I came across a paper which seems to support what I've come to consider an important underlying principle of OMT, although I only came to realize it as such over time -
On the Management of Direct and Reflected Sounds in the 5.0 Surround Sound Reproduction System, authors P. Kleczkowski, A. Król and P. Małecki http://przyrbwn.icm.edu.pl/APP/PDF/128/a128z1ap02.pdfThe paper compares 4 different approaches to reproducing a single isolated source over a 5 channel surround playback array, using source material in which the direct and reverberant components are completely separated from each other (anechoic direct source + reverberant component derived from convolved spatial IR responses) and those components routed to various reproduction channels of the playback array in 4 different ways. Conclusion is that listener preference tends towards isolation of the direct sound component to one playback channel only, with the reverberant component of the sound distributed across
all playback channels.
Inverting that and extrapolating to the recording side of things, the application to OMT is that direct sound arriving from a specific on-axis direction in the horizontal plane is ideally picked up primarily by a single directional microphone of the array while the reverberant component of that sound is picked up by all microphones of the array, including the on-axis microphone in which the direct sound is isolated as much as possible. More specific to OMT, and playback as well though not covered in this paper, is the consideration of pickup of the direct sound component, early reflection components and reverberation components as
three separate entities, the application of that on editing/mixing and in some cases through to the reproduction of these components as separate entities to various degrees. Recorded in such a way, recombination of these differentiated acoustic components can be carried through to the processing and mixing stage, and in some cases through playback itself if via a surround playback array (circling back to the focus of the research paper).
This applies to 2-channel as well as multichannel playback, but it is easiest to conceptualize in the special case of a playback arrangement where each microphone channel corresponds to a specific playback speaker.
So I see parallels here to use of a strongly-directional single forward-facing center microphone (or a tightly-focused/correlated center stereo pair), with the other microphone channels serving as ambient spatial pickup of early-reflections and reverb, and in some cases providing a welcome Hass-effect presence reinforcement, which is what I feel the OCT-like sideways-facing supercards do in a 2-channel mix when a coincident-center pair is providing L/R stereoization imaging.
Of course direct sound isolation in a reverberant space using first-order microphones placed away from the source is marginal to begin with and there will be a significant direct sound leaking into the other microphone channels. Fortunately we need a minimal amount of that for pair-wise capture and reproduction of directional imaging of sources between on-axis points anyway. It would be ideal if that interchannel "bleed" were to occur only within adjacent directional channel pairs, not spread out across more than two adjacent channels. The paper does not address stereo imaging as such, but rather a single on-center-axis source and the "spatialization" of that signal. Sure, spatialization is a
stereo aspect, but is not "stereo imaging" in the sense of reproduction of specific source position.
Further, I see each directional microphone in the OMT array serving as "center channel" for that particular on-axis direction, with all other microphones of the array serving as reverberant pickup array for that center-channel orientation. All axiis work together collectively in support of each other in this way. This requires good management of the pattern overlap between adjacent pairs around the array.
^ In other words, for each primary direction a single directional microphone picks up the direct sound, the immediately adjacent directional microphones pickup (or emulate) early reflections, and all the microphones pickup the reverberant component.