Listening to
Thelonious's comparison of Neumann KM184 / Schoeps MK22 brought to mind the somewhat different way I now think about the roll of rear facing mic channel(s) in OMT when mixing to 2-channel stereo. That's certainly how the majority of folks are using OMT, including myself for the past 10 years or so. Although I really love it and miss it, its been quite a while since I've had a surround system suited for playback of these recordings in action.
Originally, I included a rear-facing mic to feed the separate surround channels, later expanded to a rear-facing pair, which worked really well for the purpose. The tangible sense of front/back depth, the depiction of the room geometry, realistic audience reaction wrapping all around, and a far more realistic sensation of bass were all very compelling when everything was arranged correctly. I found that keeping as much direct front arriving sound as possible out of the surround channels was a big part of getting it to work right. Bear with me here, as I realize I've mentioned that a number of times in these threads.
Also mentioned numerous times was that I was super pleased to find many of the advantages translated to a 2-channel stereo mix down as well. Although the tangible surround playback aspects obviously were not the same, the inclusion still made the recordings sound far more natural and convincing to me, generally making for a more compelling listening experience. On top of that, the ability to dial in just the right amount, along with adjustment of balance and other aspects, made recording in that way super useful for me even after surround playback was no longer my primary goal. The advantage was the "sound" I was able to get by recording in that way, but also in the increased flexibility it provided in achieving a good result in general.
I still think of the rear-facing pair in terms of allowing for the addition of "front-back dimension, room and audience" to taste, but I've also now come to think of it as
allowing me to adjust the polar pattern of the primary forward-facing near-spaced L/R pair, afterward. I'm using supercardioids in both those positions to reduce acoustic crosstalk between them as much as is practical, which also keeps the array size manageable in terms of necessary spacing/angle between mics to retain a good hand-off from channel to channel moving around the array. Arguably a smooth hand-off around the sides from front pair to rear pair probably matters more with surround, but seems to me it helps in a stereo mix too, even though technically that dimension gets flattened.
Here's the thing-
Ignoring the other channels of the array (in my case a coincident center pair and a wide-spaced omni pair, which we'll just mute for this thought experiment) and starting with just the forward-facing near-spaced L/R pair,
bringing up the rear-facing pair essentially serves to broaden the polar pattern of the primary forward facing L/R mic pair. That's the main thing I hear. It's somewhat like switching a straight forward facing stereo pair to a more open pickup pattern. There are some important differences, but that's the main thing. I gain control over choice of L/R pickup pattern after the fact. Is it exactly the same? No, the forward focus of the front supercardioid pair remains essentially unchanged. The ideal spacing/angle between that forward supercardioid pair remains the same. It's forward focus remains the same. Moving toward the center the polar patterns remain supercardioid, but moving out toward the sides the effective polar pattern broadens as the rear-facing channels are raised in level, basically as much as is desirable. The same goes for the rear-image, still two supercardioids with regard to the imaging region between them, but the pattern out to the sides opens up more.. as much as I want. Can even go beyond "pseudo subcardiod pair, or pseudo jecklin-disc baffled omni pair" to the pattern facing more rearward than frontward if desirable for whatever reason. Its all a consequence of the mic angles and supercardioid polars and how they combine.
Unlike switching to mics of another polar pattern, the front and back imaging geometry doesn't change much, only off-axis pickup does. Perhaps this observation is either arcane or obvious to you, but its been rather insightful to me.
Thinking through it further- There is certainly some potential for comb filtering or phase-smearing when the foreword and rear-ward directed supercards on each side get mixed together. Essentially we have a stereo-mono-compatibly issue on each side, by mixing together the equivalent of a near-spaced stereo configuration turned sideways. The spacing between the front and rear facing pairs was originally intended to impart a slight time-of-arrival difference that increased the perception of front/back separation and helped keep the sound from in front from bleeding around back, and allowed for the use of more level in the rear-channels. That potential front/rear mono mix down compatibly on each side could be minimized by placing the rear-facing pair coincident with the front-facing pair. Which would also make the setup more compact in the fore/aft dimension. But I've not had any problems with that, I think because these four channels of the array were setup so as to "hand-off" smoothly from one to the other it works when folded down. And I suspect the front/back time-of-arrival difference imparts some attractive binaural cues in the stereo mix and makes for a more diffuse portrayal of reverberance.
So I'm sticking with the near-spacing of the front/back pairs, but other folks using rear-facing channels in their OMT setups intended for stereo mixdown might try placing the front and rear facing pairs coincident with each other.
