M-S (mid-side) recording is a two-mic approach to recording orchestras, jazz combos and similar self-balancing ensembles. M-S goes a long way toward reducing the “hole-in-the-middle” that can result from other two-mic stereo techniques, while affording extensive control over the width of the ensemble.
In brief, the M (or mid) mic exhibits a bi-directional or cardioid pickup
pattern and faces the front and center of the orchestra. The S (or side)
mic exhibits a bi-directional pickup pattern and is oriented at 90 degrees
to the M mic, facing the side walls.
By mixing the output of the S mic with positive polarity (“in phase”)
with the M mic to derive the left channel, and mixing it with reversed
polarity (“out of phase”) with the M mic to derive the right channel,
the apparent width of the sound stage can be adjusted to be wider or
narrower simply by varying the overall level of the S mic.
This technique provides results consistent with the X-Y technique
(crossed cardioid or bi-directional mics) with two additional
1. instruments in the center of the stage are not picked up off-axis as
they are with both X-Y mics (which are pointing away from the center);
2. the recording is 100% mono-compatible (e.g., for FM radio
transmission), since in mono the +S and -S signals are added together
and cancel each other out, leaving just the M component.
I have used the M-S technique extensively with good results. The first
example on this page illustrates the technique used to record Tafelmusik, a
world-class Baroque orchestra:
Here's an excellent technical introduction to M-S.
The authors also provide a comprehensive and comparative review of various stereo miking techniques here.