A Practical MS Decoder Circuit

by Jules Ryckebusch

3D"MS The =20 schematic diagram of the MS Decoder may look complicated but is actually = =20 quite simple. Both the Mid and Side signals are initially buffered by = unity =20 gain inverting buffers formed around IC1b and IC2 b. This is necessary = for=20 two reasons, first to ensure enough drive current for the following = sections.=20 Second, the final summing sections invert the signal, to achieve zero = phase=20 shift through the unit, one more stage of phase inversion is = required.

The Mid signal goes to level control potentiometer R15. It is then = fed =20 equally to the left and right summing amplifiers, which are formed = around the=20 two sections of IC3. The use of R15 (and R16) is to allow adjustment = of the=20 relative levels of the mid/side levels independent of the mic-pre = gain=20 setting. This is useful for directly feeding a recording device.

The Side signal has a little different path. After initial buffering, = it=20 is fed to the right summing amplifier via one section of dual = potentiometer =20 R16. It is also fed to a unity gain inverter formed by IC2a and its = associated =20 resistors. This inverted signal goes to the other half of the dual = ganged =20 potentiometer section. Then it is summed into the left channel via IC3a. = One=20 section of IC1 is not used. Both of its inputs are tied to ground to = keep any=20 thing strange from happening.

R17, R18, and C1-4 are for power supply filtering. The circuit will = run on=20 +/-9 to +/-18 VDC.

OK, so what are R19 and R20 doing? That is an interesting question. = These=20 are there to load down R16 so that when the levels of the Side signal = are=20 adjusted the potentiometer gives the same feel as the Mid level = adjustment =20 potentiometer R15. The op-amp summing sections are virtual grounds. This = =20 means that signals entering the op-amp see a load equal to the input = resistor. =20 All of these are 10K resistors. The Mid potentiometer feeds two of = these so=20 it is presented with a 5K load. To make the load on the Side = potentiometer =20 the same one additional 10K resistor tied to ground is added to each = wiper so=20 it =94sees=94 5K also. Here is the other interesting thing that all of = this =20 causes. Because all the potentiometers are linear taper, loading them = down =20 with a load much smaller than the resistance of the potentiometer causes = a=20 change in the characteristics of the potentiometer. This in essence, = makes=20 them respond more logarithmically than linearly, which is the way we = hear=20 anyway which is a good thing!

Alternatives:

3D"MS Well, =20 you can always break out the patch cords if your console supports phase = =20 switches. Be careful, some only have the phase switch on the channel = microphone =20 pre-amp, not the channel itself. Simply patch out of the preamp on the = =20 first channel and into the line input on the next channel. Depending on = the=20 patch bay, you may need to use a half patch technique, which is beyond = the=20 scope of this article. Before panning hard left and right, center both = and=20 adjust the level on the second channel until the side signal is = completely =20 cancelled. This means that you have the signal levels matched. Now pan = one=20 hard right and the other hard left. Most multi-track audio programs = allow you=20 to do this too. I use Studio Vision Pro and Pro Tools but I know most = others=20 have similar features. Take the side signal pre fader and send it to = an=20 effects buss. On the original channel pan it hard right. On the = effect buss,=20 phase invert the signal and pan it to the opposite channel as the = original.=20 Voila! You are done. There are usually many ways to skin a cat.

So why build or use an outboard MS Decoder? The real benefit from = having a=20 dedicated MS Decoder is that you can listen to the effect in real time = while=20 you position the microphones and make adjustments before recording. I = have in=20 the past, recorded the Mid and Side signals directly only to find out = they=20 really didn=92t work together while decoding during playback and mix = down. We=20 were recording a Leslie and fortunately after taking my lumps, I was = able to=20 re-record it. I had to buy a round of beers over that one.

Use:

Just patch the Project r MS decoder between your microphone pre-amps = and=20 your mixing board. You can also run the signal straight into a DAT or = you=92re=20 A/D converters if recording into your computer. The key to good = results is=20 the ability to monitor the stereo signal while recording. Then you = can hear=20 exactly how the final MS recording will sound. This allows = adjustments of the=20 relative levels of mid/side and microphone positions. Normally the = Mid level=20 will be fully up and the Side level almost all the way up. Depending = on the=20 amount of ambient information, you may want to back off a bit. In a=20 relatively dead space, you may want to do the opposite and back off = on the=20 Mid signal a bit.

Now here is something that is really different: Faux MS miking. So = you =20 don=92t have Neumann U87. In fact you don=92t have a figure 8 microphone = at all.=20 Well, fake it. Close mike your source with one microphone and send = that to=20 the Mid input. Then take a second microphone and place that back form = your=20 sound source to capture room ambience and send that to the Side = input. Viola!=20 Fake MS miking. It isn=92t the real thing but it sounds good. It is = completely=20 different than just panning the second microphone to a different = position=20 than the first. To hear what this sounds like just switch the Figure = 8=20 microphone to cardioid or omni. Another interesting thing to do is to = separate the two microphones used. Normally the two microphones are = as close=20 together as possible. I have recorded in a blues studio in Florida = that uses=20 a figure 8 mic as the room mike and mixes that in the same way as MS = miking=20 with the exception that there isn=92t even a Mid mic. This works = great as a=20 drum overhead too. Experimentation is the key. With that Leslie I = mentioned=20 earlier, we had the mid microphone in the bottom of the cabinet where = the=20 large speaker and rotating baffle are. The Side mic was placed about = a foot=20 from the upper rotating horns and about three feet from the Mid mic. = Sounded=20 great. Happy Soldering. Even if you don=92t build the Project r MS = Decoder,=20 hopefully you picked up a couple new tips for your audio arsenal.

Return to = previous page=20

=20 =20