M/S Matrix

[bradzatitagain]

[quote author="NewYorkDave"]You could always try PAIA's simple circuit, then upgrade to something fancier later on, if necessary.
http://www.paia.com/msdecwrk.htm
This could be built on a piece of perfboard in an afternoon.[/quote]

The PAiA M/S Mic is interesting too. It has balanced outs, which is a step up from the decoder box. I'm quizzed on how to ditch the electrets and their preamps and inplement balanced ins:

http://www.paia.com/msdecode.htm#msmic

Scroll to the bottom and hit m/s mic schematic:

http://www.paia.com/msmicsch.gif

Looking at the scat it looks like you could eliminate everything on the input sides of marked points A,C and E, right at the Mid and Side level pots. I just don't know how you'd install balanced ins (from outboard mic pres.) The pots on the Decoder in the same position in the circuit are 100k. It may be the case you'd need to keep IC1:B, IC:1A and IC:2A--on the m/s mic scat--as buffers. I profess no expertise!! (Which is to say: I dunno.)

rgrds,
Brad

 

[PRR]

> I just don't know how you'd install balanced ins (from outboard mic pres.)

No new holes really needed:

 

[bradzatitagain]

Hey PRR, you're the best, man, thanks! :!: That makes this $61 kit a very feasible answer to the M/S Matrix box cost dilemma: Wes Dooley has one for just under $600, and while I'm sure he engineered it to within a micron of its life, and it has true bypass, ssm2142 outs, and inserts for sidechain (and a case, ha!) this ought to do until I find the unattended grocery bag full of cash and diamonds at the bus stop.

 

[PRR]

Obviously you do this three times.

Or is it two? They use a goofy (wonderfully clever and inexpensive) 3-capsule MS rig. I have not really thought about how you translate that to 2-capsule MS wiring.

I've just revised the drawing (force your browser to Re-Load; you should see the PAiA copyright as well as my part changes). Put the 1uFd caps in the holes for R1 and C2: the 1uFd caps will be large and need the support. Then stick one end of the 10K input resistors in the "+" holes of C1 and M1. Let the other end of each 10K stick straight up: solder jack wires to the free ends (or if they will reach the jacks, great!).

I have not looked at the board to see if 1uFd film caps will even fit in the R1 C2 holes. I doubt they will fit properly, but if you are not going to throw the box around then you can be pretty improper and hay-wire. If it seems floppy, stick the caps down (or to each other) with silicone caulk or ShoeGoo. A little shrink-tube on the free end of the 10K may be in order.

 

[SSLtech]

Has anyone given any serious thought to this design?

Keith

 

[dayvel]

Don't think that will work. One of the 5534's should be wired as a summing amp. Wire the top 5534 as an inverting summer. Assuming the top input is the M signal and the S signal is (L-R), the output of the top 2141 goes to the summing amp and the inverting input of the differential amp. The bottom 2141 goes to the summer and the non-inverting input of the diff amp. This gives you an inverted L signal from output 1 and an inverted R signal at output 2.

Also, 5534's probably won't be stable in that circuit. Better to use a single 5532.

 

[SSLtech]

You're right. I misdrew it...

To correct the problem, I think you'd have to send R2 to pin 2 instead of pin 3. I think I should also remove R8, since it's going to drop 6dB from one leg of the sum...

Keith

 

[dayvel]

This ought to do it-

 

[SSLtech]

Dave, still gotta get rid of R8 I reckon... otherwise you'll have ((L-R)/2) - (L=R) as I see it... make R8 disappear and make R3 a wire... that should do it..

Keith

 

[dayvel]

Nope, the gain at the non-inverting pin of the bottom opamp is 1+ (R7/R4)=2, so you have to chop the L+R signal in half there.

 

[zobomix]

For mastering purpose, I am looking for something that takes a stereo signal and splits it up in to a Mid (center of stereo signal) and two Sides (S and -S which are the L and R information). Then, I would need a box that takes those three channels back into a stereo signal.

Several pro mastering engineers use this trick to eq just the center information, mostly vocals since those are always in the center. Another application is to control the width of the stereo image even though that usually screws the phase relation between the left and the right channels. It would be fun to experiment with that...

SSLTech, do you think your design could do that?

:razz:

 

[dayvel]

Actually, you only need one box. The matrix that derives Mid and Side signals from a stereo signal is the same matrix that derives a stereo signal from an M/S signal, as SSLTech mentioned earlier in this thread.

Here's how it works. One side of the box adds the signals together, the other finds the differenence between them. So when you put in a stereo signal, at the ouput you get (L+R) which is the "M" signal and (L-R) which is the "S" signal.

When you put in M and S you (M+S) and (M-S) at the output.
But
M+S = (L+R) + (L-R)
= 2L
and
M-S = (L+R) - (L-R)
= 2R
Aside from the polarity inversion, necessessitated by the inverting summing amp and level shifting due to the output buffers, this is essentially what SSLTech's circuit does.

 

[zobomix]

NICE!!!

I kind of felt that it could be something like that but was confused since I am a beginner.

Man, that's awesome. I dig this project.


:!: :!: :!:

 

[kgkeefer]

Western Electric 111C Repeater Coil was built for use with phantom telephone lines and should have excellent balance, wide bandwidth (used on both ends of class A analog leased lines, when there was such a thing) and lots of level handling (+30 dBM @ 20 Hz).

Is anyone using these old beasts in Mid-Side decoders?

Any advantage to the Sowter M-S circuit that uses FOUR 1+1:1+1 transformers instead of just a pair? The Sowter circuit passes phantom power while 2 trannies does not. Saturation is not an issue with 5 pound transformers.

Any thoughts about using UTC LS-30 or LS-30X vs. 111C REP?

Karl
Portland Oregon