[Ayreonaut]

I wrote an e-mail to Todd Welti to see what he thought of our ideas.

Daniel in an email to Todd Welti wrote:

Dear Mr. Welti,

I would like to seek your input on this. I'm an audio enthusiast, and I recently discovered your paper "Subwoofers: Optimum Number and Locations." and Floyd Toole's "Loudspeakers and Rooms for Multichannel Audio Reproduction." I also read the interview that you did with Wes Philips in 2004 on the subject.

I am very interested in implementing the solution of using four subwoofers located at the 25% points from all four walls. My room is rectangular with two open doorways. I plan to install four subs in my ceiling in an "infinite baffle" configuration. When I flew this idea past my friends at the Home Theater Shack forum they posed several reservations. Can you tell us if these objections are valid?

1. The four subwoofers will not be mutually coupled (and output may fall short.)

2. Four box subs pulled out from the walls would perform differently than four Infinite Baffle subwoofers at the same locations.

3. The subwoofers would be competing out of phase at certain frequencies depending on their spacing.

4. Phase issues will exist unless all of the subs are equidistant to the seating position. This will show up as ragged, saw-tooth response, and since it is a time-alignment issue, you’ll get different results at every listening position.

5. Experiments in sub placement are relevant only in the room they’re performed in.

Would you please share your opinion on these? You can respond to my e-mail or feel free to answer in the forum where we are holding this discussion at Home Theater Shack. Thanks for your time and efforts.

Sincerely,

Daniel

Mr. Welti was very kind to take time and respond to all of these. (Even though he's on vacation!)

Todd Welti in an e-mail to Daniel wrote:

Hi Daniel,

first off, the most recent and complete reference regarding the subwoofer work we did can be found in the May issue of the AES Journal. There is some new stuff in there that might interest you.

To address your questions:

1. The total acoustical output into the entire room space may be less, but that does not necessarily mean the output at the individual seats is less. Of course, you are using four subs. That will be in your favor.

2. I wouldn't expect the response of the subs to change too much when baffle mounting them. I assume you are crossing over at 80 Hz. The acoutical center of a subwoofer mounted right next to a wall is probably 12 to 18 inches from its reflected image, i.e. only a fraction of a wavelength at 80 Hz. So baffle mounting shouldnt make too much difference. Even if it did, it would be a difference in the overall (global) response, and would certainly not affect the seat to seat consistancy we are after. Afterthought: if baffle mounting be sure to isolate mechanically from wall.

3. This is true, but dont forget that even if they are IN phase, you can still have cancellations at a particular seat due to room modes, which will be spatially and frequency dependent (the worst kind of cancellation). This is what we are trying to avoid. There is no point in having more acoustical energy if it is only at certain frequencies.

4. I am assuming you have more than one seating position - optimized sub layouts such as the one you are interested in are only really relevent in this case. In such a case you cannot have equal distances form all subs to all seats. If your seating layout is symetrical, and your sub layout is symetrical you could at least have consistant sets of path lengths from the subs to the seats. In any case, unless the room is very dead, the room modes cannot be ignored. You cannot just think about the direct paths.

5. True only to a certain extent. It's not the dimensions so much (for example, wall midpoints are first order mode null points regardless of the dimensions). I did include an analysis of different dimensions in one of my AES papers, and found that the results held. (Sorry, I dont have the reference with me right now, I'm on vacation.) The wall midpoint solution is allways best. If the room has a large opening into another space, or wall construction that is very unsymetrical (for example brick on one wall and single layer sheetrock on the opposing wall), there may be divergences from the theoretical best sub solutions. In this case you may want to look into another technique such as is described in our recent AES paper, Sound Field Management.

You can copy this to your forum if you want.

Cheers, Todd