Mid Side?

[guinsu]

I am reading up on mid-side recording, it sounds like an interesting idea to mess with acoustic guitar. Do you need a special piece of gear to set up the matrixing? I have preamps than can invert the signal, cardiod mics and an AKG 414 that can do a figure 8. Do I need something else to set this up?

 

[tmix]

What are you recording to? A computer?

If so all you need is a microphone capable of figure 8 pattern and one that has a cardiod pattern.(and 2 channels of mic preamp)

Tom

 

[guinsu]

I am recording to a Fostex VF-16

 

[kid klash]

I am recording to a Fostex VF-16

Not knowing what preamps (and features on the preamps) you have, you may need to make/use an M-S encoder with your mics, preamps and recorder. Check out the following inexpensive kit :

http://www.paia.com/msdecode.htm

I use M-S mic'ing on drums instead of the conventional "overhead" configuration, and LOVE the spaciousness of the stereo image. It also eliminates many phase problems. It sounds good on acoustic guitar too, IMHO.

 

[mattamatta]

If you have a way to duplicate the fig8 track on your recorder and invert the polarity/phase of one of them that's all you'll need to do. If not, then you'll have to look for something to do that for you.s

 

[guinsu]

Ah, here's where I get bit by the big flaw of the VF16, no way to invert a channel. My preamps (Digimax) can invert, maybe I can run the out of one channel into another and invert it there and just try to match up all the gain. Or, since I use a patch panel, could I take 2 lines out from the panel for a channel and make a custom cable to invert 1 channel of the side? I guess I'll mess with it this weekend.

 

[JPS]

If you have a way to duplicate the fig8 track on your recorder and invert the polarity/phase of one of them that's all you'll need to do.

The M/S technique sounds interesting, but I do have a question. If you duplicate a track, then invert the phase on the duplicate, don't they cancel out the sound from each other? You would get no sound. So, why would this work in M/S recording?

 

[geet73]

The M/S technique sounds interesting, but I do have a question. If you duplicate a track, then invert the phase on the duplicate, don't they cancel out the sound from each other? You would get no sound. So, why would this work in M/S recording?

As far as I know.. the figure 8, or bi-directional, polar pattern picks up sound equally from either side of the diaphragm. When you invert the polarity.. it just shows what the opposite side picked up.

 

[JPS]

geet74

Interesting, but it was my understanding that a figure-8 pattern picks up sound from both sides of the diaphram at the same time. For example, using a figure-8 pattern in a live room will not only capture the sound in front, but also get some of the natural reverb from the rear.

I guess I am wondering if the this duplicate-a-track-and-reverse-phase trick really works for M/S, or is there something else that needs to be done. I reread the link and there was talk about panning one side trak hard left and the other hard right after phase reversal. Don't see how this would change things. I guess I just have to try it and see. I have a C3, C1 and PTLE. That should do it.

 

[Adam P]

The two inverted tracks would cancel each other out when panned to the same spot. Otherwise, they don't. If you have a computer setup you can try it for yourself...take a track (a guitar track would probably be easiest), make a copy of it, and place the two tracks in separate channels. When they're both panned dead center, they cancel eachother out, but as you pan them to the sides, they no longer do.

 

[Harvey Gerst]

The two inverted tracks would cancel each other out when panned to the same spot. Otherwise, they don't. If you have a computer setup you can try it for yourself...take a track (a guitar track would probably be easiest), make a copy of it, and place the two tracks in separate channels. When they're both panned dead center, they cancel eachother out, but as you pan them to the sides, they no longer do.

Huh? You might wanna read this first, before you make more statements like the one above:
http://wesdooley.com/pdf/technique.pdf

BTW, the "M" mic can be any pattern: omni, cardioid, figure 8, whatever.

It's also discussed in the "big thread":

https://homerecording.com/bbs/showpost.php?p=298320&postcount=403

 

[Adam P]

Huh? You might wanna read this first, before you make more statements like the one above:
http://wesdooley.com/pdf/technique.pdf

I understood his question as "won't two identical signals of opposite polarity cancel eachother regardless of individual panning and level? I was under the impression that if you were to take a track (in this case the Side mic), bring it up on a second track and invert the polarity, and bring those two tracks up to equal volume, would they not cancel each other out if they're both panned identically, for example dead center, or hard left? But once you separate them, by panning one hard left and the other hard right, they no long cancel each other?

Not trying to argue at all, I just thought I had an understanding of how it worked but apparently I'm way off, sorry.

 

[Harvey Gerst]

But that has nothing to do with how M/S works. Lemme try to explain it simpler:

If you've read the big thread, you know that a cardioid pattern is made up of equal parts of an omni pattern and a figure 8 pattern. When the two patterns are combined 50/50, you get cardioid. The positive front side of the figure 8 mic combines with the all-positive omni and increases the output from the positive direction of the signal, the front. At the same time, the negative back side of the figure 8 exactly cancels the omni positive wave and you get the familiar cardioid pattern.

Now if you turn the figure 8 mic sideways and add an omni (or anything else) facing forward, AND invert the polarity of the figure 8 mic to another channel, here's what happens:

The positive signal from the omni and the original positive signal of the figure 8 combine to give you all the left side information. At the same time, the positive signal from the omni and the inverted (now positive, but inverted) signal of the figure 8 combine to give you all the right side information.

The cool thing is that it's still a combination of omni and figure 8 on the left, so you still get the usual cardioid pattern if you shut off the right side inverted "S" signal. It makes a cardioid pattern, pointing left. Shut off the left figure 8 pattern and bring up the inverted figure 8, combined with the omni, and the mic is now a cardioid, pointed to the right.

The level of the two figure 8 signals gives you the stereo spread, while the center mic gives you the mono information. If you collapse the stereo image by lowering the "S" levels (or centering the pan pots, the "S" information cancels out or disappears and you're left with a perfect mono signal.

If you think of M/S as Mono/Stereo rather than Mid/Side, it might make things easier to understand: The M mic track (from the forward facing microphone) contains the Mono information, and the two S mic tracks (from the sideways facing figure 8 microphone) contain the Stereo information.

Using a cardioid for the Mid mic results in the two side patterns acting like X/Y cardioids BUT, they collapse prefectly to mono.

The phenomena you're describing is caused by inaccuracies in the signals due to room reflections and differences in the signal due to other factors - it doesn't contain any stereo information. The "S" signal in M/S does - it contains all the stereo information, once properly decoded.

 

[Adam P]

Harvey:

I apologize for my lack of clarification. My initial post was in response to this:

If you duplicate a track, then invert the phase on the duplicate, don't they cancel out the sound from each other? You would get no sound.

I was just trying to explain that the two inverted tracks would not cancel each other out when panned Left and Right. I didn't mean to imply that that the Left and Right channels from a M/S decoder would cancel out if summed to mono. I realize that the portions of the two signals that come from the side mic would cancel, leaving only the signal picked up by the mid mic.

I get the impression that a lot of people believe that two signals that are identical except that they are polar opposites will always completely cancel one another out. Correct me if I'm wrong, but wouldn't total, 100% cancellation only occur if the two signals are A) at the same level and B) panned to the exact same spot?

I realize this is taking the discussion away from M/S technique and more towards simple phase cancellation, but I think that a better understanding of phase cancellation would help some to understand how the M/S technique works.

Thanks for the explanation and the links, as well.

 

[Harvey Gerst]

In a perfect world, yes, they would still cancel out, but the distance between your ears, plus various imperfections in the system, plus room reflections, all combine to leave artifacts floating around due to imperfect mixing in the air.

But the discussion about M/S has nothing to do with that phenomenon; it's all about sums and vectors, due to microphone pattern interactions.

If you take a mono signal, invert one track and pan them wide, you'll hear the signal and it won't be completely canceled out. BUT, if you leave everything else alone, and move the speakers till they're right next to each other, you'll hear more cancellation, mainly because you're taking the room out of the picture. Leave the speakers wide apart, put a mic between them and you'll get X level; move the speakers next to each other and you'll see the mic level drop.

Again, it has nothing to do with M/S; it's a byproduct of room/system problems.