Double Bass Array (DBA)

[StigErik]

Measurements are here :

Dipole vs Box speakers

Ok—here is my first in room response curve sweep for the Enterprise Omnis. I used the mic in my iPhone for the test. I took the measurement from the centered listening position: Distances: Front wall: 6 feet Side wall: 4.5 feet Speaker separation: 6.5 feet Speaker To listener: 6.5 feet...

audiosciencereview.com

 

[architect]

have you tried reversing 6 of the 12 subwoofers like a checkersboard?
for the dipole 12” woofers you could do the same
distortion levels should even get lower wouldn’t they ?

 

[StigErik]

I know, have done that before so am familiar with the benefits.

 

[neRok]

Sure. It looks like ***, but sounds very nice!

Front end - including MG 3.7 with 4x12" midbass dipoles, MG 3.6 center channel with two 4x10" midbass dipoles. The whole thing is usually hidden behind drapes and a projection screen, so it's a bit less ugly than this (see pic #2).

View attachment 346991


View attachment 346992

Rear end, also showing the surround speakers - MG 3.6 plus 4x12" midbass dipoles - and hush-box for my projector. The large stack of firewood is actually "skyline" diffusors.

View attachment 346989

I came to ask about the installation details, but I can see them here. Very nice. I'm surprised it works so well because this relationship of subs to wall is meant to be the worst option, the best being flush in/with the wall, and next best being subs firing towards wall. But your results show that these minor details are dwarfed by the improvements from having the plane wave.

The actual difference is 5dB, not 9.

You connected 2 subs per channel to the front but 3 per channel at the rear. Perhaps some/all of the SPL difference is load (ohms/wiring) related? Or maybe not, I'm not 100% on the effects of wiring vs ohms vs power vs SPL.

Dips are the result if there are no modes or the ones that are there are not coupled to speakers/LP ("null").

It's all reflections, and modes are reflections where the frequency divides cleanly between 2 surfaces. The speaker or LP location along the mode determines if the overlapping sound/reflections boost each other or cancel each other. These handy visualisations show this.

This particular null might be more of an SBIR problem (which is also reflections), based upon LP to back wall. It's redundant when considering the plane wave because the back subs effectively cancel the reflection.

 

[StigErik]

I came to ask about the installation details, but I can see them here. Very nice. I'm surprised it works so well because this relationship of subs to wall is meant to be the worst option, the best being flush in/with the wall, and next best being subs firing towards wall. But your results show that these minor details are dwarfed by the improvements from having the plane wave.

You connected 2 subs per channel to the front but 3 per channel at the rear. Perhaps some/all of the SPL difference is load (ohms/wiring) related? Or maybe not, I'm not 100% on the effects of wiring vs ohms vs power vs SPL.

I can share a little details about the installation. Attached here are two simple drawings of the room layout that should explain most of it. The shape of my room with its cathedral ceiling is not optimal for DBA. The subwoofers should also be placed flush to the front and rear walls, which they are not. There are also five very large speakers in the room that will disrupt the plane wave. Even so, it works very well with some tuning.

I also mentioned that the subwoofers are hung (like a pendulum) from the ceiling using rubber rope and rubber straps. This is done to avoid vibration transfer from the subwoofers to the structure of the room itself. The longer the pendulum, the lower will the resonant frequency of the pendulum itself be as well. Assuming gravity is 1G, a 2 meter pendulum will have a resonance frequency of just 0.35 Hz, way lower than the frequency range of the subwoofers.

I have different number of amp channels and wiring schemes for the front and rear arrays, and the gain of the amps used is also different. That really got me confused! I've always thought that something around 9 dB var the correct attenuation of the rear array in my room, but measurements of the real SPL coming from each subwoofer (done at 1 cm from the speaker cones) showed that is more like 5 dB.

 

[StigErik]

The drawings above are not entirely accurate anymore. I've moved the sofa about 1 meter further back, and moved the front L R C speakers a little further away from the front wall as well.

 

[StigErik]

It's all reflections, and modes are reflections where the frequency divides cleanly between 2 surfaces. The speaker or LP location along the mode determines if the overlapping sound/reflections boost each other or cancel each other. These handy visualisations show this.

This particular null might be more of an SBIR problem (which is also reflections), based upon LP to back wall. It's redundant when considering the plane wave because the back subs effectively cancel the reflection.

I've investigated this a little, and the dip I see at 34 Hz looks like an SBIR thing to me.

 

[neRok]

I've been thinking about DBA some more. I'm wondering how the "noise pollution" side of things is? So what is the bass volume like in other rooms, or outside the home, etc? I'm guessing it's a bit lower than it would be with normal multi-sub, but still fairly loud?

 

[StigErik]

My experience is that it can be slightly less loud in neighboring rooms or outside the house than regular single- or multisub solutions, which I've had in the past. Don't have any numbers, just a feeling based on fewer complaints ...

 

[Justdafactsmaam]

Then there would have to be a null for the rear array too.

I think the cancellation of the front array at the rear array prevents the nulls and peaks.

Individually both the front and rear arrays would excite peaks and nulls.

 

[StigErik]

I've been playing around with an idea I've had. Below the lowest standing wave mode, we really don't need the active absorption from the rear array. If we instead make the rear array to be in phase with the front array we can gain some more maximum SPL. To accomplish this, I use the second subwoofer output of my AVR in reversed phase, do some low-pass filtering on that, and route it to the rear array only.

I've also fiddled a little with the EQ so the results are perhaps somewhat smoother than before. Anyway, here's the new response of the array with and without help from the rear array below about 20 Hz.

 

[Keith_W]

Flat down to 10Hz! So impressive.

 

[StigErik]

Thank you!
The results I’ve gotten here is really a showcase for what DSP is capable of. It would be impossible to achieve it without DSP.

 

[Bjorn]

How does it measure at other locations in the room? And what about stomach/leg area?

 

[Oddball]

My experience is that it can be slightly less loud in neighboring rooms or outside the house than regular single- or multisub solutions, which I've had in the past. Don't have any numbers, just a feeling based on fewer complaints ...

While I have nothing similar to your exuisite setup and my multi-purpose, and very irregularly shaped room could not really withstand high SPL below 25hz (rattling of various non HT intended structures), positioning 2 subs up-front and 2 subs back eventually alleviated in room as well as "neighbouring rooms" problems. While bass array is intuitive setup (at least visually), in my room it did not really produce optimal graphs, and did not sound at its "best" (both at MLP), so took me quite a while to get there.

I still have to keep below 25hz SPL under check, but can do that with higher SPL (i.e. no in room rattling and less neighbouring room complaints). I would say that even a primitive bass array setups with only 4 subs could help and they could sound "best" all factors considered.

 

[neRok]

Hi @StigErik. Would you be able to share some measurements of your bass array system (in mdat file)? With a common timing ref, I would like to see ones for the front subs on their own measured at the listening position, the same for the rears before array related DSP changes, then the rears with the DSP applied, and then finally the end result measured from listening position. I just want to get a better feel for how it all comes together based upon initial sweeps.

 

[StigErik]

Will look into it. Getting a good timing reference is a bit problematic. I don't have any speakers without long delays, since they all run FIR filters. I could dig up a passive speaker of course.

 

[neRok]

I wouldn't think that sort of delay matters because REW just waits there listening for the tweet, so as long as the tweet stays the same time wise, then it's a legit reference point/time.

 

[Elgrosso]

DBA are always so impressive, I need to try!

Since there is a big attenuation on the rear array with an MLP closer to the rear wall,
do you guys think that smaller subs could be used there? (by Sd or Xmax)

I want to try with 2.2 and then 4.4 subs but in theory I would need 6.6 and that’s just a lot, maybe even more since I have an L shaped room (but small L leg).
Mismatching by re-using what I have would help, I’m thinking 12'' in front and 10'' for example.

 

[Keith_W]

How does it measure at other locations in the room? And what about stomach/leg area?

@StigErik I would like to bump this question. I would like to know as well.