Mid-side stereo technique

When using a Mid-Side technique (cardioid facing forward and figure-8 above with diaphragm at 90 degrees, so picking up the sides), it seems like you might have more control over the mix by setting the relative volume of sources on the left and right. Would this be true? I want to record three vocalists live but I'm not sure we would get the blend perfect, and it would be nice to raise or lower the volume of the person on the right or left side.

Also, would it be safe to assume that any pro mix/mastering studio would be able to use the MS matrix, and is that necessary to create the stereo image?

How do you think this technique compares to recording three vocalists with a spaced pair of mics? How about two cardioid mics? I want to pick up some room ambience.
Kristin Andreassen

Reply: The mid-side (MS) method of stereo miking uses a cardioid mic aiming straight ahead, coincident with a bidirectional mic aiming to the sides. (Actually that cardioid mic can be any pattern, but cardioid is most commonly used). The mid-side stereo microphone technique lets you adjust the stereo spread (stage width) -- anywhere from mono to wide stereo -- either during recording or during mixdown.

You send the mid and side mic signals into a matrix box. Here's what it does:
The matrix is a sum and difference network. It includes an M/S ratio knob.
Mid + Side = left signal. Mid - Side = right signal. So the output of the matrix is regular left-right stereo. MS simulates a coincident pair of supercardioid mics with a variable angle between them to control the stereo spread.

To change the spread during the recording, connect the stereo-mic outputs to the matrix box and connect the matrix-box L–R output to the recorder. Use the stereo-spread control (M/S ratio) in the matrix box to adjust the stereo spread.

To alter the spread after the recording using a matrix box: Record the mid signal on one track and the side signal on another track. Monitor the output of the recorder with a matrix box. Back in the studio, run the mid and side tracks through the matrix box, adjust the stereo spread as desired, and record the left and right outputs.

To alter the spread after the recording using a DAW:
1. Record the mid mic on track 1; record the side mic on track 2.
2. Copy or clone track 2 to track 3. Be sure the waveforms are aligned.
3. Pan track 2 hard left; pan track 3 hard right.
4. Reverse the polarity of track 3 or use an “invert polarity” plug-in.
5. Group tracks 2 and 3 so their faders move together.
6. To change the stereo spread, vary the levels of tracks 2 and 3 relative to track 1.

If you increase the amount of side signal, (that is, decrease the Mid/Side ratio in the matrix box) the stereo width increases, and vice versa. This does not affect the relative volume of the sound sources, just the perceived position of their images between the two monitor loudspeakers.

I'd think that any pro studio would know how to use a matrix box. If not, this article tells how. Not all studios have a matrix box, though.

Suppose you're recording a trio of singers positioned left-center-right in front of the stereo mic array. You'd run the mid-mic and side-mic signals through an MS matrix, either during recording or playback. If you use the matrix while recording, you need to set the M/S ratio to achieve the desired stereo spread over your monitor speakers before recording. If you use the matrix during mixdown, you'd record the mid mic on one channel, record the side mic on another channel, and feed both channels into the matrix during mixdown, when you can adjust the stereo width as desired. There might be an MS matrix plug-in available for DAWs.

Suppose you have set the stereo spread so that you hear one singer from the left speaker, one in the middle (phantom center image), and one from the right speaker. You record the matrix outputs -- left and right signals -- on a stereo track in a DAW. By varying the left-right panning of that stereo track, you can make the left or right singer louder; however the image of the center singer will shift left or right. I think that if you turn up the mid mic signal before it enters the matrix, the center singer might get a little louder, but the stereo stage will become narrower (that is, the images of the side singers will move toward the center). It's a compromise.

I think it's a lot easier just to give each singer a mic, pan each mic where desired, and adjust the level of each mic as desired. If you want, add some high-quality stereo reverb to each track so that the singers sound like they're in a shared space.

If you use a stereo mic instead of individual close-up mics, the best way to control the balance of the sound sources is during recording, by moving each singer toward or away from the mics until the monitored balance sounds good. You could still adjust the stereo width during mixdown, but you'd need to determine the balance during recording. Any stereo mic technique picks up whatever balance is happening at the mics; the balance is hard to change after the recording is done.

Like any coincident-mic technique, Mid-Side gives sharper imaging than spaced pair for off-center sound sources. However, if the spaced-pair mics are not too far apart (maybe 2 to 3 feet), and the singers are hard left, center and hard right, the image sharpness would be about the same as with Mid-Side. Spaced pair is of course cheaper than MS, and spaced pair gives a warm sense of ambience.

With a spaced pair, if you turn up the left mic to make the left singer louder, the center singer's image will shift partway toward the left between the monitor speakers.

Here are three ways to place two cardioid mics to record in stereo. Each method has pros and cons.

1. One directly over the other, angled apart about 120 degrees (each mic 60 degrees either side of center). This gives sharp imaging and is mono-compatible. It's tricky to mount the mics that way, though.
2. Angled apart 90 degrees (45 degreees either side of center) and spaced 11 inches horizontally. This gives pretty sharp imaging and some sense of "air".
3. Aiming straight ahead and spaced about 2 to 3 feet apart horizontally. This gives fairly sharp imaging, and the best sense of "air" or "warmth." It also captures the most natural timbre of the center singer's voice, because the mics are aiming at the singers.

I'd start with the mics maybe 3 feet away to pick up some room ambience. Listen to the playback, and if you want to hear more of the room, move the mics about a foot farther and try again. If the sound is too reverberant or muddy, move the mics a foot closer and listen again. If you want the freedom to add digital reverb during mixdown, mike close (maybe 1.5 feet) so that you don't pick up the room. Good luck!