Interesting examples of the application of "Blumlein shuffling" have been posted by Voltronic over in the
Acoustic Recording Thread, followed by my analysis and some discussion of what's going on with it. To avoid the oversaturation of that thread with a discussion of how shuffling might be applicable to multi-mic arrays such as OMT setups, I'm continuing that discussion here.
Here's where I left off over there-
Thanks for starting the conversation on shuffling with your examples. Super interesting stuff. Hope I haven't gone too deep into this for the others following.
A few of my conclusions-
- I'm now convinced that a not overly wide spacing in combination with shuffling is an optimal arrangement for a stereo pair of omnis when used on their own.. in most cases - there are always exceptions and I think those tend to arise in taper situations more so than typical recording scenarios. The oddity of superwide 20-30 foot omnis splits sometimes working great for PA amplified stuff without suffering a "hole in the middle" is an example of that.
- This is great news for tapers! Less wide spacings are far, far, far more practical, achievable, and less-imposing.
- Application of shuffling may make for an improvement of any taper recordings made with a coincident or near-spaced pair. That's most of them.
- Thinking about how it might apply or not to multi-microphone arrays, meaning those which include a microphone, pair, or more mics placed in the center between a wider spaced pair.
Forgive one additional indulgence in thought experiment about that before I let this go.. actually, strike that. Probably best if I take that over to the OMT thread.
My first thought is that in a multichannel mic array which is to be summed to stereo (specifically 3-position PAS and OMT) we'll still need to get the microphones of the channels that are going to be summed far enough apart to mix without conflict. The mics feeding the channels to be summed either need to be
close enough together or far enough apart* to avoid potential problems that can arise when positioned somewhere between - the sort of all-or-nothing, inverse Goldilocks problem of mono-compatibly. Shuffling may not be of help with that. We can experiment to find out of course. And fertile ground for trying that is remixing multichannel OMT recordings in which less than the desired amount of spacing was used for whatever reason.
Second is how and where shuffling might be applied. To keep it simple, let's assume a 3-channel arrangement which consists of two omnis with a cardioid in the middle. The addition of the cardioid in the middle, along with providing other benefits, fixes many of the things that are problematic with a wider omni spacing, but it's introduction also
requires that the spacing be made wider. Alternately it might be a 3-mic triplet of angled directional mics that are able to use somewhat less spacing. Either way, in this thought experiment I'm using a narrower spacing than I'd would otherwise prefer, and attempting to compensate with the application of shuffling.
I see three potential ways of applying shuffling to the raw 3 channel L/C/R output from the array when mixing that down to 2-ch L/R stereo:
1) Shuffling is applied to just the L/R outside mic pair prior to mixing. ..Or to to the L/R mixed stereo output. Either way the result should be the same as the contribution of the single center mic is made identically to both L/R channels, effecting only the Sum not the Difference. This is the essentially the same as adding shuffling to any existing recording.
2) Shuffling is applied to L/C, and C/R, but not to L/R.
3) Shuffling is applied to L/C, C/R, and to L/R.
#2 and #3 are the interesting ones unique to a three channel array. Fun to think about what's going on with those.
Might it address the potential for imaging conflicts in a 3-channel, 3-position array from having three different SRAs in play in such an arrangement? One for the L/C pair, another for the C/R pair and a third for the L/R pair. As posted about here an in a few other threads a few months back, we can arrange for the L/C and C/R imaging segments to "hand-off smoothly to the other" by way of "steering" their SRAs via forward spacing of the center mic. If all the mics were placed along a single line those two SRA's would overlap each other in the center rather than the inner edge of the two SRAs lining up to affect a smooth hand off from one to the other across the center. The problem is there's also an SRA associated with the L/R mic pair, and that SRA is considerably narrower than the combined SRA of L/C + C/R.
The question is, does shuffling alter SRA? If only at low frequencies where the shuffling filtering is active? Can we use that to our advantage in this situation? Can we get the combined SRA L/C + C/R closer to the SRA of L/R with a clever application of shuffling?
I'll have to think about this some more.
