[Duke]

Apologies in advance to readers of Robert Green's Yahoo group... much of this is re-cycled...

Often discussions of "which sub for Maggies or Quads" includes recommendations of dipole subs, and for good reason:  Dipoles blend well with dipoles!  Let's take a look at WHY dipoles work well, because I think there's a common misconception.  And let's look at multiple monopole subs, and see what they do better and what they do worse, and see if there might be a "best of both worlds" approach (spoiler alert:  I think there may be two "best of both worlds" approaches). 

Okay let's start by dividing the bass region into two zones (and I realize these aren't new concepts for most of you):  The "modal zone", where we get those familiar and unwelcome room-interaction peaks and dips; and the "pressure zone", down below the modal zone, where the wavelengths are so long relative to the room dimensions that all of the first reflections occur within 1/4 wavelength of each other, which tends to put all of the in-room bass energy in-phase.  We will come back to these concepts later.

Another concept we want to look at is "de-correlation".  This concept is less well-known, but it is the primary acoustic smoothing mechanism for in-room bass in both dipole bass systems and distributed multiple monopole bass systems.  We get good de-correlation when the in-room bass energy is adding in semi-random phase.  De-correlation is desirable because it results in smooth bass, because no particular frequencies have enough energy to stick out like sore thumbs (or "peaks").  To help you visualize what we mean by the term, good "de-correlation" would be like the foot-falls of a highschool football team running out onto the field.  On the other hand, strong "correlation" would be like the foot-falls of the Wehrmacht on parade.  The thing to remember is, good de-correlation = smooth in-room bass. 

And smooth bass is "fast" bass, literally and perceptually.  Because room + sub(s) = a minimum phase system, in-room peak = slower energy decay at that frequency = "slow" bass.  So if the perceptual goal is "fast bass", the road thereto is the one marked "smooth bass".

In general, dipole bass really is smoother than monopole bass!  In 2002 researcher James M. Kates published a paper in the Journal of the Audio Engineering Society showing that dipoles have smoother in-room bass than monopoles.  This wasn't news to the Gradient and Quad and Maggie and SoundLab and Apogee owners out there, but it WAS long-overdue officially-sanctioned peer-reviewed validation. 

Now most people who talk about the "speed" (in-room smoothness) of dipole bass attribute it to the figure-8 radiation pattern exciting fewer room modes.  But if only few-and-far-between room modes are excited, that's an indication of strong correlation (Wehrmacht)!   And we'd expect those few-and-far-between peaks and dips (especially the peaks) to stick out like sore thumbs.  So, I think that attributing the in-room smoothness of dipole bass entirely (or even mostly) to the figure-8 radiation pattern is a misconception.  Kates' data shows smoother in-room bass for dipoles, which implies good de-correlation, so something else must be going on.  Here's what I think is happening:

The primary smoothing mechanism for dipole bass is, the phase relationship between the frontwave and the backwave energy.  The figure-8 radiation pattern also plays a role here, so it does matter, but not all by itself.  Because of the radiation pattern, the normal-polarity frontwave goes off in one general direction, and the reverse-polarity backwave goes off in the opposite direction.  After a few room boundary bounces the twain shall meet again, but their phase relationships are now generally semi-random - which means good de-correlation - highschool football team!

[Not that I'm necessarily the world's biggest fan of highschool football teams, whose members were often the dispensers of wedgies and noogies to us little nerdlings, but at least football gave them someone else to hit and hit hard.]

In general, it takes two intelligently-distributed monopole sources to get the same approximate in-room smoothness of a single dipole source.  The energy from the dipole source sums in-room in semi-random phase because the frontwave and backwave start out in opposite polarity and are launched in opposite directions, and the energy from two intelligently-spaced monopole sources also combines in semi-random phase because of their physical distance from one another and dissimilar physical distances from most of the room's boundaries. 

So the same mechanism - semi-random phase summation - is at play in smoothing the in-room response of both dipole bass systems and Swarms in the modal region.  But the two approaches totally diverge as far as what happens down in the pressure zone!   If the energy in the pressure zone is all in-phase (like with four monopole subs), then it sums in-phase.  This can result in boom, because in-phase summation is inherently 3 dB louder than the semi-random-phase summation we were getting up in the modal zone.  And because of the ear's heightened sensitivity to relative loudness at low frequencies (which is shown in the bunching up of equal-loudness curves south of 100 Hz), a 3 dB emphasis below 40 Hz can sound as big as a 6 dB emphasis up in the midrange region.  (This is also an argument for most systems having significant room for improvement in the bass region.)

With a dipole, at very long wavelengths the situation is kinda the same but the outcome is totally different.  In the region where all of the reflections are happening within 1/4 wavelength, we have half of the energy with one polarity and half with the opposite!  The net sum would be complete cancellation ( making it a "no-pressure zone"?).  In practice the cancellation isn't complete, but it is severe, and as a result dipole bass tends to have poor impact unless a) the room is very large (pushing the "no-pressure zone" down very low in frequency) and b) the dipole bass system actually does go very low (which requires a very large baffle and/or a lot of EQ). 

So intelligently-distributed monopoles can match the in-room smoothness of dipoles in the modal zone as long as we use enough monopoles.  But down in the pressure zone, the monopoles tend towards having about 3 dB of excess energy, while the dipoles tend towards complete cancellation (virtually no bass energy).   Neither situation is ideal, and subjectively "too much low bass" can be even more distracting than "not enough low bass". 

But, the Swarm has one final trick up its sleeve!  We can reverse the polarity of one of the four subs!  Now as we move down into the pressure zone, we have 3/4 of the energy in normal polarity and 1/4 of the energy in reverse polarity, so we end up with neither the low-end over-emphasis of the pure monopole (all woofers in-phase) approach, nor the trending-towards-complete-cancellation of the dipole approach.  And by reversing the polarity of one sub, we are also improving the de-correlation up in the modal zone.  So we end up with smooth bass all the way down, with no over-emphasis anywhere.  And remember, smooth bass is "fast" bass.

Okay, is there another approach that can give similar results?  I think so, and it's something Siegfried Linkwitz has been doing for decades:  Use dipole bass down to the transition between the modal and pressure zones, and use monopole bass down in the pressure zone.   Both approaches work well.

Not all rooms will have a significant pressure zone - open-floorplan rooms won't - so not all of the distinctions made above will be equally applicable.  The best answers are probably arrived at on a case-by-case basis.   But I think that a good distributed multisub system, well set-up, is competitive with a good dipole system across the modal zone and superior to both dipole and pure monopole (all woofers in-phase) down in the pressure zone.

Duke

[TomS]

Good read Duke. Thanks for re-posting 

[opnly bafld]

Thanks for sharing Duke.

[Duke]

Thank you both very much!  I admire your determination if you actually managed to read all of that.  I hadn't realized how long it was until it actually posted.

[poseidonsvoice]

Duke,

Thanks for posting! Correlates with what I have experimented completely! In fact, to satisfy my own curiosity and dipole bass aficionados, I experimented with "multisub dipole bass" specifically with Rythmik drivers (so theee would be no argument from naysayers with regards to "quality" and benefits of "servocontrol"). What did I find? Although it sounded good, it failed in the "impact" and "realism" department compared to my current setup that is using multiple high efficiency subs (95 dB in my case) in the modal region, and 1 or 2 subs (depending on room size, preference) for the 1st mode. A simple FR curve wasn't able to discern the differences between the 2 setups (as they both measured great) nor did other measurements like waterfall, etc...I imagine it will come down to things I can't measure and can't prove either like more freedom from compression, lower THD & IMD of the drivers used at that particular SPL, etc, etc...

Best,
Anand.

[AJinFLA]

I've posted this a couple times already:

https://secure.aes.org/forum/pubs/conferences/?elib=17270

An overview of about 30 bass studies.

[THROWBACK]

A fascinating and well-written paper. In the past I have owned big dipoles (double KLH-9; Infinity Servo Statik; Acoustat 4; Sound Lab; Maggie. OK, so I'm restless). My attempts to mate monopole subs with them never came off well, for reasons that Duke has outlined quite cogently.  Now I have OB main speakers (GR Research LS9s) mated with a three-stack of GR Research (Rhythmik) OB servo woofers per side. Aaah! Much better.

But I'm still restless. Although my system presents a good "startle" factor, I'm thinking it could have more. Duke's post plus Anand's suggest that a combination of monopole and dipole subs might be  worth trying. So maybe I should just add a couple of monopole subs.  But, boy! That sounds like a huge complication: amps (with volume controls), cables, positioning. And can my preamp handle a third output? This may be an itch I'll never really be able to scratch.

[Jonathon Janusz]

THROWBACK, if you can swing it, I know from hearing basically your same setup (Mockingbird LS-X with two triple stack OB servo subs) at RMAF, with and without a pair of sealed servo subs firing out of phase from the back of the room, the result from adding the sealed subs is well worth it.  I would from my experience compare it to taking what you have now, with all its goodness, and adding the foundation in the bass the boxed LS-9 brings to the table compared to the same speaker OB.  Evolutionary rather than revolutionary, but definitely not subtle.

[Duke]

Duke,

Thanks for posting! Correlates with what I have experimented completely! In fact, to satisfy my own curiosity and dipole bass aficionados, I experimented with "multisub dipole bass" specifically with Rythmik drivers (so theee would be no argument from naysayers with regards to "quality" and benefits of "servocontrol"). What did I find? Although it sounded good, it failed in the "impact" and "realism" department compared to my current setup that is using multiple high efficiency subs (95 dB in my case) in the modal region, and 1 or 2 subs (depending on room size, preference) for the 1st mode. A simple FR curve wasn't able to discern the differences between the 2 setups (as they both measured great) nor did other measurements like waterfall, etc...I imagine it will come down to things I can't measure and can't prove either like more freedom from compression, lower THD & IMD of the drivers used at that particular SPL, etc, etc...

Best,
Anand.

Thank you Anand!

Hopefully one day measurements will catch up to perception... right now measurements are usually excellent tools for telling us why we hear what we hear, but your experience is an example that there is still ground left to be covered.

I have these two somewhat similar bass guitar cabs and when you compare the measurements, one should have "more bass", but perceptually the other one does.  This has been reported by all of my customers who own both or have heard both side-by-side, except for one... and that one is the guy who does measurements for a magazine.  I suspect that his ears are calibrated better than all the rest of ours, but serious bass players are extremely good at this sort of thing, so that doesn't explain the majority's perceptions.  It would be really nice if I can figure out how to optimize for best perceived low end. 

It seems to me that sometimes the better the "grip" the motor has on the air, via larger cone area for instance, the better the impact.  A heavy cone can still work if the motor is powerful enough, but power-to-weight ratio seems to matter too, which trades off against cone area.   I've tried woofers that looked good through this lens and were distinctly "mushy" down low, so it's more complicated than that.  Your approach of modal-region-optimized high efficiency woofers + pressure zone woofers may be the way to really nail it. 

[Duke]

I've posted this a couple times already:

https://secure.aes.org/forum/pubs/conferences/?elib=17270

An overview of about 30 bass studies.

Thank you AJ!

My membership expired during the Great Recession a few years ago, and I haven't renewed it yet.  I'll buy the article if it's really valuable.  Can you give me a few sentences that describe what you learned from it?
 

[AJinFLA]

Thank you AJ!

My membership expired during the Great Recession a few years ago, and I haven't renewed it yet.  I'll buy the article if it's really valuable.  Can you give me a few sentences that describe what you learned from it?

Multiple subs in mono results in smoother amplitude, but loses ability to recreate spatial info in recordings.
Gradient (cardioids) showed the highest immunity to modal issues/placement in rooms and offered lower decay times. Lower decay times increased discrimination of spatial effects.
The least researched (2006) was 5(+) full range channels, each driven by appropriate signals (like JJs PSR).

Btw, AES membership ($125/yr) gives full access to entire library. No need to purchase individual papers any more. I can't think of $125 better spent if one has the slightness interest in "Audio".

cheers,

AJ

[JohnR]

the "pressure zone", down below the modal zone, where the wavelengths are so long relative to the room dimensions that all of the first reflections occur within 1/4 wavelength of each other

Hi Duke, could you clarify the "pressure zone" definition? The above seems to suggest that the room might need to be 1/8 or less of a wavelength in the longest dimension (i.e. if I sit on top of a sub located on the short wall).

[Duke]

But I'm still restless. Although my system presents a good "startle" factor, I'm thinking it could have more. Duke's post plus Anand's suggest that a combination of monopole and dipole subs might be  worth trying. So maybe I should just add a couple of monopole subs.  But, boy! That sounds like a huge complication: amps (with volume controls), cables, positioning. And can my preamp handle a third output? This may be an itch I'll never really be able to scratch.

IF your main amp is sending a fullrange (rather than highpassed) signal to your main speakers, perhaps you could build a voltage divider network which would derive a line-level signal from the amp's output, and use that to drive one or more subwoofer amps.

[Duke]

Multiple subs in mono results in smoother amplitude, but loses ability to recreate spatial info in recordings.
Gradient (cardioids) showed the highest immunity to modal issues/placement in rooms and offered lower decay times. Lower decay times increased discrimination of spatial effects.
The least researched (2006) was 5(+) full range channels, each driven by appropriate signals (like JJs PSR).

Thank you AJ, very interesting.   I'm under the impression that very few recordings actually have stereo bass.  Some Swarm users are using two amplifiers, probably in stereo in most cases, and driving the two subs on one side of the room with one amp and the two on the other side with the other amp.  At least some of them are also using the phase controls to put the subs on one side of the room in phase quadrature (90 degrees apart) from the subs on the other side, which is a technique suggested by researcher and acoustician David Griesinger for synthesizing hall ambience from recordings with mono-summed bass.

Very interesting that cardioid sources are the most immune to modal issues!   Since cardioid subwoofer require two woofers per unit (one of which is used to partially cancel the other) along signal processing and separate amplifiers, it's probably relatively expensive compared to implement.   

I wonder how four monopole subs with one in reverse polarity would have compared.  Oh well, we'll probably never know, as that's a pretty obscure technique.

Hi Duke, could you clarify the "pressure zone" definition? The above seems to suggest that the room might need to be 1/8 or less of a wavelength in the longest dimension (i.e. if I sit on top of a sub located on the short wall).

My definition of the "pressure zone" was quite sloppy!  Thanks for catching that!!

Here is a better one:  The "pressure zone" is the region below the frequency that has a wavelength of twice the longest internal room dimension. In this zone, sound behaves very much like changes in static air pressure.  The longest room dimension is actually the oblique dimension, going from one bottom corner to the opposite top corner. 

[Tyson]

I used to have some crazy bass systems that could do very high output down to 16hz, and I found that I didn't actually like it very much.  It pressurized the room and made me basically feel like I was under water.  Very uncomfortable feeling for me.  I ended up with OB (servo) bass because it gave a lot of punch, but specifically did NOT pressurize the room like a high output box(es) solution.  Plus, most of the 'percussive' bass is 50hz and above, that's where the 'kick you in the chest' feeling comes from. 

Even though my subs were pretty crazy, they were nothing as insane as my friend Thomas' system, the "12 Shivas Dancing" infinite baffle which was able to put out prodigious bass, to say the least.  And same problem, for me - the pressure on my inner ear was just downright uncomfortable.  It felt like I was swimming.

[Duke]

I used to have some crazy bass systems that could do very high output down to 16hz, and I found that I didn't actually like it very much.  It pressurized the room and made me basically feel like I was under water.  Very uncomfortable feeling for me.  I ended up with OB (servo) bass because it gave a lot of punch, but specifically did NOT pressurize the room like a high output box(es) solution.  Plus, most of the 'percussive' bass is 50hz and above, that's where the 'kick you in the chest' feeling comes from. 

Even though my subs were pretty crazy, they were nothing as insane as my friend Thomas' system, the "12 Shivas Dancing" infinite baffle which was able to put out prodigious bass, to say the least.  And same problem, for me - the pressure on my inner ear was just downright uncomfortable.  It felt like I was swimming.

Thanks for posting your experiences with uberdeep room-pressurizing subs.  I don't feel so bad now that my subs only go down to about 20 Hz. 

By any chance did you ever try something like the four-subs-with-one-in-reverse-polarity setup?  Not that there would have been any obvious reason to do so, just curious.

[Tyson]

No, I think at the time no one had even thought about a swarm approach, let alone tried it. 

As you know from my RMAF reporting, I do really like the AK Swarm 

Mainly my point was more of a "be careful what you wish for", especially with high output low bass.  For me, room pressurization was a very real negative.

But I think a lot of people haven't had a super high output low bass system and might end up with a nasty surprise if/when they finally get one.

[AJinFLA]

Thank you AJ, very interesting.   I'm under the impression that very few recordings actually have stereo bass.

Paper 41 of 45 in the study I linked addresses just that http://www.aes.org/e-lib/browse.cfm?elib=13358, as did this thread in diyaudio http://www.diyaudio.com/forums/lounge/247583-can-human-ear-really-localize-bass-18.html.
Some caveats: The vast majority of popular music recordings don't. Usually acoustic music types that might actually be performed in actual spaces. Jazz, classical, etc.

which is a technique suggested by researcher and acoustician David Griesinger for synthesizing hall ambience from recordings with mono-summed bass.

Griesingers work is referenced also, as are other decorrelated maximum separation (+/- 90 degree) lateralization methods. He is one of the early advocates of smooth amplitude alone (free of peaks) being not the sole arbiter of bass SQ.

Since cardioid subwoofer require two woofers per unit

They do not. There are many examples of flow resistance enclosures that generate cardioid response, ME Gethain, Music & Design (John K), http://kimmosaunisto.net/CardSub/CARDSUB.html etc.

It should be also noted from the studies that there is significant differences in ability to lateralize among the population.
So as usual, YMMV, depending on many factors, including music preferences.

cheers,

AJ

[Jonathon Janusz]

Very interesting that cardioid sources are the most immune to modal issues!   Since cardioid subwoofer require two woofers per unit (one of which is used to partially cancel the other) along signal processing and separate amplifiers, it's probably relatively expensive compared to implement.   
site top corner.

If you are after a cardioid pattern from a subwoofer, wouldn't it be more easily achieved with a U-shaped open baffle (a boxed sub with an open back) or maybe a H-frame with different depths to the wings front or back?  This would eliminate the signal processing and one amp per sub?

[poseidonsvoice]

If you are after a cardioid pattern from a subwoofer, wouldn't it be more easily achieved with a U-shaped open baffle (a boxed sub with an open back) or maybe a H-frame with different depths to the wings front or back?  This would eliminate the signal processing and one amp per sub?

Jonathan,

I built a pair of U-frame cardiod subs designed by John Kreskovsky, several years back. It definitely required EQ and of course amplification. The driver was specifically designed for dipole and cardioid applications. So I have no idea where you are coming from with the notion that a U-frame and H-frame eliminate or even minimize the need for signal processing. Of course they sounded quite good but due to the relative inefficiency of the drivers used, they were SPL compromised, and as such, I moved on.

I think some degree of signal processing and/or EQ will always be needed to achieve a flat bass response in room. It however is minimized the more subs that are used in the room as long as the subs are somewhat asymmetrically placed and not grouped together.

Best,
Anand.