Multiple Small Subs - Geddes Approach

Pallas said:
My current setup (unlike my previous one) violates Geddes' dictum that one subwoofer be above the centerline of the room. However, the mains as you can see are well above the room's centerline, so there are three 12" sources playing in the bass.

But don't forget the other keys of the Geddes approach, which are to run your mains full-range (LFE+Main and DoubleBass are common names for this mode on modern receivers) and set them up along the lines of Markus's excellent page. That includes both the ideas about placement and use of crossover frequencies that seem awfully high by the misguided standards of conventional wisdom.

The "above the centerline" is not a dictum, but a recommendation. I've often done without it, it just worked better that way in the simulations. But the real world is more complicated and we can never get such a clean setup as in a simulation. Hence, the actual field adjustment becomes the key. I don't think that you will get the optimum without some measurements, although listening can be sufficient at times. Markus procedure is correct and it should at least be tried. But, each room is different and each time I do this it comes out different. Its always an improvement, but nearly always different levels, etc.
 
Multiple subs

Cal Weldon said:
I am yet to hear a multi sub system that didn't sound cluttered and muddy. I will stick with one bass unit thank you.

There are two or three schools of thought on this, but I believe most everyone agrees that muliple subs in a small room is a good idea.

Outdoors, this would not be the case. You would want everything within 1/4 wavelength, if possible. Where not possible, you would still want to setup where interference were minimized, to put the lobes where you wanted them to be, in the listening area.

Indoors the problem is every sound source is accompanied my multiple virtual sources further than 1/4 wavelength away. They're the reflections from the room boundaries. So the idea is to average the sound field using a handful of sound sources. It is a way of using dense interference to smooth the sound field.

You can position the subs in any number of ways to do this. Welti suggested a handful of arrangements. Geddes suggests a random arrangement, one in a corner, one above mid height and another in a random spot. I like modeling the room to get a rough idea where to put the subs, then measuring to get it just right.

 
There can be only one school of though because it's just a physical phenomenon. Modeling the room is IMO a waste of time. You only get the first few modes right and then everything becomes sort of random in a real room. For calculating the first modes a hand calculator is sufficient.
Earls method is trial-and-error but doesn't take much time. You can place the subs anywhere you want and still get best results. At least as good as the location of the subs allow.

The only faster method I know of is that:
http://www.jbl.com/home/products/pr...age=ENG&Country=US&Region=USA&cat=EQU&ser=PER
 
The schools of thought I'm referring to are these:

1. Single monopole sub
2. Single dipole sub
3. Multiple subs arranged per Welti
4. Multiple subs arranged per Geddes

I think most everyone agrees that Welti or Geddes multisub placements make more uniform bass response in a small room than single sub installations. But there are differences between Welti and Geddes approaches. The main one is that Welti suggests symmeytrical placement and Geddes suggests pseudo-random placement.

If you use measurements to setup your subs, then I suppose modeling becomes less important. But I think it is a worthwhile step. It is easy to do and can help you know where to start.
 
There are differences in Welti's approach and mine. But the basic fact is that Welti only tried symmetrical placements, not random and I tried both. He recommends four as optimal, I recommned three, but in different setups. Its also true that Welti's modeled room was ideal - no doors or furniture, perfectly rigid walls, etc. and this is never the case in the real world. So how do you do Welti's recommendation? Well you can't really and it ends up being basically random anyways.

And I agree that I wouldn't bother to model the room if I were going to use measurements in the end to set it all up. I never worry about "where" the modes are, I'll find these soon enough when I measure the setup. And then I know exactly where they are.
 
What's with the attitude, Markus? I've usually considered you to be fairly level headed over the last few months of reading your postings. Your last few posts seem fairly uncharacteristic of you.

I've read both Welti's and Geddes' approaches, studied them in detail, modeled them both, tested them both, measured them both. I've tried various techniques and crossover approaches, more like what Geddes likes to call it, "blending" slopes of overlap.

I'm basically a horn guy, and have been championing uniform directivity loudspeakers for decades. I raised some eyebrows on some of the forums about five years ago when I said basshorns weren't the way to go for home hifi. It was like heresy for a horn guy like me to say I preferred direct radiators for subs, at least for home hifi. I make a pretty massive hornsub, but I never recommend it for home use. I always suggest multiple subs instead, usually made from a simple direct radiating LAB12 woofer.

So please don't mistake me for some guy that just stubled onto this idea. I've been doing the multisub thing for years.
 
markus76 said:
Sorry Wayne, but I really don't see any difference in "SFM" and "Gedlee Multisub".

Well, in a general context, you're right. They're both multisubs. That's the most important part, really, and in that sense I would agree. Especially in light of the fact that the more sound sources are used, the less important placement becomes.

Perhaps we should focus on the multisub feature, and not make this a Geddes verses Welti thing; however, Geddes has long made a point that the pseudo-random feature makes his proposition different from Welti's.
 
Wayne, Welti doesn't dictate symmetrical setups for SFM. So for me there is no "Geddes verses Welti thing". There's only one effect on the low frequency response when adding multiple subs. Varying their position, level, phase and bandpass let's you smooth the frequency response at a particular listening area.
 
gedlee said:
But then there is that SoCal Effect:D

Wondering just what freezin' yer keister in Michigan does to ya? One plus about cold weather though, it makes a warm body close by that much better.

:angel:

Back on topic, I'm working out a multisub install myself right now. We'll be looking at a group of sealed 12"s, (2x modified classic JBL AlNiCos modded for higher Qts and lower Fs along with cone stiffening/dampening in 7.5 ft^3 cabs) and 1 Castle Acoustics Classic (nice sub, small sealed 12).

I have a complex room with a lot of openings, so my plan is to put 2 at floor level, with varied distance to the sidewalls, and one a little higher up in the room to distribute the floor-celing modes.

I've had multiple subs before but they were less "subs" and more "low woofers", as they were playing up more like 70, 80Hz. I'm inclined to keep the subs low and have mains that go low too. While the geddes approach with less filtration makes some sense, what it doesn't appear to address is localization from distortion artifacts and higher frequency content, so I'll stick with 'mains' to 40 or thereabouts and subs below that.

Per one of the posters in this thread, it will look funny to have subs smaller than the mains (vented 15"s), but I'm inclined to think that the cabinet volume will make up for it :nod:
 
badman said:
While the geddes approach with less filtration makes some sense, what it doesn't appear to address is localization from distortion artifacts and higher frequency content, so I'll stick with 'mains' to 40 or thereabouts and subs below that.


Well actually I have dealt with that and talked about it before. This is exactly the reason why I use bandpass subs as they have acoustic LP filters which takes away any HF effects that might cause the subs to be localizable - which can happen. Closed box doen't have this advantage, otherwise I'd just use those.

With "mains to 40" and "subs below that" you won't get any multi-sub benifits at all. There would be no point in doing it.
 
markus76 said:
Wayne, Welti doesn't dictate symmetrical setups for SFM. So for me there is no "Geddes verses Welti thing". There's only one effect on the low frequency response when adding multiple subs. Varying their position, level, phase and bandpass let's you smooth the frequency response at a particular listening area.

SFM is an approach that uses signal processing in addition to source placement. It's newer than Welti's original work, where he modeled room/speaker setups with MatLab and then made acoustic measurements of a variety of configurations. His conclusions at that time were that subs should be placed at two wall midpoints, four midpoints or four corners. Those were what I would call "Welti configurations" and are what I believe Earl considers them to be also, from discussions here and elsewhere.

Again, I tend to agree that we should look at this as a multisub thing, not Welti and not Geddes but rather a generic approach that uses multiple subs to smooth room modes. Each room is different, so there probably shouldn't be a cookie cutter approach other than to say use multiple subs and to set them by measurement.

The thing is, historically, Harman International promoted symmetrical configurations suggested by Toole, Welti and others. Geddes promoted a similar setup but with a pseudo-random placement instead of a symmetrcial arrangment. Both have moved some, but Geddes, in particular, made it a point to suggest that the two approaches were different.
 
gedlee said:



Well actually I have dealt with that and talked about it before. This is exactly the reason why I use bandpass subs as they have acoustic LP filters which takes away any HF effects that might cause the subs to be localizable - which can happen. Closed box doen't have this advantage, otherwise I'd just use those.

With "mains to 40" and "subs below that" you won't get any multi-sub benifits at all. There would be no point in doing it.

Well, apart from the increased displacement ability of multiple units, and reduced thermal effects (at least that's my take) of multiple voice coils vs. one larger one. Also, let's not forget that -12dB @ 80Hz wouldn't be a ton of attenuation, and is what a typical sub crossover will give you, so you'll have smoothing contributions due to continuing output over the corner frequency up in the upper bass.
 
badman said:
I've had multiple subs before but they were less "subs" and more "low woofers", as they were playing up more like 70, 80Hz. I'm inclined to keep the subs low and have mains that go low too. While the geddes approach with less filtration makes some sense, what it doesn't appear to address is localization from distortion artifacts and higher frequency content, so I'll stick with 'mains' to 40 or thereabouts and subs below that.

This is very important. There are two competing priorities in any mutisub setup.

One priority is to smooth room modes with dense interference, which requires distributed sound sources in the modal region. The modal region can actually extend up pretty high, into the lower midrange in small rooms. The Schroeder frequency is the point where room modes become close enough to be indistinguishable, and so marks the approximate point where wave motion in the room begins to act as a reverberent field. This can be as high as 200Hz in small rooms.

The second (but equally important) priority is to prevent localization problems, which means you have to low-pass the subs lower the further away from the mains they are. In some medium size and larger rooms you also have to deal with inverse square falloff clues if the subs are too far from the mains. Once you get into rooms that large, multisubs aren't generally a good solution anymore but there is a sort of grey area transition between "big" and "small" rooms.

It is attractive to find solutions that allow the mains to go low enough to provide adequate bass, providing two bass sound sources built-in. Sometimes subs can be placed just a few feet away, far enough to provide smoothing of the midbass frequencies but close enough that they can overlap up to a fairly high point. Sometimes large format midrange and woofer subsystems can be made to overlap enough to provide this smoothing, which also smooths floor bounce. Subs placed further should usually be low-passed lower. The further away they are, generally, the more effective they are at smoother lower frequencies anyway. This also helps in regards to preventing localization.