What I found interesting (one thing of many) in the talk was the suggestion (which makes decent sense) that the "family of curves" will predict the in-room response with very high correlation for a very wide variety of listening venues when confining to the higher frequencies. The room modes, on the other hand, which dominate at lower frequencies, are highly variable and cannot be predicted except on a case-by-case basis. This adds some insight into Ian's comments that they would prefer to leave the higher frequencies pristine, out of the speaker.
Of course, one cannot really deal with room modes with an equalizer anyway (as Andrew has emphasized), since they are highly dependent upon location. A danger seems to be that one would try to "fix" an isolated local null that the microphone picks up, which is caused not by the speaker or by the sub but by the room. Then, the effect is corrected globally in the EQ, causing the opposite error at most local listening positions.
Dr. Toole's suggestion seems to be, moreover, that some of these very local interference effects are psycho-acoustically innocuous to human listeners, and that the cure can be worse than the disease. I suspect that this is a very real, correct, and practical concern. Especially at high frequency, any local interference effects will exist within a very confined region and making some room-wide compensation for artifacts picked up a one specific point would be absolutely the wrong thing to do.
Without a doubt, it seems that starting with speakers that have a flat frequency response and a smooth transition for the directivity (this is manifest by proxy in the gently / smoothly down-sloped sound power curve, if I understand) is essential. Additionally, the flatness of that curve should be maintained over a wide dynamic range, such that highly dynamic passages are not artificially subdued (compressed) in certain frequency ranges. This is what I interpret as "linearity" (increases in signal strength are manifest proportionally in output strength at "all" master volume levels and frequencies), and the point of the HP woofers. There are just some problems that start at the source, and cannot be corrected downstream. It would seem that quality speaker components and intelligent human engineers are not in danger of being made redundant any time soon.
On the other hand, there are certainly colorations that a room will induce, and to the extent that they are universal/global in origin (or can at least be sensibly averaged over), it also makes good sense that a carefully calibrated EQ could help out quite a bit. There is no shortage of user testimonial that the high end Audyssey filters (XT32 and Sub EQ for example) do quite substantially and very positively impact the delivered in-room experience. It should be emphasized that these systems sample quite a few points to get their suggestion for the global calibration, and are much less sensitive to local dips and interference coincidences than simpler approaches. I'm looking forward to playing around with it when my system is finally set up.
DSQ
