[youngho]

I have a few more questions after looking at the primer on bass mentioned in my previous post and the experiments with multiple subwoofers in http://www.harman.com/wp/index.jsp?articleId=1003

Mode cancellation seems to depend on the elimination of nodes by placing a second signal source on the opposite side of the listener at the same distance as the original. Does mode cancellation also disrupt standing wave formation and reduce the resonant peaks, as well?

For example, a listener sitting in the middle of a 10 ft wide room would be sitting in the node for the corresponding odd-order axial modes, like at 56.5. He or she might experience a relative null at the frequency, depending on the other modes of the room.

Subwoofer A placed directly against the wall directly to the left of the listener, would sit in an antinode for all axial modes corresponding to the 10 ft width and excite them accordingly. If subwoofer B is placed directly against the wall directly to the right of the listener in opposite polarity and same phase, B should cancel out the node for this odd-order axial mode at the listening position, right?

However, the listener is sitting in the antinode for the second axial mode at 113 Hz, so he or she may experience a relative peak at the frequency. Does B cancel out this peak, as well?

I guess, my essential question is whether "mode cancellation" actually is "node cancellation"?

Thanks for any assistance for this question, too!

Young-Ho

[youngho]

After spending much time squinting at the graphs in the "How to get the bass right" paper, and trying to compensate for things not quite lining up like I'd expect (possibly due to PDF layout), I think that mode cancellation probably does not do anything about even-order modes but that it seems to help with odd-order peaks, given the decrease in the first-order width peak seen with the listeners to the sides. Perhaps I'm wrong.

This would suggest that "mode cancellation" really does cancel odd-order modes, with respect to peaks and nulls?

It was remarkable to me how the peaks and nulls at different listening positions didn't line up in the graphs as I'd expect, possibly due cancellations occuring at slightly different wavelengths as the listener moved closer to one sub relative to the other or (less likely) to the walls themselves providing some absorption and making the effective width and/or length slightly longer, depending on the position of the listener (?).

Looking back at the Harman paper on multiple subwoofers, it seems to me that the corner position was best for a single unequalized subwoofer because it excited as many modes as possible. Midwall position for the single unequalized subwoofer resulted in failure to excite odd-order modes for the axis corresponding to that particular boundary, so there were fewer peaks, the average was lower than for corner placement, and standard deviation, Max-ave, and Max-min increased accordingly. In this room, with this listening position, it seemed as though the conventional wisdom of sticking the subwoofer in the corner really wasn't a bad place to start.

I certainly look forward to having the opportunity to own a home with a dedicated listening room that allows for the placement of and experimentation with two separate subwoofers with extensive parametric equalization. I'd also be interested in experimenting with adjusting phase and level to compensate for closer proximity to one subwoofer versus the other in a front-back configuration or setting the phase of one subwoofer 90 degrees out of phase with the other in a left-right configuration as per  Lexicon's Bass Enhance feature.

Someday...

Young-Ho

[Ethan Winer]

Young-Ho,

> It was remarkable to me how the peaks and nulls at different listening positions didn't line up in the graphs as I'd expect <

All peaks and nulls are not modal. You also get comb filtering at predictable distances from every boundary, and that's in addition to the modal peaks and nulls.

--Ethan

[youngho]

Thanks, I forgot about comb filtering. The first height mode doesn't seem to *quite* line up at different listening positions, and the peak expected from the second width mode seems to shift slightly upward in frequency when the second subwoofer is added, but I think I'm squinting too hard.

Young-Ho