Double Bass Array (DBA) - The modern bass concept!

[FoLLgoTT]

I'm really surprised to see that the "Double Bass Array" (DBA) is not very popular outside Germany. I searched for key words, but found only one thread in the AudioCircle forum.

Maybe someone is interested in this concept, since it is by far the smartest way to get a linear frequency response and completely get rid of room modes.
The DBA was described first in a paper from Klein+Hummel which unfortunately exists only in german. I hope my english is good enough to give a small overview here.


Configuration

A DBA consists of two identical arrays of subwoofers. One on the front and one one on the rear side.
The subwoofers have to be mounted on special positions on the wall. For example if you have 4 drivers in one array (that means overall 8) ranged in a square, their correct place is at 1/4 and 3/4 of the wall's width and height. Like this (my room in an early stage):



The side walls work like mirrors and have the same effect like more equidistant bass sources. This completely eliminates the room modes between the side walls and between floor and ceiling.
With this order the front array produces a plane wave which propagetes through the room. When it arrives on the rear wall the second subwoofer array creates the same signal but with inverted polarity. So both waves compensate each other and no reflection on the rear wall occur. The bass is completely free of modes!

Of course it only works, if the rear array is delayed by the time the sound needs to travel from the front array to the rear (delay = speed of sound / room length). Such a delay can be achieved by using cheap DSP equalizers like the Behringer Ultracurve 2496.
The level of the rear array usually has to be a bit lower than the level of the front array, since there is always a bit loss in real rooms when the wave propagates. But with measuring equipment the best setting is easy to find.

The subwoofers itself should have a low depth so reflection from the mounting wall get minimized. Ideally the drivers are build directly into the wall.

Of course a DBA also works with more or less drivers per wall. It is only important that the distance between 2 drivers is twice as long as the distance between the driver closest to a side wall and the side wall itself.
Both dimensions can be considered completely independent of each other.
For example if you want to use only 2 drivers per array, they have to be mounted on 1/4 and 3/4 of the room width and on the middle between floor and ceiling.
Denser driver grids conclude in a higher frequency where a plane wave will still be formed. With common room dimensions 4 drivers per array are enough to ensure a plane wave up to the LFE cut-off frequency.


Conclusion

The big advantage of this concept is to be completely free of room modes and to get exact the frequency response of the simulation. The maximum sound pressure of the DBA is defined as the number of subwoofers in one (!) array times the maximum sound pressure of one single subwoofer.
Another advantage is that the bass is fine on a large area and not limited to one seat. This makes it suitable for large home cinemas. Also the subwoofers can be integrated perfectly because of their small depth. A simple curtain is enough to hide them completely (my front looks like this now). I have seen other solutions which integrates them into self-built shelves in the living room.
And with a DBA playing "Sokoban" at home is over!

The only disadvantage I know of is the adaption to the front speakers, since their sound pressure usually lowers with 6 dB when doubling the distance and DBA's sound pressure stays constant in the whole room. So if you adjust a linear frequency response for the first row of seats the second will get slightly more bass (or in fact lower sound pressure from the fronts).

The costs of a DBA strongly depend on the components you want to use. My solution is a very cheap one with about 1100€ for the whole setup. But with more and bigger drivers the costs can easily explode to a multiple of that.



Frequency response of the Simulation (with same lowpass and equalizing like the real one):





1/3 octave smoothed measurement in the middle of my room (looks like a fake, but it is not! ):

[clubfoot]

Where did you take the measurements from, i.e. seat location or infront subs?

[FoLLgoTT]

Originally Posted by clubfoot

Where did you take the measurements from, i.e. seat location or infront subs?

The microphone was placed in the middle of the room (about 3m away from the front) in a height of about 1m. The room's dimensions are 5,6 x 3,8 x 3 m.

[coyotemoon]

Interesting....

But I am not sure I understand the real world advantage. You are using perhaps eight sub woofers in what seems to be an average room size, all just to eliminate room mode issues, that can be handled for most users by competent placement, room treatment or even active equalization and room analysis.

However, I do bet it sounds very nice!

Tom N.

[armystud0911]

yeah, but I have never seen a natural response curve that flat, very impressive if all this information is accurate.

[FoLLgoTT]

Originally Posted by coyotemoon

You are using perhaps eight sub woofers in what seems to be an average room size, all just to eliminate room mode issues, that can be handled for most users by competent placement, room treatment or even active equalization and room analysis.

Playing around with one or maybe more subwoofers just changes the impact of the modes and strongly depends on the listening position. In fact it is not more than trial and error and the result can never compete with a DBA or good room treatment.

Equalization only is a bad compromise since it only affects amplitude and not reverberation. For true audiophiles not an option.

The only working way is room treatment. The downside is that it is pretty complex and bass absorbers take extra space in the room. You have to measure a lot and higher knowledge about acoustics is needed. Only few users can do this.
But in non cuboid rooms room treatment seems to be the only way to go, because a DBA doesn't work here.


The real world advantages of a DBA are

• flat frequency response in a large area (perfect for home cinema)
• no modes and reflections at all
• very simple to build
• takes only space on the walls (no big boxes on the floor)
• can be integrated to invisibility (if done right you get a high WAF )
• sophisticated concept (you can tell your friends "I have an active absorber at home!" )

[FoLLgoTT]

Originally Posted by armystud0911

yeah, but I have never seen a natural response curve that flat, very impressive if all this information is accurate.

I can ensure you it is. But I have to admit that if a sofa would stand under microphone's position the frequency response is not that flat anymore. Every big furniture will decrease the DBA effect a little.
Right now on my listenings position (about 4m from the front) it doesn't look that flat, but even there it looks much better than with one or more conventional subwoofers placed in the room. You can go around and the bass never really changes audibly, pretty astonishing!

[catapult]

Interesting concept! The only problem I can see is it only fixes the sub but the mains also have modal problems that aren't addressed. Maybe a DBA sub would be a good match for dipole mains pulled well into the room.

Edit: Nils, do you have a room curve with less smoothing? 1/3 octave can cover many sins.

[FoLLgoTT]

Originally Posted by catapult

Interesting concept! The only problem I can see is it only fixes the sub but the mains also have modal problems that aren't addressed. Maybe a DBA sub would be a good match for dipole mains pulled well into the room.

In my setup the DBA plays the bass of all speakers up to 100 Hz (with 24 dB/oct like most receivers do). So the fronts don't have problems with room modes. Of course the reflections of frequencies over 100 Hz have to be treated specially.

Edit: Nils, do you have a room curve with less smoothing? 1/3 octave can cover many sins.

Here is a measurement with 1/24 octave smoothing and a longer FFT. As you can see the plane wave can only be formed up to about 90 Hz (distance between two drivers = half wavelength), so interferences begin to occur. A denser grid would perform better at higher frequencies.

[catapult]

That 1/24 octave measurement looks really nice.

[WallyWest]

Sounds like a three dimensional interpretation of the Harmon white paper's recommendations.

[catapult]

Originally Posted by WallyWest

Sounds like a three dimensional interpretation of the Harmon white paper's recommendations.

It's a bit different in that it uses active cancellation to make it sound like the room is open to the outdoors at the back. The idea is to launch a planar wave off the front wall with a delayed, 180 out of phase wave off the back wall, timed to fire when the front wave reaches it. So the wave just dies when it hits the back wall.

You could do about the same thing with a few feet of fiberglass on the back wall and a front wall IB with the drivers arranged to launch a planar wave.

[TheEAR]

Great post,and this setup is a true departure from what most are used to when dealing with room acoustics.

[crackyflipside]

Could the drivers in the rear be mounted in reverse and in same polarity (with some milliseconds of delay)?

[WallyWest]

Originally Posted by catapult

It's a bit different in that it uses active cancellation to make it sound like the room is open to the outdoors at the back. The idea is to launch a planar wave off the front wall with a delayed, 180 out of phase wave off the back wall, timed to fire when the front wave reaches it. So the wave just dies when it hits the back wall.

You could do about the same thing with a few feet of fiberglass on the back wall and a front wall IB with the drivers arranged to launch a planar wave.

Ok, I kind of skimmed through it first and didn't get the whole concept.

One question though, the OP says that the array's SPL stays constant through the room. And that's because the exact location of the drivers sets up reflections from the side walls (and floor/ceiling I guess) that reinforce the sound? So basically this design eliminates all room mode problems except for the rear wall. Which you can either treat with this nifty cancellation system, or a ****load of fiberglass.

Nifty.

[J_Palmer_Cass]

Originally Posted by FoLLgoTT

I can ensure you it is. But I have to admit that if a sofa would stand under microphone's position the frequency response is not that flat anymore. Every big furniture will decrease the DBA effect a little.
Right now on my listenings position (about 4m from the front) it doesn't look that flat, but even there it looks much better than with one or more conventional subwoofers placed in the room. You can go around and the bass never really changes audibly, pretty astonishing!

Just wondering, are those subwoofers (drivers) in a sealed enclosure, or is that an IB setup?

[catapult]

Originally Posted by crackyflipside

Could the drivers in the rear be mounted in reverse and in same polarity (with some milliseconds of delay)?

I don't think it matters much which way you physically mount the drivers. The thing is, whichever way you mount them, when the voltage goes (+), the air pressure goes (-) on the rear drivers. And the rear signal is delayed by a DEQ or whatever by however many milliseconds it takes for sound to travel from the front wall to the rear wall.

Originally Posted by WallyWest

One question though, the OP says that the array's SPL stays constant through the room.

I don't think that's strictly true. It's like a line array (except this is a 2D line array or a plane array) where you're always in the nearfield. In the farfield, SPL falls off at 6dB when distance doubles. In the nearfield, SPL falls off at 3dB when distance doubles. So the falloff with distance is less but it's still there. That's why the volume control needs to be a bit less on the rear speakers so they exactly cancel the front speakers' sound that has travelled all that distance.

[WallyWest]

Originally Posted by catapult

I don't think it matters much which way you physically mount the drivers. The thing is, whichever way you mount them, when the voltage goes (+), the air pressure goes (-) on the rear drivers. And the rear signal is delayed by a DEQ or whatever by however many milliseconds it takes for sound to travel from the front wall to the rear wall.

Slightly off topic, but can the BFD 1124 do that? Digital delay that is. Or do you have to go to the 2496 to get that feature?

[FoLLgoTT]

Originally Posted by WallyWest

One question though, the OP says that the array's SPL stays constant through the room. And that's because the exact location of the drivers sets up reflections from the side walls (and floor/ceiling I guess) that reinforce the sound? So basically this design eliminates all room mode problems except for the rear wall. Which you can either treat with this nifty cancellation system, or a ****load of fiberglass.

Yes, that's exactly the way it works.

Originally Posted by catapult

I don't think that's strictly true. It's like a line array (except this is a 2D line array or a plane array) where you're always in the nearfield. In the farfield, SPL falls off at 6dB when distance doubles. In the nearfield, SPL falls off at 3dB when distance doubles. So the falloff with distance is less but it's still there. That's why the volume control needs to be a bit less on the rear speakers so they exactly cancel the front speakers' sound that has travelled all that distance.

I think it is a bit different. Point and line sources lose SPL whith larger distance, because the surface of the wave (sphere/cylinder wave) grows and the sound pressure is always allocated on the whole surface. So the energy is always the same it is just spread on a larger area.

But the surface of the DBA's plane wave remains always the same when traveling through the room. In an ideal infenitely long room (with solid side walls) it could travel until the end of time with the same SPL.
As WallyWest said, the trick is the special placement which let the side walls work like extra bass sources (mirroring from the real sources). So in fact a DBA works like an open transmission line with constant intersection.

The loss in real rooms can be explained by not ideal conditions. For example in my room there is a rather thin wooden door in the side wall which vibrates a bit.

Originally Posted by J_Palmer_Cass

Just wondering, are those subwoofers (drivers) in a sealed enclosure, or is that an IB setup?

A time ago I used closed boxes. But since there is no room gain with a DBA (you practically get free air conditions) I had to use a +12 dB shelving filter in the lower end. This dropped max SPL a lot, because of high cone excursion. So right now all subwoofers are vented designs.

[J_Palmer_Cass]

Originally Posted by FoLLgoTT

Quote:
Originally Posted by J_Palmer_Class
Just wondering, are those subwoofers (drivers) in a sealed enclosure, or is that an IB setup?



A time ago I used closed boxes. But since there is no room gain with a DBA (you practically get free air conditions) I had to use a +12 dB shelving filter in the lower end. This dropped max SPL a lot, because of high cone excursion. So right now all subwoofers are vented designs.

OK. It's hard to get a perspective on size via your pictures.

What size drivers did you use, and what are the rough dimensions of those wall mounted subwoofers (H X W X D)? Also, with that many speakers, do you need high amplifier power, or do finess that issue wiring speakers in a special series / parallel manner?

Are the ports facing certain directions (like all up), or do they face different different directions for different speakers?

I like the concept myself, and I may try it in the room I am working on.

[FoLLgoTT]

Originally Posted by J_Palmer_Cass

What size drivers did you use, and what are the rough dimensions of those wall mounted subwoofers (H X W X D)? Also, with that many speakers, do you need high amplifier power, or do finess that issue wiring speakers in a special series / parallel manner?

I use 25 cm (10") drivers in 100 liters each. The boxes' dimensions are 100 x 80 x 17 cm. I could reduce the depth to 14 cm, but in the case of not enough max SPL I wanted the ability to change to 30 cm (12") woofers. But for my room it is enough (according to simulation about 115 dB down to 20 Hz).

Two of the drivers (8 Ohm) are wired in parallel and two pairs are wired in a serial manner. That results in an impedance of 8 Ohm which is no problem for most power amplifiers.
I use one stereo power amp with 500 W RMS. The front array is driven by the left channel and the rear array by the right. Amplification ist no problem at all with vented enclosures.

Are the ports facing certain directions (like all up), or do they face different different directions for different speakers?

The ports are all faced to the side walls not far away from the drivers. This seems to be no problem, because the wavelength is very long at 24 Hz and always result in constructive interference.

[Lindahl]

Very cool concept. I don't think it's feasible for my room, but I like it a lot. I can't imagine trying to do this with the 45 lbs 15" monsters I recently purchased, or even any of the smaller drivers people use here. But, for 10" drivers, no problem, and it sounds like it worked great. Maybe for my next theater...

[Jonomega]

nice concept, and thats a good "natural" 1/24 octave graph!

[Mark Seaton]

Good stuff FoLLgoTT,

This is a good solution to what I've observed many times over...

Placement in a room can easily deal with interactions in two dimensions of a closed room. The part that throws the wrench in the mix and causes all of the headaches is the back wall.

In looking at a room with this approach, you will be better served to consider the walls as reflective boundaries instead of looking at things from a room mode perspective. An ideally solid boundary effectively creates a mirrored image of the real sources. If you have a symmetrical spacing in one dimension and have mirrors on opposing sides, like two parallel walls, the virtual behavior is akin to an infinitely repeating array. How close the sources are placed to eachother and to the boundaries determines the upper frequency at which they couple to form effectively a homogeneous source. If you have side walls and floor-ceiling, you can then have effectively a planar source like described above. To the degree that the walls are more reflective than transmissive, this behavior will be observed in the real world.

I have not personally tried the cancelling idea at the rear wall, although it makes sense so long as you can clearly predict the interaction of both sources with the room in the cancelling dimension. In other words, if the subwoofer spacing isn't close enough for the upper frequency limit, the behavior will be more location dependent.

I'll have to do some more looking into this and experimenting. Of course the trade off here is potentially needing more subwoofers for similar SPL, especially at very low frequencies, but I could certainly see this being a useful option.

The idea that comes to my mind is in using some creative filtering to use this approach down to the lowest length mode and then either cut off the rear sub, or use an all-pass filter to shift the two subs back into polarity at the lowest frequencies so we don't give up VLF gain in the room. It certainly makes a strong case for a two way subwoofer for above and below the modal regions.

[andy_puiu]

Fascinating idea, and it makes perfect sense.

Two implementation thoughts that occured to me...

- A partially finished basement would be perfect for the fiberglass insulation option suggested above. Build a false back wall loaded with fiberglass, let the sound pass through this (attenuated), where after bouncing around back there and dissipating some more, it would have to again pass through your false fiberglass wall and get attenuated more. Wouldn't be ideal, compared to the active wave cancellation, but would save half the cost (half the drivers, half the amp eliminated) and also eliminate the need to tune the volume and time delay for wave cancellation.

- Could this work with a sonosub? Two tall sonosubs, with a driver at both ends? Might be hard to build/unstable, but assuming 8 ft ceilings, put a driver 2 feet off the floor (bottom) and another at 6 foot elevation (top). Maybe do cutouts and covering on the tube to make it look like it goes floor to ceiling but actually opens up at both ends. I don't know, just thinking out loud.

[andy_puiu]

We obviously need a better, catchier name for this... there hasn't been much interest/activity in this thread.

[cameronthorne]

I'm definitely interested, but I've only been lurking. It sounds like a fascinating concept, but nothing I would be able to use or even investigate for quite some time. I have only open, lossy rooms.

If I ever build my dream house with a dedicated room, this would be an interesting thing to combine with an infinite baffle alignment.

[crackyflipside]

Originally Posted by andy_puiu

We obviously need a better, catchier name for this... there hasn't been much interest/activity in this thread.

An Acronym!

Well it already has that; a DBA sub.

[FoLLgoTT]

Originally Posted by andy_puiu

We obviously need a better, catchier name for this... there hasn't been much interest/activity in this thread.

I think my titles often sound a bit too dry or too technical to get many clicks. Maybe something like "Ultimate subwoofer setup! The search is over!!!" would get the attention of every user, but I don't like such titles...

[FoLLgoTT]

Another interesting idea appeared at a german Hifi forum. The idea is to use only a Single Bass Array (SBA) on the front wall and to add a time shifted signal with inverted polarity to the original signal.
So the wave will be reflected at the rear wall exactly one time and then it will be canceled by the additional signal.

This works so far, but only down to the first room mode as you can see in the frequency diagram. At lower frequencies the room works as a compression chamber and the sound will not be reflected by the rear wall anymore.

A possible solution could be applying a high pass with a cut-off frequency of the first room mode to the additional signal. Sadly with IIR filters the whole phase response changes and doesn't fit to the original signal anymore. I played around a bit with delay and level, but I could not find working settings. Maybe a bit more time is needed...


SBA only:




SBA with additional inverted signal with inverse polarity:




DBA: