Optimal source with regards to room acoustics + bandwidth

[thadman]

Loudspeakers are often discussed separate from the room. However, this may lead to non-optimum solutions, as they both constitute the total system.

There are a sundry of theories claiming optimum response.

Below the room fundamental, source location is irrelevant as room pressurization is uniform. Dipoles do not apply as they constitute a velocity source.

Within the modal region, distributed sources (monopole or dipole) have been observed to yield a subjective, optimum response. Arrays (ex. DBA) may offer an interesting alternative.

Above the Schroeder frequency, disagreement persists. Directional, pressure point-sources (ex. waveguides) as contributed by members such as Geddes and Le'Cleach have been observed to yield a subjective, optimum response. Directional, velocity point-sources (ex. dipoles) as contributed by members such as John K and Linkwitz have been observed to yield a subjective, optimum response.

The room's interaction with the source is not constant and as such, I believe a combination of the aforementioned formats may be required to yield the globally optimum solution.

Dr. Geddes has advocated a transition from monopole to distributed monopole in the LF (<100hz).

Linkwitz has advocated a transition from dipole to distributed monopole in the LF (<50hz).

However, perhaps a non-traditional alignment may yield further improvements in response.

Reflections (ex. floor bounce) have been observed to yield a non-optimum response as the distance from obstacles approaches the reproduced wavelength (~50-500hz). Digital filters (ex. FIR) do not lend themselves to this type of problem as it constitutes a global irregularity in response. Furthermore, due to the restrictions of the space, waves do not expand as they would in free-space. An analog can be taken akin to the pressure wave expanding through the classical horn. A transient-perfect response contributed by a point-source may not be possible throughout such bandwidths. Assuming this to be true, we must reform our idea of the optimum transducer if we are to approach the global optimum solution.

I might advocate a transition from distributed point-source / line-source --> line-source --> point-source.

Any thoughts?

[thadman]

C'mon, I'm sure members have considered this topic before