Adam Dean Ellis

If you have spent any time at all investigating the various stereo microphone techniques you have undoubtedly come across the concept of M/S recording. The technique requires one cardioid mic and one bidirectional, or figure eight, mic. The microphones are placed so that there are three “lobes”, two focused to the sides ( the figure eight mic), and one focused forward ( the cardioid), and fits in between the side lobes, e.g. mid.

Here is a picture borrowed from electronic musician.

The stereo image is produced by manipulating the phase differences and similarities between the two sources.  But first, lets dig into our knowledge of basic acoustics and remember that two otherwise identical sound waves that are 180˚ out of phase will cancel each other completely. Acoustic annihilation! If the sound waves are only partially out of phase there would still be some cancellation, rather a net loss of acoustic energy. If the two waves are completely identical and in phase the would sum, for a net acoustic gain.

As we can see from the picture above, sounds originating from a frontal source would arive at the two microphones at different times, thus puting the electrical signals generated by the microphones out of phase.  We also have to consider the indirect sounds, reflections off of various surfaces in the room, that are arriving even later than the direct sound that give us our sense of space.

To take advantage of the various phase interactions, and produce a stereo image, we need to sum the Mid and Side, and send this new signal to the left channel. For the right channel we need to sum the Mid and Side again, but this time we need to reverse the phase, or flip the signal 180˚. The effect is like taking a positive number and adding a negative number to it…

Left = M + S

Right = M + (-S)

The center of the image is of course regular ole stereo phantom center and, magically, it consists of the sound source that was directly in front of our M/S pair.

Since we are building a stereo image out of the various phase information our microphones have captured, there is no reason why we could not deconstruct a stereo image via the same kind of phase information, manipulate or change it, and put humpty back together again. Mastering engineers do this frequently in order to apply processing that they do not necessarily want to affect the entire mix. For example, a mastering engineer is given a a mix of a pop song, and for whatever reason the vocal is way to loud, or is very sibilant. Since it is a pop song, we can assume that the vocal is panned dead center. The solution may be to use M/S processing to separate the stereo recording into Mid and side channels, and then simply lower the level of the mid channel, which is the center of the stereo image. When its put back together into a new stereo recording the vocal will be quieter, but so will everything else that is panned dead center. If the snare was in the perfect place in terms of balance we may have created more problems in our attempt to fix the vocal. Still, if lowering the level does not do it, try a multi-band compressor, an EQ, or anything else your heart desires. One more thing to note is that when you lower the level of the mid channel you will also increase the perceived width of the stereo image, and vice versa.

There are alot of ways you can go about acheving the M/S process including a variety of off the shelf plugins that just do it for you. If it makes your life better, please,  go that route. However, you can achieve the same, if not better results, with the fetures provided in the most modest version of Pro Tools. Although this article uses Pro Tools as an example, I feel certain that you can pull the same thing off with most any multi track D.A.W. platform that allows internal busing.

Once you have your session created in Pro Tools you will need to create one stereo audio track, and five stereo aux inputs., then label and color code the tracks as follows.

1. (ST audio track – Blue) “Source”
2. (AUX input – Green) “Mid”
3. (AUX input – Green) “Side”
4. (Aux input – Yellow)   “Left”
5. (AUX input – Yellow) “Right”
6. (AUX input – Red) “Decoded

You can use any colors you would like, just make sure to group the tracks as we did above. Next insert one of the time adjuster plugins – under insert–>Multichannel Plugin —> Delay —> Time Adjuster (short) – into each of the five AUX tracks..  You can use any stereo plugin that will allow you to reverse the phase (ø) of either the left or right channel independently.

From here we need to begin to route our tracks to their appropriate destinations, and this is where it can get a little tricky at first.  However, we have already arranged our tracks so the signal flows basically from left to right. Start by assigning the “source” track to “bus 1 – 2.” Next, set the input of both the “Mid” and the “Side” tracks to the same bus. Now, your source track should be feeding both the “Mid” and “Side” tracks.  Now, assign the output of the “Mid” track to “bus – 3.” Please not that you are assigning a stereo bus to a mono output, which results in the summing of the stereo input, or to put it into language we understand, this is Left + Right = M.

Next assign the output of the “Side” track to”Bus – 4,” again summing the channels of the stereo source track. However, this go around we will use the time adjuster plugin to reverse the phase (ø) of the right channel. This equates to Left – Right = S. If you’re still with me, this should look an awfull lot like the expressions M+S = L, and M – S = R. Essentially the same thing is happening.

For our “Left” and ” Right” tracks we will set the input to “bus 3 – 4,” the outputs to “Bus 5 – 6,”and pan both sliders for each track left and right respectfully. Next, via the Time adjuster plugin, reverse the phase of the right channel of the “Right” track. Essentially, we are taking the M and S and re-encoding them to a new stereo pair.

Finally, set the input of the “decoded” track to “bus 5 – 6″, and the output to “interface 1 -2.” This is not really doing anything beyond providing a single fader for the final stereo output. You will find that piling all these tracks on top of one another will require to you bring down some of the starting levels. I usually begin with the source tracks fader at -3, and the Mid, Side, Left, and Right at -6. I generally leave the “decoded” track at unity.

Notice that even though we inserted the trim plug in all but the source track we only used two of them. Plugins, regardless of how hard they are working, will add latency to the signal, which will cause phase shift. The effects are usually very painfully obvious. Give it a try by removing the plugin from either the “mid” or the “side.” Also, if you command – click the numbers below the faders you can cycle through information such as peek level, fader level, and latency which is represented in terms of samples. Our first track should read zero, and the rest should read about 4, meaning that these tracks are about 4 samples behind tracks with zero latency. Regardless of what it is, if tracks 2 – 6 have a matching number, you’re good.

Now you can import the stereo track of you choice, and if you have never heard it before, prepare to be amazed. Have fun. But while going crazy and pushing the limits can yield some very productive results, attempt to make subtle adjustments once you get it out of your system. Try to bring impact to the music, or change the blend of the instruments, while leaving the sonic quality intact. If you are using a commercially released album you may be hard pressed to improve it. Still, breaking down a stereo mix into its true anatomy can be very revealing. I find myself in mix and tracking sessions often using the perspective I gained from just playing around with M/S processing , and undoubtedly my mixes improved.

And now that you have been through all that, here is my simple thanks to you.

Free Download: M/S Processing Template for ProTools 7.4

Now go play!

Adam