker wires because there is less current draw by the
WIN # 5) Freqs below 80
are now NOT causing transient intermodulation distortion with the higher
freqs (and vice versa) in your power amp. Cleaner still.
So let's start with the simplest method:
a passive, "filter" that blocks below 80 Hz from going to your
|Fig 9. Here's the Marchand
XM46SB PASSIVE Filter.
Here's how it's connected to a typical 2-channel system.
|Fig 9. The passive filter
used as "half" the crossover
So you roll off the mains at 80 Hz,
24 dB/octave; and you set the filter in the sub the same way.
To use an active filter (if it has 2
or more sections we can now call it a CROSSOVER), you have choices like
many versions of the Marchand
(solid state, tube, balanced, unbalanced, 1 way, 2 way, rotary knob, precision
stepped attenuator...) and some others.
Putting the active, 2-way crossover
in your system is done like this:
|Fig 9. The active crossover
Since ALL the filtering is done IN the
crossover, you turn OFF the filter in the sub. For fine level matching adjustments
you typically have a HIGH and LOW knob on the crossover to play with.
|Fig 10. One variation of
a Marchand crossover showing volume control knobs.
So all of this seems easy and yet with
ANY of the passive or active crossovers we have not YET addressed the critical
issue of the group delay in the sub. SO even though we have made everything
lovely in the frequency domain, the INHERENT delay in the sub is still there.
What are our options?
FIXING THE GROUP DELAY
We CANNOT change (or fix) the inherent
group delay in modern subs. That leaves us with two choices IF WE ARE
INTENDING TO BE FANATIC !
OPTION 1) We can move the
sub closer to our body about 9-10 feet or so, and then use the phase control
on the sub itself to fine tune the match. This is not necessarily as crazy
as it sounds. We do this successfully in studios all the time. Of course
this might not work in your particular room.
OPTION 2) We must introduce
an equivalent delay TO THE TOP (mains) to match the inherent delay in the
sub; then we can super fine-tune the match by using the phase knob on the
Some notes about phase knobs: If you
have a SWITCH on a sub labelled "phase" that is wrong. It is not
phase; it is POLARITY. Phase is ANY NUMBER of degrees shifted, from 1 degree
to 360 degrees to 3600 degrees and so on. Polarity is either 0 degrees or
180 degrees, period. (see Fig.1 and Fig.3 above) If you have a phase
KNOB on a sub, the circuit is usually designed to only ADD DELAY. You cannot
take away the inherent delay in the entire electro-mechanical physics of
the sub, but you can ADD further electrical delay. Some subs are calibrated
in electrical degrees of waveform at 80 hz, because 80 hz is ALWAYS the
magic frequency. Therefore IF the knob says 180 degrees it is actually adding
6.25 msec of delay to the sub signal; this is the equivalent of moving the
sub 7 feet FURTHER AWAY.
So how do we add delay to
the "top"? We would have to introduce a real processor
to do that. The options are a device like the DEQX,
or the Mcintosh version, here.
There is also the BSS Studio processor here.
All of these are audiophile grade devices. That means that UNLIKE
all the "digital" speaker gadgets intended for use in
nightclubs and rock n roll systems, these do NOT operate at 44kHz,
(or even 48kHz) and you will NOT be disappointed with what the
"processing" has done to your precious highs. Many of
the so-called "professional" units are perfectly suitable
for a noisy bar or a rock touring PA system but you might be very
disappointed if it is your intent to use them in a critical audiophile
listening/monitoring situation. That means beware of $99 - $299
In the instance of Home Theater processors,
there is an easy method. We can take advantage of the somewhat flawed concept
of "speaker distance settings" to perfectly fix the sub timing
issues. Simply set ALL the top speakers ( L C R Ls Rs) to 7 feet where
they belong, and set the sub distance to 18-19-20 feet.Now, because
all consumer equipment operates backwards (!!!) you are introducing 10 msec
delay TO ALL THE TOP SPEAKERS. Now you can fine tune the phase control on
the sub to add a bit more delay to perfectly match the mains and
the results should be spectacular. My test CD and the two different procedures
to accomplish this are all carefully explained here: www.soundoctor.com/testcd.
Once you correctly place the sub(s)
in your room so they correctly couple to your desired area, cross over the
mains to the sub correctly, and correct the timing issue your results will
be everything you hoped for.
IN SUMMATION ( pun intended...
I am therefore not suggesting that everyone
force themself to be so fanatic an audiophile, or to necessarily get this
crazy when setting up a sub. But I AM SUGGESTING that you should know ALL
the possible options and then you can decide just what is best for your
particular situation. Is it overwhelming? Yes. Is it a lot of work? Yes.
Did you just spend "a lot" of money on a subwoofer or two and
expect bang for the buck? Yes. Is it going to adjust itself? Sorry, no.
ONE LAST BIT OF RELIEF
Even if you CAN'T get the timing of
your sub to match your mains as closely as it can be done, there IS a saving
grace: re-read the paragraph above
about using 2 subs. Notice that humans actually LIKE the fattening up of
the bass loudness envelope in time. Therefore even IF your sub is 12 msec
late, and you are one wavelength off, as long as you get that delayed wavelength
to line up with the bass coming out of your mains, your frequency response
will be pretty good and you won't have any awful objections, again, assuming
you get as much else right as possible.
Enjoy your audio journey!