I am having a little trouble wrapping my head around this technique. I would like to experiment with it a little, but need a helping hand...

So, you have two mics, one facing the source in Cardioid and another perpendicular to the source in Fig-8. The next bit is where it loses me. Using a DAW, I will have three tracks, one cardioid and one for each side of the fig-8, one with the phase reversed. But, where does the second fig-8 track come from? It has an output on a standard 3 pin XLR and is still, to all intents and purposes, a mono microphone, but has two side it detects sound from.

Also, can i use two cardioid mics rather than a fig-8, or would there be too much off-axis sound being captured?

Thanks!

--------------------
Learner
'I don't care if it's got 90 tracks, what does it sound like baby!' - Ray Charles

Quote audioface:

---

So, you have two mics, one facing the source in Cardioid and another perpendicular to the source in Fig-8.

---

Actually the 'M' microphone that faces the source can be anything between omnidirectional and figure-8.

Quote:

---

But, where does the second fig-8 track come from?

---

You use the same as for the 1st track: it comes from the 'S' microphone (which must be figure-8 in any case).

So from M and S you can get L and R like this:

L = M + S (which means L is a mix of M and S)
R = M - S (which means R is a mix of M and polarity-reversed S)

You need at least 4 tracks (M, S, L and R). Some people use also a specific track for the polarity-reversed S channel. If we call this fifth track 'S2', then R = M + S2.
The minus ("-") sign corresponds to phase inversion.


Actually, you can mix the balance between M and S so it should rather be:

L = a*M + b*S
R = a*M - b*S

where a and b are the two attenuation factors you apply to the M and the S tracks


If you are using Reaper it is easy to route tracks to others and to reverse phase in the routing. Otherwise maybe there are good freeware decoding plugins, but I never tried any.

Quote:

---

Also, can i use two cardioid mics rather than a fig-8, or would there be too much off-axis sound being captured?

---

The problem would not be off-axis sound but rather insufficient capsule coincidence

Best
-j

Mid can be any pattern you want - though cardioid is the most popular.
Side is always a figure-8 with the +ve lobe facing left.

You record the two tracks.

To make normal stereo: M+S=L / M+(-S)=R

The easy way is to use the Voxengo MSED Plug-in in your DAW to turn this into stereo.

--------------------
John - Sound-Link ProAudio
President - Federation Internationale des Chasseurs de Sons

Quote audioface:

---

I am having a little trouble wrapping my head around this technique.

---

You're not alone -- it is a bit of a mind-twister initially... and then eventually it all clicks together and suddenly makes perfect sense!

Quote:

---

So, you have two mics, one facing the source in Cardioid and another perpendicular to the source in Fig-8.

---

Yes, but as others have said, the forward-facing Mid mic could actually have any pattern you like, depending on the kind of stereo acceptance angle you want from the decoded array.

With the Mid mic as a fig-8, the decoded array works out to the equivalent of a pair of crossed fig-8 mics (the classic Blumlein pair, in fact). With a cardioid centre mic you essentially have something approximating crossed cardioids. And with an Omni centre mic you have back-back cardioids. Adjusting the relative sensitivities of the Mid and Side mic allows some degree of variation of stereo acceptance angle (and therefore perceived stereo width) and shape of the equivalent crossed-pair's polar patterns.

The other thing to bear in mind is that the side of the fig-8 mic that is in the same polarity as the Mid mic MUST FACE THE LEFT SIDE OF THE SOURCE. If you get this the wrong way around, the decoded stereo image will come out left-right reversed too.

Quote:

---

where does the second fig-8 track come from?

---

It's a duplicate of the first fig-8 track, with identical gain but a polarity inversion. The normal fig-8 track should be panned hard left, so that the left side of your mix bus combines the Mid mic signal (which is panned centrally) with the in-phase output of the Side mic. The duplicated but inverted Side track should be panned hard right, so that the Mid mic signal combines with the inverted side signal.

Adjusting the level of the +S and -S tracks (which must be ganged so that they track together perfectly) relative to the Mid track will adjust the perceived stereo width. No Side signal at all gives pure mono (just the Mid signal, panned centrally). As you increase the +S/-S signals the image will start to open out. Push the +S/-S levels too high and you'll end up with the classic 'hole-in-the-middle' effect and an unpleasantly 'phasey' sound!

Quote:

---

Also, can i use two cardioid mics rather than a fig-8, or would there be too much off-axis sound being captured?

---

Not as an MS array -- but you could use them as a normal crossed cardioid stereo array.

For Mid-Side applications the Side mic always has to be a proper figure-8 mic, mounted coincidently with and at 90 degrees to the Mid mic.


it's hard to describe this in words, but I'll have a go -- it's easier if you draw it out on paper, and then it makes a lot more sense.

Consider a Mid-Side pair using a cardioid mid mic. The first thing is to break that cardioid down into its component elements: a cardioid pattern is derived from the combination of an omni pattern and a forward-facing fig-8 pattern with equal sensitivities, ad the + side at the front.

If you add those two patterns together at the four cardinal points:

Front = positive fig 8 + omni = lots of output signal
RH side = nothing from the fig-8 + omni = modest output signal
Rear = negative fig8 + omni = complete cancellation, deep rearward null, no output at all!
LH side = nothing from fig-8 + omni = modest output signal

Okay... so let's apply the same kind of analysis to the Mid-Side array using the cardioid mid mic, but analysed from the 45 degree left and right angles -- the nominal axes of a typical L-R coincident pair.

Front Left 45 degree angle = omni + positive of forward fig-8 + positive of side fig-8

To decode MS for the left channel we add the Mid and Side elements, so the two fig-8s sum to produce another fig-8, but angled at 45 degrees with the positive lobe front left. Add that to the omni element and you end up with a cardioid facing 45 degree front left.

Front right 45 degree angle = omni + positive of forward fig-8 + negative of side fig-8

To decode the right channel we subtract S from M (or add the inverted Side signal to the Mid signal, so invert the side fig-8 and the positive and negative lobes swap sides.

Combine the two fig-8 elements now and you have another fig-8 with the positive love facing 45 degrees right. Add in the omni element again and you have a cardioid facing 45 degrees right.

So the decoded equivalent of a cardioid/fig-8 Mid-side array is a pair of crossed cardioids at about 90 degrees.

Hope that helps to make it clearer! As I say, drawing it out on paper usually helps.

One final bit to think about: if you extend lines from the centre of the mid-side array forwards, passing through the points on each side where the fig-8's polar pattern crosses the mid mic's pattern, the angle between those two linens denotes the stsreo acceptance angle.

If you make the side pattern larger (ie increase the sensitvity of the side mic or increase the side level relative to Mid), then the stereo acceptance angle is reduced. If you have a small band on a stage and the acceptance anngle is reduced, the band will occupy a greater proportion of the entire stereo image -- so the image appears to get wider or 'more stereo'

Conversely, if you reduce the level of the side mic, the stereo acceptance angle gets wider and the band will appear to occupy a smaller proportion of the stereo image -- reducing to mono (which it does when there is no Side signal at all).

Hugh

--------------------
Technical Editor, Sound On Sound

Great article on MS recording here - http://designingsound.org/2011/08/tim-nielsen-special-ms-recording/#more-1 0857

And another, more techy, but good - http://www.neumann.com/download.php?download=lect0016.PDF

I'm a recent convert to MS recording (from XY), though I can't speak for it's effectiveness in music recording, by recording the undecoded separate mid and side signals you can have so many options afterwards in post.

Quote Lew_Tapes:

---

Great article on MS recording here - http://designingsound.org/2011/08/tim-nielsen-special-ms-recording/#more-1 0857

---

It's an interesting piece, and from the perspective of a location TV/film sound recordist where the concept is more prominent and popular for a host of practical reasons.

However, as is surprisingly common, there is a degree of confusion about some aspects in his text -- one of the more obvious being his claims that decoding to LCR doesn't suffer from the image narrowing effects of using divergence to fill a centre channel during a film dub... but actually it does because it's doing exactly the same thing! Oops!

hugh

--------------------
Technical Editor, Sound On Sound

Woah! Ok, after reading the more detailed replies, I can sort of see why it resulted in my getting a headache... It reminds me a lot of the A-Level Mechanics I just did, in that it takes 'simple' arithmetic (adding and subtracting) and makes it painfully hard! I sort of get it now though...

So, I send the signal from the +ve Edited by HR to avoid further confusion -- he meant Side mic (Left), panned hard left, to a new track in REAPER (I will use this as a reference DAW, as it is what I use...), polarity reverse it, pan it hard right, then link the faders to change together... Increasing gain = increasing stereo width

I will definitely be printing off this page, as well as those articles Lew_Tapes, so thanks to all. Next on the shopping list, a multi-pattern condenser... Presumably, using mics of a similar design gives better results? 3 AKG 414s set accordingly (I wish!), for example? The reason I ask is, I would, at some point soon, like to buy a Ribbon mic, thus utilising its fig-8 pattern straight off.

--------------------
Learner
'I don't care if it's got 90 tracks, what does it sound like baby!' - Ray Charles

Edited by Hugh Robjohns (27/08/11 09:12 AM)

Quote audioface:

---

Woah! Ok, after reading the more detailed replies, I can sort of see why it resulted in my getting a headache... It reminds me a lot of the A-Level Mechanics I just did, in that it takes 'simple' arithmetic (adding and subtracting) and makes it painfully hard! I sort of get it now though...

So, I send the signal from the +ve M mic (Left), panned hard left, to a new track in REAPER (I will use this as a reference DAW, as it is what I use...), polarity reverse it, pan it hard right, then link the faders to change together... Increasing gain = increasing stereo width

I will definitely be printing off this page, as well as those articles Lew_Tapes, so thanks to all. Next on the shopping list, a multi-pattern condenser... Presumably, using mics of a similar design gives better results? 3 AKG 414s set accordingly (I wish!), for example? The reason I ask is, I would, at some point soon, like to buy a Ribbon mic, thus utilising its fig-8 pattern straight off.

---

NO NO NO!

The Mid mic. is panned centre
The Side mic. (+ve lobe facing left) is panned full left
Polarity reverse the Side mic. and put that in another channel panned full right.

--------------------
John - Sound-Link ProAudio
President - Federation Internationale des Chasseurs de Sons

Quote audioface:

---

The reason I ask is, I would, at some point soon, like to buy a Ribbon mic, thus utilising its fig-8 pattern straight off.

---

Single-diaphragm figure-8 microphones are the best for MS (and its extensions in multi-channel recording).

Condenser microphones of this kind are rare, and most of them are expensive (Schoeps, Sennheiser MKH30, Josephson, etc).

Ribbon microphones can be less expensive and are still single-diaphragm (see the Beyer M 130 or the Coles 4038 for example). But many ribbon microphones roll off the high frequencies a lot, so these high frequencies will be mono after MS decoding (provided that the M microphone can record them of course). Also, if you are using an old ribbon microphone with a modern condenser microphone, you may have some sensitivity matching problems (not sure if it is really a problem in practice if you have the right preamplifiers).

There are also some dedicated MS microphones, see the ones used in movies for example.

Of course you may still have decent results with figure-8 microphones that are made of back-to-back cardioid capsules (like are most multi-pattern large-diaphragm microphones). But honestly, I have no idea, I never tried it in practice.

I do not record at the moment because I do not have the time nor the room for it, but I have always been interested in MS and other encoded recordings. In the past I have rented very expensive microphones for this (renting is inexpensive here). If I had to buy a good fig-8 microphones for MS I would probably go for two M 130 (Blumlein MS array).
I remember some people telling that the Samson active ribbon microphone is pretty good.
Maybe some Chinese figure-8 microphones are good too, in this case this is the least expensive solution.

-j

Sorry John! I seem to have upset you. What you described is what I meant, so sorry I didn't explain it that clearly. I left out the Mid mic as that is the bit I understand. I should have put that in. So yes, as you say, Mid = centre, +S = L and -S = R. Now I get how to set it up too, so I understand the theory behind the MS micing. Now, to put it into practice...

And with regards to mic choices, I have an Oktava MK-012 (well, two to be exact), so was planning on using that for the centre mic. With regard to the Side bit and ribbons, I was looking at the sE R1. Being a big fan of sE and wanting an inexpensive(ish) ribbon, this seems ideal for me. I know by being stingy, I sacrifice quality, but the R1 is, from what I understand, pretty good, and a certain company we all know and love has a good deal on them. In italics in case I'm not allowed to plug a company...

--------------------
Learner
'I don't care if it's got 90 tracks, what does it sound like baby!' - Ray Charles

Quote Bob Moose:

---

Single-diaphragm figure-8 microphones are the best for MS (and its extensions in multi-channel recording).

Condenser microphones of this kind are rare, and most of them are expensive (Schoeps, Sennheiser MKH30, Josephson, etc).

---

The Ambient EMESSER ATE 208 is a single-diaphragm condenser fig.8 that is not expensive - about £600 I think.



It's made for Ambient by MBHO, so should be excellent for the price.

I will be getting one in a week and plan to run it through its paces and write a review for SOS (if they'll print it).

--------------------
John - Sound-Link ProAudio
President - Federation Internationale des Chasseurs de Sons

Quote Bob Moose:

---

Single-diaphragm figure-8 microphones are the best for MS (and its extensions in multi-channel recording).

Condenser microphones of this kind are rare, and most of them are expensive (Schoeps, Sennheiser MKH30, Josephson, etc).

Ribbon microphones can be less expensive and are still single-diaphragm (see the Beyer M 130 or the Coles 4038 for example). But many ribbon microphones roll off the high frequencies a lot, so these high frequencies will be mono after MS decoding (provided that the M microphone can record them of course).

---

All fair enough... but there is a trap with many ribbons, in that a lot these days are designed deliberately to have different tonalities front and back -- a bright side and a dark side, if you like. While this provides handy flexibility when recording guitar amps, it messes up the stereo imaging when used for MS. So pick your ribbon side mic carefully.

While a single diaphragm mic is the ideal for side mic applications, the spacing between diaphragms in double membrane (multipattern) capacitor mics isn't a problem at all in practice, and a good many single-bodied MS mics actually have physically separated back-to-back cardioid capsules for the purpose, still with pretty acceptable imaging.

I've had perfectly acceptable results with AKG C414s and Neumann U87s as MS pairs before now, although I prefer the Sennheiser MKH 30/40 combination overall.

Hugh

--------------------
Technical Editor, Sound On Sound

Mmmmh tasty... Sadly, while £600 may be inexpensive to some, pricing is very subjective, and £600 means no food for a good few weeks maybe once I have had a play with the technique and have found some really good uses for it, an investment like this would be worth it. Until then though, I will stick to setting it up with mics I have access to through associates of mine, or that I actually own personally!

And with regard to ribbons, I think trying before buying will be the best way to do it, so thanks Hugh. That is something I had not thought of...

I feel like I could take some mics into the world and set up a MS pair now, so thanks for all the help!

--------------------
Learner
'I don't care if it's got 90 tracks, what does it sound like baby!' - Ray Charles