Mid-side stereo with Pro Tools?

[gwailoh]

How does one work with mid-side stereo signals in Pro Tools? Is there a decoder plugin, or is there some other technique required?

Thanks!

[PTUser NYC]

There's probably a plug in, but if you don't have one, you can still do it in Pro Tools. Even if you do use a plug in to actually do it, this discussion is useful to understand what M/S is.

SET UP THE MICS

For M/S recording you use two microphones, with their capsules very close to touching. The idea is to get them to be virtually in the same place, so their outputs will be very similar.

One mic is an omni or cardiod mic, and it faces the center of the stereo image of the sound source you are recording. If you want the source to end up in the middle of the stereo image, point the first mic right at it, if you were hoping to place something on the side, point this mic at the center of the image, which might be up to 90 degrees away laterally from the source, (left or right).

Imagine, for example, using a U87, hung, mounted with the capsule below the body, perpendicular to the floor, and rotating its "point of view" on a horizontal axis, like a camera panning left or right across the horizon. Point the mic at the center of the stereo picture you want to capture. This omni or cardiod mic is the "Mid" mic.

Now for the directional component. A second mic, which is bi-directional (figure 8 type) in nature is used. If the first mic was hung, then the bidirectional mic is right side up, so that the two capsules are as close as possible, one under the other. The bidirectional mic is set so that from a bird's eye view, looking down, the "front" of the bidirectional mic is rotated 90 degrees to the left (counter clockwise) of the center line the first mic is "looking" at. If the "Mid" mic is pointing North, then the bidirectional mic is pointing West, 90 degrees to the "left". This is the "Side" mic.

THE THEORY

At this point, lets consider what happens in each microphone.

When a sound comes from the front of the M/S pair, the Mid mic, (which is pointing right at it), records the sound as a "positive polarity" event. The compression side of each audio wave is causing the capsule to move backwards, and create a positive electrical voltage, the wave in Pro Tools goes up from the zero line. It will make the speaker move out toward our ears first when played, recreating the compression wave in the air that hit the microphone.

As happy as the Mid mic is, the Side mic gets very little of the direct sound here, because its null is pointing directly forward, since we set it up facing "hard left".

Add them together, and you get roughly just the mid mic. Not much from the Side mic.

Now imagine a sound source coming directionally from the leftish side of the array. It makes the same deal with the mid mic (especially if its in omni), but now the bidirectional mic is picking more of it up too. Since the bidirectional mic is pointing left, sounds from the left are being recorded as "positive polarity". The waveform in Pro Tools is going positively first, just like they did in the Mid mic.

Add them together and you get level from both. This time you get basically the same Mid mic performance, PLUS the now present Side mic. Since the capsules are very close together, and the resulting signals are very similar, the result is louder than the center recording was. Things to the left get louder!

Imagine the other situation - where the sound source is on the right hand side. Mid mic is still the same, "positive polarity", but the sound is now coming into the BACK of the bidirectional mic. Now, while the "sound" is still similar at both mics, the mid mic is recording the situation as a "positive polarity" event, but the bidirectional mic is seeing the sound BEHIND IT, and this means polarity inversion from its point of view. That means that the waveform will be very similar to the mid mic's signal, but just "upside down". Look at it in Pro Tools, and see it moves downward from the zero line. This is a "negative polarity" event.

Add these two signals together, and you can imagine that they would cancel eachother out to some degree. The resulting sound is actually softer than the centered recordingwas . This get SOFTER when they move to the right.

When you add the two mics together, everything gets louder as it moves left, and softer as it moves right.

Now imagine if you flipped the polarity of the bidirectional mic's track. Obviously the opposite effect would occur. Things to the left would now get quieter, as the opposite polarities cancel eachother out, and things would get loder as they moved to the right, and the two waves became coherent.

This is how M/S works. With the Mid mic track panned center, and the Side mic panned hard to the left, with a polarity inverted copy of the Side mic track (at equal volume to its upside down twin) panned hard right.

Now things on the left get louder in the left speaker, and things form the right get quiter in the left speaker AND things from the left get quieter in the right speaker, and things from the right get louder in the right speaker.

Stereo!

M/S DECODING IN PRO TOOLS

Put the mid mic on a single mono audio track.

Create two more mono tracks and put the bidirectional "side" mic on each of them.

Flip the polarity of the second side track with audiosuite.

Drag the two side mic audio files into a stereo track, and delete the now empty 2 audio tracks.

Now you've got one mono track with the Mid mic on it, and one stereo track with the Side mic on the left, and the polarity inverted Side mic on the right - all "hard panned" (Panned all the way to the left and right sides).

Bring up the Mid mic, panned center, and then bring in the stereo Side mic tracks to taste. As you bring up the volume of the Side mics, find the sweet spots where convincing realistic stereo images are heard. If the original levels were the same, -6db for the Side mics relative to the Mid mic is one good place to check.

Once you find a good relationship, group the two tracks so they always follow volume changes together, in the same perspective.

VAMP AND FADE

You can see that as the source moves to the left, it gets louder in the left side, and quieter on the right side at the same time, and vice versa for right located sources. We're just combining both of the mono experiments from earlier, and panning each result to one side.

This technique can generate VERY detailed and realistic pictures of instruments in space. It is especially good at capturing ambient sources with a lot of room, like drum kits, roomy amps, string and horn sections, anything where an organic sounding stereo picture of the space is called for.

One additional benefit is that when your mix is summed to mono, the two side mic tracks cancel eachother out, and the mid mic speaks for itself. This mono compatability is rare for most stereo techniques where phase intereference from the two sides of the image can thin a sound out or make it sound comb filtered and plastic when it gets summed to mono.

[gwailoh]

Thank you! What I was looking for was how to flip the polarity of the second side track, since there's no polarity button on the standard virtual channel strip. Much appreciated!

[PTUser NYC]

Thank you! What I was looking for was how to flip the polarity of the second side track

Actually, I kinda figured by the way you aksed - but for other folks, who may not understand it all, I put all the info there.

Polarity flipping:

1) try "INVERT" audiosuite

2) most EQs have a "phase" button. As long as all the tracks are equally delayed, its OK.

[Fokke van Saane]

Waves has a MS decoder.

[Don Geppert]

The INVERT plug is the way to go. The bi tracks will cancel TOTALLY (in mono) if the volumes are equal. It's a beautiful thing.

PTUserNYC ... great detail on the theory and practical of MS technique inside and outside of ProTools. Extremely well written.

[PTUser NYC]

PTUserNYC ... great detail on the theory and practical of MS technique

Sure thing!

It may sound naive, but if we all take the time to help a few people out here, then there will be enough information to search that we can all learn almost anything about recording here on the ol' DUC.

I can't even begin to describe how much I still learn here every year. There's a tremendous amount of excellent digital audio theory floating around in this archive as one good example.

Thanks to all who have opened my eyes along the way, and tomorrow.

[Chris Cavell]

PTUser NYC,

I'd like to clarify one small point. MS micing techniques are not limited to omni and cardioid mic patterns for the M microphone. Generally, the M mic is any pattern b/w cardioid and figure 8 (card, hyper, super, figure 8), but can be used "creatively" with patterns b/w card and omni. A cardioid mid will mathematically translate into coincident cardioid XY after m/s processing.

similarly:
Hyper Card mid -> xy with hyper cards
super card mid -> xy with super cards
figure 8 mid -> blumlein

Pattern translations in the other direction (towards omni) aren't always accurate across the spectrum depending upon the size of the capsule (a byproduct of poor off axis frequency response inherent in large diaphragm microphones that is exaggerated the more omni directional the pattern is).

An example:
It would not be uncommon for someone documenting a symphonic performance to place an m/s microphone abnormally close to the orchestra in order to keep the mic stand out of the aisle. This can create an overly "dry" recording with the standard cardioid Mid pattern...so the engineer will dial through the pattern box toward figure 8 to pick up more ambience from the rear of the hall to compensate.

Cheers,
Chris

[PTUser NYC]

Like I said, something new, every day.

Thanks for the explanation!

I suppose that the reason I've liked using an omni pattern in some cases in the past has been because I wanted more room ambience. I don't do classical recording, so if I'm using an M/S pair for ambience, I'm probably hoping for something surrealistic anyway. More room to compress! In this situation, I don't always used matched microphones either, so again, more "creative use" than audio journalism. For rock drums, I have liked a U87 in omni for the mid mic, and a Royer R121 ribbon for the sides. Obviously, I didn't get much of a "transparent" recording, but it is a great 3D sound.

In a more "truthful" application, like a string quartet, I might want to be looking at some flavor of cardiod for the mid mic, and hope to match the two mics. Heck, two ribbons might be nice too, even if the mid mic would be Bidirectional, especially for a brass section perhaps?

[Chris Cavell]

Absolutely; I just wanted to point out the most traditional approaches. M/S recording is an incredibly fun technique to experiment with. (I wrote this in a thread the other day: a really fun method to play with is XY large di omni's sent through a double matrix...very cool sound...a good "alternative" vocal micing technique, and fun for a drum room mic setup...the lows remain faily centered and most of the width and ambience is expressed in the highs)