Up to four subwoofers for maximum performance--What do you think? (1 Viewer)

[Phuong]

Has anyone has read Harman Audio's research on subwoofers?

http://www.harman.com/wp/pdf/multsubs.pdf

It's a pretty easy-to-read essay that addresses the issue of bass optimization while taking into account real-life limitations. I remember that Dr. Hsu has often recommended placing the subwoofer right next to the listener for maximum bass accuracy. Unfortunately, such ideal physics doesn't translate into something practical. Not many people want subwoofers and wires in the middle of their living room, which is probably why that idea has never taken off. Back in the 1990s, I figure Tom Nousaine legitimized corner placement of subs on the basis that it will excite the maximum number of room modes and generate maximum SPL--ie. it's the best compromise. From what I gather, the main tradeoff of corner placement is that it yields fairly even frequency response only for the person seated at the calibrated listening position--i.e the sweet spot. I think most of us are willing to live with this tradeoff.

Apparently, the folks at Harman have discovered the magic formula for subwoofer performance (in a rectangular room), which is four subwoofers, one against the midpoint of each wall, creating even frequency response for virtually everyone in the room (a larger sweet spot) while producing ok SPL (though less than 1 in the corner or 2 at opposing midpoints).

The second best method for balancing even frequency response (though I see it as the best it provides almost as much SPL as 1 in the corner and a sweet spot almost as large as 4 placed at the midpoints) with max SPL is 2 subwoofers, one at the midpoint of opposing walls. Floyd Toole also recommends equalization when this is all done (using a parametic equalizer). Obviously four subwoofers is less attainable for those of us who must work for a living and stack other things in our home theaters besides gear, but it would be interesting to hear from anyone who has tried this method.

The validity of their research—who am I to question their scientific credentials—is making me wonder which subwoofer would be best to buy if I wanted to eventually add another so I'd have two subwoofers at opposite wall midpoints instead of just one in the corner.

What do people think? Two boxes or two cylinders? I am currently looking at buying a SVS 20-39 PCi. Or just buy the best single sub I can afford right now, and worry about which pairs to get when I'm lucky enough to afford such a setup?

 

[douglas-b]

How big is your room?

 

[Jack Gilvey]

What I got from the paper is that it's not aiming for the most even FR response, but the most uniform (same peaks and dips) over a wide listening area. This is so eq can flatten it for multiple positions.
Didn't see the SPL comparison. I think it was referring to total room gain, not absolute SPL.

 

[Chu Gai]

One of the differences between what Nousaine did and what Harman did was that Tom actually took measurements with the sub in different positions while Harman used software to model the responses. At the time of Nousaine's work, his own personal experiences with the software modeling programs were that they were not correlatable to actual measurements.

 

[Phuong]

Actually, Harman did both software modeling and actual measurements (in a 20 x 24 x 9 room with multiple AKG-C98 microphones at ear level) to verify that real world results adhered to their theoretical models, which it did. Jack, you're right that absolute SPL is not specified. Their term is LF factor, with 1 in the corner having the highest.

I'm not here to defend Harman's methodology or findings, but was wondering if people have heard about these findings before and if not what impact this might have on people's approach to building a high quality home theater. As for me, it was a shocker in theory to see the tradeoffs of one in the corner vs. two or four at wall midpoints. In practice, I've never had complaints, but then again I never thought much could be done about extreme dips in the bass region and that there was nothing "better" than corner placement for creating a larger bass "sweet spot" for multiple listeners.

Currently, I run a single VTF-2 in the corner of the 13 x 16 x 8 living room of my apartment in case anyone is interested. The only reason I'm upgrading is because I will be giving the sub away to a friend. SVS 20-39 PCi is the next likely sub, and I'm looking to joining to club, and one day (a long way from now) to buy two and configure them according to Harman's findings.

 

[Tim_Stack]

Unless your living room is Radio City Music Hall, I can't imagine what someone would need with four subwoofers.

 

[SteveCallas]

Multiple subs in multiple locations isn't always needed to get a flat response. You should measure your current response in the corner to see what exactly is going on - measuring again with the second sub will be the only way you can know for sure which spot(s) work best for your room. I have my sub in my front left corner, about 9-10' from my seat, which isn't supposed to give the flattest response, yet I get a near perfect one without EQ:

I do benefit from a large open room with vaulted ceilings and non parallel walls though. So basically, you'll have to measure. RoomEQ Wizard is free to download and only requires an RS spl meter (I suggest digital), a sound card, and some cables.

 

[Jack Gilvey]

I printed out a copy long ago and fully intend to try out a couple sealed subs at opposing wall midpoints as my next project (bought a Behringer DEQ2496 towards this end) . I can't do front/back as in the paper, so I'm hoping right/left works as well...don't see why it wouldn't.


This paper is different from most in that it's not recommending placement for flattest response. The use of eq is assumed/intrinsic, and the recommended placements are meant to achieve a uniform FR across an area (it may or may not be very flat...that's not the purpose). Any applied eq would then effect this whole area, unlike the usual situation where applied eq can wreak havoc in a neighboring seat outside a small "sweet spot".

 

[Chu Gai]

I'll offer the following comments by TN on this.
Harman's approach appears to be highly idealistic and quite problematical for the home user to implement successfully.

 

[Wayne A. Pflughaupt]

A number of years ago I took multiple location readings with two subs in a 16’ x 14’ x 8’ room – both in the same corner, in two front corners, catty-cornered, and one in a corner with the other centered on the opposite wall, etc.

The readings for the separated subs were simply abysmal – so bad not even equalization could have helped. Without a doubt, the best readings came from the single corner location. Not perfect by any means, but easily smoothed with an equalizer. The only viable separated readings I got (i.e. that could be reasonably equalized) were from the both front corners arrangement. However, the penalty was a loss of maximum SPL.

This experiment, along with Mr. Welti’s, is only good to you if your room happens to have identical dimensions. Every room is different, so it’s best to experiment and take measurements. That said, the first place to try is typically going to be the corner.

Regards,
Wayne A. Pflughaupt

 

[HoraceHungry]

Multiple subwoofers can do the trick. At least that was the result in a recent norwegian test. They tested 1 Velodyne DD12 placed at the centre of the front wall, and then they added a second DD12 at the centre of the back wall.

The pictures speaks for themselves





The entire test in norwegian can be found here

http://www.audiophile.no/Publisering.asp?Id=378

 

[Chu Gai]

My norwegian is very rusty. Any idea how to get an english translation?

 

[Edward J M]

I use dual subs at the 1/3 and 2/3 points on the front stage with extremely good results in a 12x18x8 room. Aside from a mild bump at 65 Hz (fixed with a PEQ), the response is very smooth and linear without any sharp suckouts or peaks. And that's with the mains in the loop (something every in-room FR sweep should include IMO) and 1/24 octave resolution.

The theme in this thread should be 1) YMMV; 2) experiment with sub placement and measure the FR to determine what works best in your particular room.

Also bass traps can really improve the SQ and will help reduce the severity of any modal peaks/nulls. And after bass traps, a PEQ can also help greatly but remember not to boost true nulls - try to first address those with placement/traps.

 

[HoraceHungry]

Chu: If the article contained anything interesting but the pictures I would have translated it. But apart from the pictures they just reiterate the Harman study and talk up the DD series

 

[Mark Seaton]

I whole heartedly agree. In my experience, the larger, more irregular, and lossier the room, the more likely corner or a single center/front placement is preferred. As dimensions get smaller and boundaries more rigid, you can usually find placements for 2-4 subwoofers that will yeild smoother in-room response or better response across more seats.

 

[Edward J M]

Because of comb filtering, etc.? Sure, checking each speaker individually is easy. Actually with dual subs, one should check:

1) Sub #1
2) Sub #2
3) Sub #1 along with its respective speaker
4) Sub #2 along with its respective speaker
5) Subs 1 & 2 combined
6) Subs 1 & 2 combined with the left main
7) Subs 1 & 2 combined with the right main
8) Subs 1 & 2 combined with both mains

This systematic approach allow the user to see the contribition of each component and how they interact with each other. It takes time, but it can be worth it if you are seeing FR problems and not sure where they are stemming from.

 

[Tom Vodhanel]

>>>Back in the 1990s, I figure Tom Nousaine legitimized corner placement of subs on the basis that it will excite the maximum number of room modes and generate maximum SPL--ie. it's the best compromise. From what I gather, the main tradeoff of corner placement is that it yields fairly even frequency response only for the person seated at the calibrated listening position--i.e the sweet spot. I think most of us are willing to live with this tradeoff.

 

[Mark Seaton]

While I agree it is easy, it is just as easily overlooked or forgotten, and a common mistake I see when someone has spent a good deal of time EQ'ing and thinks they are done.

The only addition to your list is the inclusion of the center channel. In any movie Dolby Pro Logic setting, users can input a common signal to L & R channels and get output from the center channel. If trying to match levels between channels, remember that the summed signal to the center channel will be 3dB louder than the same signal sent to L or R channels.