[denverdoc]

Not so much a question as an endorsement for the following link: http://www.harman.com/wp/index.jsp?articleId=1003

I'll try to summarize the article linked: a JBL/Harmon researcher uses computer modeling to determine the best placement for subwoofers in a generic rectangular room, using from one sub to a virtual infinite array in effort to provide reasonably uniform and flat bass FR to a centrally located area which might be used for HT use. Results are then verified w/in reason in a real room with SOTA measurement.

Conclusions: 2 are definitely better than one, under some circumstances 4 better yet. After that very diminishing returns.

Using these guidelines I took a pair of high powered active subs and placed them in one of the highly recommended, but counterintuitive, locations and was astonished to discover bass ripple (20 to 100Hz, measured at 1/10 octave intervals) was reduced from +/- 5 dB to at most 1/2 that--all w/o tube traps or any other kind of room treatment! A huge null from 70 to 45 and then a big 36 Hz peak which persisted throughout  the conventional wisdom locations and had me thinking about Helmholtz sponges were practically eliminated!

Moral of the story IMO: whenever, wherever possible, relieve mains of <75 Hz duties as the constraints of proper imaging, avoidance of early time-smearing reflections are ALWAYS at odds with smoothest bass FR. To say more, would relieve you of reading the paper.

Secondly, if you can meet first criterion, may save huge anount on bass absorbers, esp low WAF tube traps and the like.

Thirdly, by far the most elegant solution involves use of IB subwoofers. I'm putting two in my crawl space which, unless I have misunderstood the specs written up by T. Nousaine in some glossy mags for big players like Genelec and Velo 1812, will out perform in EVERY respect for a fraction of the cost. Reading the assault on Michael Green and the cultism which surrounds, I can w/o reservation recommend one cult-saved me, as opposed to costed, 5 to 10k   --the cult of the infinitely baffled.

J smolley, aka denverdoc

[zybar]

Not so much a question as an endorsement for the following link: http://www.harman.com/wp/index.jsp?articleId=1003

I'll try to summarize the article linked: a JBL/Harmon researcher uses computer modeling to determine the best placement for subwoofers in a generic rectangular room, using from one sub to a virtual infinite array in effort to provide reasonably uniform and flat bass FR to a centrally located area which might be used for HT use. Results are then verified w/in reason in a real room with SOTA  ...

So which locations did you use?

Where did you have them previously?

George

[ctviggen]

And, can you tell where the subs are?  If so, I don't think I could stand this.  I can barely stand it with a single sub behind my RF speaker -- even with level matching and lowering the crossover to 60Hz, I can still easily point out the location of the sub.

[ctviggen]

Darn!  I can't do this anyway -- I have no place to put the subs, as I have three doors along one wall.  Unless I buy 4 subs, I can't do this.

[Ethan Winer]

JS,

> an endorsement for the following link <

Yes, that's a great article, and there's no question that multiple subs can give a flatter low end response than only one. But a more uniform response is only half of the problem in small rooms. Those "low WAF" bass traps are still needed to reduce modal ringing, which is every bit as damaging to the music as the skewed LF response.

--Ethan

[zybar]

And, can you tell where the subs are?  If so, I don't think I could stand this.  I can barely stand it with a single sub behind my RF speaker -- even with level matching and lowering the crossover to 60Hz, I can still easily point out the location of the sub.

Bob,

I am crossing over my subs between 70-90Hz and they can't be located.

What do you hear that gives away the sub?

George

[markC]

I cross my subs over @ 40hz, (I have a pair of cheap ass Yammy's), and the only way that I can tell that the subs are/were in the mix is when I turn them off. I run my speaks full range, (obviously), and the subs just fill in that last 1/2 octave or so.

[denverdoc]

Some responses to questions above:

First: I tried the speakers in the usual suspect locations--because the towers attach to the subs I had less freedom than separate subs would allow--tried them near corners, away from corners, asymmetrically loading with one near a corner and the other speaker more in free space.

None of those arrangements came close to matching the overall smoothness of the bass response with the subs centered on opposite walls.

The Harmon paper suggests that the walls be front and rear, but on basis of symmetry I decided to use left and right walls as front and back would have caused some small difference in arrival time.

In this arrangement, the distance from mains and the two subs all more or less inscribed a circular arc. No, no cues whatsoever as to location. (Here the trick is picking both a low enough xover freq, and a commensurately steep filter--also the cabinets and baffle must not multiplex the frequencies covered, and the drivers/vents need to be free of extraneous chuffing/turbulence).

As to still needing traps, not sure whether I follow Ethan's argument re need for traps. The whole point of this exercise is to reduce modal ringing which is what perturbs the FR in the first place.

Put another way, Harmon paper showed: with a virtual array of several thousand subwoofers in a room, the FR asymptotically approaches that of a single driver in anechoic free space. The variation in FR from seat to seat also becomes much smaller. What might be a bit counterintuitive is that bass output actually drops with a larger number of woofers in the room and this is evident in the real world trials of this same paper with as few as 4 drivers.

I believe this paradox has to do with the reduction of room gain with multiple (>=4) subs. In other words, like any other engineering issue, ain't no free lunch--use the corners if you want to maximize room gain, but pay the price in big FR ripple. Add large N woofers to the room to get smoothest FR, and there is a definite loss of efficiency. My belief is that with an optimally placed pair, and judicious parametric EQ, the traps may not be necessary, period.

[mac]

by far the most elegant solution involves use of IB subwoofers. I'm putting two in my crawl space which, unless I have misunderstood the specs written up by T. Nousaine in some glossy mags for big players like Genelec and Velo 1812, will out perform in EVERY respect for a fraction of the cost.

IB woofers are subject to loading the room the same way a sealed box woofer would.  Nothing wrong with that, but placement within your room, as you pointed out, is critical.  IMO dipole woofers have a huge advantage of minimizing room mode excitement (due to their radiation pattern) no matter where they are located in the room.  If your primary goal is max spl the dipole solution isn't the most cost effective to implement.  Cheers, mac.

[zybar]

So you did midpoint L/R vs. midpoint F/B?

Can you post measurements?

I thought the article did show that you lost a little output at the bottom, but gained a more uniform response with the midpoint vs. corner placement.

I am going to try the L/R midpoint vs my current placement (near the corners, about 39" from the front wall and 24" from the side walls) and see what happens.

Unfortunately, I can't do F/B placement since I have my screen on the front wall.

George

[pjchappy]

I'm a BIG endorser of room treatments. . . they made a HUGE improvement in my place.

My question. . . can bass-traps make one sub (in one corner) disappear?  Or, is that part of their purpose?  

I heard a system that had one SVS sub in the corner and crossed over low.  I still could isolate it. . .the SPLs comin from that side of the room were just so much more. . .


Thanks!

p

[ted_b]

So you did midpoint L/R vs. midpoint F/B?

Can you post measurements?

I thought the article did show that you lost a little output at the bottom, but gained a more uniform response with the midpoint vs. corner placement.

I am going to try the L/R midpoint vs my current placement (near the corners, about 39" from the front wall and 24" from the side walls) and see what happens.

Unfortunately, I can't do F/B placement since I have my screen on the front wall.

George

George,
You have your subs 2 feet out into the room from the side walls?  Your pix make it look like they are snug against the side walls, about 2 ft back from the front.  New placements?

Ted_B

[zybar]

I need to post a more recent picture.  

Right now the subs are where I specified above and the main speakers are 38" from the side wall and 20" from the front wall.

George

[denverdoc]

George,

Sorry, should have saved the measurements, but can give you the highlights:

Prior to L/R wall centering--output from 100-85 w/in a dB or 2, then followed by a deep notch of >= 6dB with maybe some milder superimposed wiggles depending on which placement that reached a minima round 40, followed by a peak at 31 of 3, and then relatively uniform output to 22, with subsequent rolloff. Your basic roller coaster ride.

With Harmon placement: 100-75 more or less flat, followed by 3 to 4 dB notch with trough around 45-48, then gentle rise to reference level reached by 32, and uniform output to 19Hz, No lower freqs available. But sense was actually could have plumbed a few Hz deeper.

I could kick myself for not trying the f/r placements, but I sold the speakers and was so friggin delighted with the l/r wall placement I just didn't take the time. Given that I am gonna be carving pretty good sized holes into my floor, a really dumb oversite . But in defense, i was also overly biased against having a bass source behind me--as i have since thought thru the project in more detail, I will likely end up adding a pro audio limiter to protect the 18"ers from bottoming out as excursion and not thermal limitations are predominant in infinite baffle deployments, and as most of these have delay options, could have dialed in the 15'/1100 ft/sec delay for the back woofer and used a pair of amps instead of just paralleling the 2 drivers off one amp. May still do it as i have a couple of winks to think on this.

And of course the beauty of IB, is no screen issues or unsightly subs of any description to clutter up the room!

J

[denverdoc]

Mac,
Missed your post while typing my last. You're absolutely right that the IB's load the room the same way any other sub would. You also know from PC's that I am big on dipoles and dropping a bunch of change for a DIY pair. In the end, I decided that in spite of their many advantages, the demands made by HT and some music sources, like well recorded jazz at venue playback levels, creates a point of seriously diminishing returns for high spl between 15 and 50 cycles for dipoles. Hence i changes horses at the 60 Hz stream for the benefits of IB, which are many. Many ways to skin a cat!
J

[youngho]

if i understand it correctly, i think that the mid-wall placement of the subwoofer is beneficial in avoiding excitation of odd-order axial modes corresponding to the axis that it's in the middle of, that the pairing of the subwoofers results in axial mode cancellation which should eliminate those odd-order nulls which would normally occur from standing waves (does it eliminate peaks and even-order nulls, too?), and that the lesson is probably that your infinite baffle sub should be located in the center of the room (though that won't help the null that will occur halfway between the floor and ceiling).

bass traps will still be beneficial to address resonances and any existing even-order waves. also, this wouldn't address floor-ceiling axial modes or even-order axial modes corresponding to the horizontal axis perpendicular to the line between the subwoofers.

also, i don't think that low-end was significantly compromised by having two subs. the lf factor was decreased, but this metric was "corrected" for the number of actual subs.

young-ho

[denverdoc]

Youngho, Thanks for the thoughtful reply. As I understand it, and i make no claims to being an acoustician, the mid-wall location will break the room into shorter dimensions so to speak as there are now pressure maxima existing not only in the corners but mid room as well, so we have in effect cut the pipe in half. This will then shift axial resonances/nulls upward in frequency, and hopefullly by distributing the energy over a greater number of wavelengths, reduce rippling. As to the opposition of the two subs, I believe that in addition to recreating the beneficial effects of the first placement on the second wall--otherwise the opposing wall would twang pretty much behave as if corner loaded--there may be some additional advantages. Of course any such arrangement will have nulls and were I seated on axis of the opposition, I'd be even more concerned--but this is exactly what the paper argues for!

But I started this thread, less interested in theory which after all are usu based ona parallelopiped space that comes not close to describing my own listening room. I figured I would share my own enthusiastic observations, thinking other readers might reap similar benefits! And these were measured benefits, not like wow i have more slam, better extension, etc.

As to the LF being compromised, my take on the paper was that this became evident with 4 or more subs, and of course, it is "normalized"--put a 100 subs in any room, and one can get the foundation to rock and roll. Matter of bang/$$ ! My reading was that 2 will nearly reap the FR smoothing of four with nearly as much potential output.

I hope this clarifies my earlier post, and as to IB mid ceiling room null, why would this be any different than conventional subs? Seems like folk are really invested into having tube traps. Fine if you can stand their looks and the outrageous prices secondary sources ask for PU foam or whatever.
Personally if a certain freq cannot be eq'd out of downright nastiness will build a Helmholz resonator tuned to that f and either wall or ceiling mounted.
J

[youngho]

I'm no acoustician, either.  I took a year of physics in high school and another in college, but waves were always particularly confusing to me.

My understanding is this:

There are pressure minima and maxima throughout the room, depending on the order of the mode and the axis involved. In the center of the room, for example, with a subwoofer in the corner, the listener will be sitting in the null point (or node) of all odd-order modes and the maxima or peak (or antinode) of all even-order modes.

The mid-wall location is helpful because it places the subwoofer in the center of the axis of the wall that's it's up against, meaning that it will not excite any odd-order modes corresponding to this axis. It will continue to excite even-order modes because it's at an antinode for these modes. It's not exactly that you cut the pipe in half. Rather, by putting the speaker in the middle of the pipe, it doesn't create odd-order standing waves, but it does create even-order standing waves, just as it did when it was at the end of the pipe.

I guess the second sub is helpful in terms of mode cancellation by virtue of its location 1/2 wavelength away from the first. It shouldn't create any new nulls, but it should get rid of some of the existing ones as noted previously.

I absolutely believe your experience. I wish I had a suitable room to try it.

With respect to the mid-ceiling null, you might negate it by adding a second IB sub in the middle of the ceiling! Alternatively, depending on the type of subwoofer, a Hsu subwoofer might avoid this if it places the driver at the midpoint of the floor-ceiling axis.

Equalization is certainly more cosmetically acceptable than bass traps, as you point out. I just wanted to mention why bass traps might still be beneficial.

Sorry to hijack the thread,

Young-Ho

[ctviggen]

Too bad, Denver Doc, that you live in Colorado, or I could bring some bass traps to test.  If I ever buy two subs (which I'm trying to do right now), I can test your theory with and without bass traps and publish the results.  However, I won't actually be able to live with my subs this way, as I have too many openings along one wall and a sub would close off one of the openings (not to mention get in the way).

[zybar]

Too bad, Denver Doc, that you live in Colorado, or I could bring some bass traps to test.  If I ever buy two subs (which I'm trying to do right now), I can test your theory with and without bass traps and publish the results.  However, I won't actually be able to live with my subs this way, as I have too many openings along one wall and a sub would close off one of the openings (not to mention get in the way).

Bob,

If you want to burn a few hours, come on over and we can do measurements with and without the Minitraps in the room and at different locations.

George