I was trying to experiment with a mid-side on my acoustic guitar and ran into some problems. Tell me what I'm doing wrong.
I used a At 4060 for the center mic and a 414 TL II for in figure 8 for the sides. I went to tape without flipping any polarity switches. I took the returns from tape of the side mic, multed it and ran it into two discrete channels (hard panned). The problem is that whichever side I flip polarity on cancels and the other (the one that is in phase with with the center mic) is boosted. This leaves my balance way off kilter. What am I doing wrong?

the content of your 'side' mic is too similar to the content of your 'mid' mic. Try moving the 'side' mic further away.
I often use M/S in the non-traditional sense by putting my 'mid' mic up close on the instrument, and the 'side' mic really far away, to pick up lots of reflections. This is my 'reverb'. Sounds a lot more natural than using a reverb effect.

Are you sure your 'side' mic is positioned correctly....i.e. the lobes of the figure 8 pattern are perpendicular to the direction of the cardiod mic? Sort of sounds like you might have the figure 8 mic rotated 90 degrees from where it should be.

Experimentation aside,

The only real advantage to using M-S is mono compatibility, thus is extensive use in broadcast, public radio, etc...

If you have no need for mono compatibility go with other true stereo techniques. I recommend Blumlein, x-y coincident card., and ORTIF, you'll get better and more realistic imaging and balance than M-S.

The other adavantage to using M/S technique, other than mono compatibility, is the ability to alter and adjust the "width" of your stereo image by changing the ratio of your mid to your side. If you record each channel sperately (rather than decoding to tape) you can adjust this later when you mix.

Another advantage is that you can get a much more accurate stereo image with two completely different mics (ie not matched, or even the same type or model) in m/s than in any other conicident pair (X/Y, bleumline, ortf, etc). For example, using a beyer M260 and an AT4050 (in cardiod) in an X/Y configuration isn't going to be very accurate or balanced accross the stereo field. Using a Beyer M160 and an AT 4050 (in fig 8 ) in an MS situation will at least give you a realistic stereo image, since information from both mics is equally added to each side of the stereo field.

Another cool thing about MS is that you can use diferent polar patterns for your mid mic to create different "virtual" pairs. For example, if you use an omni as your mid mic, the result is very similar to a bleumline pair.

mike

Hey rawballpower,
just as said before, make sure that your two microphone are ankled at 90 degrees from each other. MS-stereo is a stereo technique that involves two coincidenced microphones, that means they should be as close together as possible. If you move the two microphones appart you'll get something that might sound good, but it's not Mid-side stereo, which was invented by Alan D. Blumlein in 1935.
Before you pan the two matrixed/decoded S-signals to the left and right, make sure that they cancel each other out completely, when panned to the middle. After that pan them and then bring in the mid signal.
I think it shouldn't be too much of a problem that you don't use microphones with identical frequency responses. I've read an article by David Josephson in which he describes an MS-pair that consisted of an RCA 44 for the side-, I think it was, and a Neumann tube mic for the mid-signal. He knows so much more than I do, I trust him.
Good luck with your stereo guitar.

So, what do you guys typically use M/S for? I'll only use it if I'm sure I want something straight up the middle in the mix. I find that other stereo techniques are better suited for panning off center and still maintaining a nice sense of space (albeit not true stereo). So, I'll sometimes use it for FOK or acoustic guitar, but I usually with go with X/Y or ORTF if I want stereo.

used a At 4060 for the center mic and a 414 TL II for in figure 8 for the sides

Try testing out the 414 with a pair of headphones to make sure the pattern is correct. Sometimes mics like that malfunction in sneaky ways. Also make sure the two mics are perpendicular (90 deg.). Close micing isn't the most flattering use of this method. Typically m-s is good for choir, piano, ensemble, overheads, and room micing. It is important to remember that the side channels (+-) must be exactly the same level.

the content of your 'side' mic is too similar to the content of your 'mid' mic. Try moving the 'side' mic further away.

Like Max said. The m-s matrix will only work if the mics are a coincident pair (same distance from source sound). That is the only way for the signals to cancel out/add evenly.
This is my favorite stereo micing technic next to the Blumlein method. I especially like micing the entire drum kit in a dead room with it.

Incidentally, we are offering to sell supped-up Electrical Audio M-S Matrix boxes on a build-to-order basis. A bunch of people have expressed interest in buying our boxes over the years, so I figured I'd start making them.
If your interested PM me.

I've used MS often (close micing) and I've fought the skewed image some. I had attributed it to the fact that the M and one side of the S are basically in phase and they sum, whereas the inverted S and M form a difference... thus shifting the image. But now I'm thinking it's related to close micing.

any thoughts?

I've used MS often (close micing) and I've fought the skewed image some. I had attributed it to the fact that the M and one side of the S are basically in phase and they sum, whereas the inverted S and M form a difference...

Both S and inverted S form a sum. Just a different sum (left or right).
Both should have the same energy. A bidirectional mic as the side mic perpendicular to any mic will have the same level when added in or out of phase. This is of course if the sound source has equal energy on each side of the mic setup.