Skip to main content

Just curious if anyone has messed around with stereo or surround recording using 3 mic / 4 track "double MS".

Comments

RemyRAD Wed, 04/05/2006 - 22:05

Well I have used 2 pair of MS microphones in the past. I don't do surround, yet. I haven't really tried creating a MS surround microphone system but there is the Holophone microphone system that I believe offers a similar capability?

What is your desired application for said microphone system? I think a Blumlien crossed figure 8 set up might be more applicable to what you desire?? Of course that is just for stereo again but for surround you may want to try 4 Omni microphones where 2 are in phase and 2 are out of phase in a pseudo figure 8 Blumlien style configuration and then you would be ready for surround? Keep them close and/or on top of each other if you try that.

Coming and going
Ms. Remy Ann David

RemyRAD Thu, 04/06/2006 - 09:50

aracu? First you're talking "surroundsound" then you're talking about "mix down to stereo afterwords". You are obviously desiring to experiment and you should but first you need to understand how MS Works. Having a surround signal and collapsing into stereo isn't what I would exactly call practical or purpose driven? If you want an MS based surround, then I would use 2 pair of MS microphones, one facing forward and, one facing to the rear of the hall or room that you are recording in. I think you could also try a single middle Omni, facing upwards towards the ceiling with 2 pair of figure 8 directly below your Omni middle while utilizing your console or software for the matrix decoding? Again, your figure 8's I believe would cancel when collapsed to stereo? It certainly gets my brain percolating about attempting what you have suggested for a strictly surround-based recording?

Which way did he go?
Ms. Remy Ann David

Simmosonic Thu, 04/06/2006 - 17:20

You can create back and front MS using three capsules, recorded to three tracks, and then decoded to four channels of surround (or whatever).

First, you'll need two bidirectionals and an omni, similar to Remy's suggestion. here's the method:

One of the bidirectionals faces forward at the ensemble being recorded. The other bidirectional faces sideways, so the ensemble is in the null and the + side of the capsule is facing the left side of the ensemble. Now you need to incorporate the omni. All three mics need to be coincident and as close together as possible. If you have a small end-address omni (DPA, Earthworks, etc.) put it in the middle, between the two bidirectionals.

Now, here's the decoding method:

1) The sideways facing bidirectional mic is the S signal.

2) Mix the omni and forward-facing bidirectional to get the desired polar response. With those two responses (omni and bidirectional) you can create just about any polar response you want. For example, with each at the same level you'll get a cardioid response. Increasing the relative level of the omni will create polar responses between cardioid and omni (i.e. wide cardioid, subcardioid or hypocardioid, depending on who you talk to). Increasing the relative level of the bidirectional will create responses between cardioid and bidirectional (e.g. supercardioid, hypercardioid, etc.). All of this maths assumes you have the gains and sensitivities of each mic adjusted correctly in the first place...

3) The signal created in (2) above is the M signal for the forward stereo signal. It can now be combined with the side facing bidirectional to create an MS stereo signal of the ensemble. You can keep adjusting the balance between the forward-facing bidirectional and the omni to get just the right stereo coverage of the ensemble.

4) Now, taking a split of the omni and the forward-facing bidirectional signals, reverse the polarity of the forward-facing bidirectional and repeat steps (2) and (3) to create the rear-facing MS pair. But, of course, you'll adjust the balance between the forward-facing bidirectional and the omni to create the desired stereo coverage of the rear 'room' sound, rather than the ensemble sound.

That ought to give you four signals (2 front, 2 rear) from three capsules recorded to three tracks. When collapsed to mono, you'll be left with the omni capsule signal plus whatever the difference in level is between the forward-facing bidirectional in the front and rear channels.

So, you've got a mono-compatible coincident trio for recording in stereo and/or surround. Nice! This is conceptually what the Soundfield mic does, although it adds a bidirectional facing up/down as well, adding more flexibility.

But, personally, I don't think coincident is the way to go for surround recording...

aracu Thu, 04/06/2006 - 19:52

How about if you used a stereo mic for the two bidirectionals, and
mounted the omni upside down, slightly over them, to avoid blocking
the capsules. Can't resist asking how you would mix in a fourth
bidirectional mic pointing up and down, recorded to a fourth track.
Do you think it would work well with an AKG 426 stereo mic and
one (or two) Senn 800's, or would those mics be too different from
one another.

Simmosonic Fri, 04/07/2006 - 00:25

aracu wrote: How about if you used a stereo mic for the two bidirectionals, and mounted the omni upside down, slightly over them, to avoid blocking the capsules.

What do you mean by mounting the omni upside down? What mic are you thinking of using? Is it an end-address or side-address?

The real problem you'll have is due to the fact that you can't possibly get all three mics in the exact same place, so you'll have to be very careful when aiming the array so that the sound you're trying to capture is perpendicular to all capsules. But if you've got a stereo mic, then by all means use it...

aracu wrote: Can't resist asking how you would mix in a fourth bidirectional mic pointing up and down, recorded to a fourth track.

Resist! Resist!

Oh, alright then... The concept is the same, but now you can change the direction of the mic, so it can 'see' up/down as well as left/right and front/back. You combine the up/down bidirectional with the omni to create the desired up/down response, then MS it with the S bidirectional.

On the simpler version of the Soundfield, this allows you to make the mic either side-address or end-address.

By the way, the Soundfield microphone overcomes the problem of getting all the capsules in the same position in a very clever manner. Instead of using three bidirectionals and an omni all crammed into one space, it actually uses four subcardioids (IIRC) in a tetrahedral array. With a bit of clever processing, this can be used to create the equivalent signals to three bidirectionals with an omni in the centre, and they are all perfectly coincident. Smart...

aracu wrote: Do you think it would work well with an AKG 426 stereo mic and one (or two) Senn 800's, or would those mics be too different from one another.

Well, because it's MS, it is a bit more forgiving of the individual mic tonalities and so on. I mean, they will be there, and they might result in something like the sides being brighter or duller than the centre, but you won't get any left/right imbalances; any differences will be distributed the same on both sides.

All I can suggest is that you give it a try... All this theory is wonderful, but if it sounds bad it is bad.

aracu Sat, 04/08/2006 - 21:04

And you can't monitor very easily, if at all, what your possible end
results will be. It would be nice if there was a Soundfield type monitoring system that you could plug any mics and pres of yout choice into, being able to adjust the parameters to see what your post options would sound like for that particular situation.

Simmosonic Sat, 04/08/2006 - 23:05

aracu wrote: And you can't monitor very easily, if at all, what your possible end results will be.

It would not be hard if you were using a laptop or digital mixing system that had MS decoding available for individual track pairs...

You'd have to start with a 'preset reference' mix, such as the MS ratios at 1:1, and the omni/bidirectional balance for the M signal also at 1:1 (i.e. cardioid). That'd be the right starting point and would allow you to put the mic array in the right position to capture the music itself. Then you can work from there, fine-tuning the M response and MS balance for each pair as required. The rear pair would not have to mirror the front pair, by the way. The front pair is set up to capture the ensemble correctly, in terms of width and depth and so on. The rear pair is set up to capture the ambience of the room correctly, it could be considerably wider or narrower than the front pair, and it could use a different M pattern. Whatever floats your boat...

aracu Sun, 04/09/2006 - 19:01

Simmosonic, your brain has really been working overtime with such a detailed description! Great info. I will try it with my 4 channel Sound Devices. It allows you to listen to a monitored decoded signal of one pair at a time, of two MS pairs, although without being able to adjust the monitored ratio of M to S. It also has Soundfield B-format monitoring,
meant to be used with a Soundfield mic.