Quick M/S Question

[mosrite]

Can anyone think of any reason why, when using M/S, you couldnt place an element squarely within one of the 'side' mics lobes? Will this simply mean that that element will appear hard panned when you matrix down for the mix or are there other implications?

Thanks,
Mosrite.

[tINY]

What do you mean by element?

If you mean a sound source, then no. If a sound in one of the lobes will end up in both sides. Turns out this is more like your ears work anyway.

Remember that if the sound is in the side lobe, it still gets picked up by the cardioid somewhat. And remeber that L = M+S and R = M-S. That means that the signal from the figure eight is in both sides, but out of phase.

The difference is the relationship of the "middle" signal. On one side, it'll add and on the other side it will subtract. If you are in the center of the lobe of the 8, you'll be down about 6dB on the card.

I guess that if the gains are set for it, you could end up with a "hard panned" sound at one location if you calibrated your mic pair for it. I don't remember anybody doing this as a matter of course. Usually, they'll set the relative gains so it sounds good in the room.






-tINY

[mosrite]

Thanks Tiny although Im still a bit confused by this M/S thing.

If the sound source is in the left lobe (or +S) of the figure of eight and there is nothing in the right lobe except room (-S) then the only thing adding the sound source to the right lobe is the Mid mic? So if the sound source were directly in the left lobe and very out of range of the Mid mic would it not then appear panned left when things are decoded?


Thanks again

[tINY]

No, The "S" is one signal. the minus sign in front of it just means that it is out-of-phase on one side compared to the other.

Like a figure 8 microphone. If you have something on one side of the mic, you get a signal. The only thing that changes when you move to the back side of the mic is the polarity of the sound.

What you are doing with the M+S and M-S is changing the phase of the "S" signal when you combine it with the "M" signal.

Remember that if you have an identical signal out-of-phase with another and combine them in equal parts you end up with no signal (unequally, and you end up with less than the louder signal). In phase, and they add.

Look at this and see if it helps:





-tINY

[Yannick]

When not 100% familiar with MS recording, it is easy to forget that the left side of the S and the right side ("-S") are 1 and the same signal.

They are NOT S and -S ...
The bidirectional mic has only one diaphragm and has no clue as where sound comes from. (do not confuse the S mic with two cardioids back to back, with separate outputs)

Just imagine an MS setup with two bidirectionals (Blumlein, but rotated 45deg, and through an MS matrix). If you put a sound source at 90deg, the M mic pics up nothing (theoratically) and the S a loud, in phase signal. BUT, hereafter it goes through the MS matrix and you end up with S in the left speaker and -S in the right ...

However, this can work if you use a omni M mic for example.

[mosrite]

Thanks guys. While im familiar with other stereo techniques M/S is something im investigating fresh. Almost got my head around it though! The thing that confuses / confused is the idea that there is no seperation of information of the Side mic. To my mind there is different information going into both lobes but I suppose these cant be conveyed as it is just one signal? Getting there!

[ShamansDream]

This is the capsule in the ECM-MS5 by Sony.
Three capsules matrixed to a left and right signal.

When shopping for a stereo mic in the late 70s I auditioned that Sony and a Shure VP-88 - the VP88 sounded so much better i bought one right away. Using three cardioid capsules is not really MS - it is a fake way to imitate it. No wonder I hated the Sony - besides lacking low end it has a custom style of imitation MS!

DO rerad the Wes Dooley articles, they contain basic information essential to understanding. I believe there was a very good article from the BBC many years ago too.

[Improv]

I found it helpful to think of the Mid to Side relationship in terms of Carrier (mid) and Modulator (side). What you hear is the interaction of the two.

Start by simplifying it to two channels... mid and one side (+S). This is not that different from an X/Y stereo setup, only your patterns are different and you are choosing your cardioid side as your "baseline"... then the side mic is the delta, or difference, from the mid mic.

Now, the third, out of phase channel, shouldn't be too hard to integrate into the picture at this point. You kind of have a stereo image going with your mid and +s channels, so you take that side channel modulator, flip its phase, and pan it the complete other direction. Now you have this second "stereo image" that is the complete opposite of the other "stereo image" only on the other side. As something moves from one "stereo field" to the other (i.e. from left to right or vice versa), the phase relationships between the mid and side channels, and the way you have it all panned, takes care of all the image trickery for you.

I love m/s.

Maybe not the most technically sound way of thinking about it, but like I said, it's helped me to get my head around it at least.

[mosrite]

Improv,
Excellent explanation, thats a good way of thinking about it. I suppose it could be argued that with MS you dont attain the same level of seperation of sources within the stereo spread that you would with say XY. Could it be said that with MS you get a stronger stereo depth though? (through a combination of a strong centre image and the 'sitting back' of the figure of eight mic?)

[Improv]

the thing about M/S is the flexibility. You can really hype the stereo channel for a "surround your brain" sort of vibe, and still not lose mono compatibility (though the levels will obviously change in mono, the harmonic character will remain the same). It's a different sort of stereo than any other techniques (duh... they all sound different) but I find it to be more natural and much eaiser to work with on mixdown.

Keep kicking myself for forgetting to try it on piano... anyone here do m/s piano?

I have done MS piano. When there is s strong likelihood of mono listeners, on radio...

This Friday I'm recording a small vocal group (www.kitka.org) in a church. I'm planning to use a set of close mics and a tall stand with MS pair for more of the ambience. I'm thiking it might be a good time to try using two Figure 8 mics, instead of cardioid and figure 8 - I have a matched pair of 4050s and would like to see if it does a better ambience capture... Anyone here tried that who can say anything about it?

<L>

[Improv]

using figure-8 on mid will do probably just what you think it will... diffuse the middle slightly. Basically it makes it sound in mono more like it does with the sides in in stereo.

[Matti]

These might help to demonstrate MS (ST250/SP442 DEMO) and Soundfield methods http://www.soundfield.com/downloads/software.php
Matti

Good question! I bought them second hand as a pair, and they come in a case with spaces for both mics and shockmounts. He said they were matched - but I don't see it on the AT site either.

Glad you asked though - I see that mine are the AT4050-CM5 which is discontinued! It has better specs than the current non-CM5 model, so I guess you could call them "vintage" - at any rate I am very happy with them, from orchestra to vocal to overheads, they are my premium best pair among 60 mics I use for both studio and live work.

differences here:

<http://gearslutz.com/board/attachment.php?attachmentid=18818&stc=1&d=1145903589>
<http://gearslutz.com/board/attachment.php?attachmentid=18819&stc=1&d=1145903589>

<L>

[pkautzsch]

Theoretically, there should be absolutely no difference between XY and MS.
What is actually being done when adding the S to the M signal (or subtracting it) ist the same as happens in dual diaphragm mics. The shape of the resulting pattern depends on the way the signals are combined. Since the same signals are combined in two different ways, you get two different patterns (which are symmetrical to each other). This would be the same as using two mics with that pattern each. There is no phase difference between L and R because there is no time difference, and the phase inversion that is introduced by the fig-8 is complemented with the phase inversion in the matrix.
Stereo separation does ONLY happen by level differences in any coincident setup like XY and MS. Thus, you DO obtain the same level of separation when having a cardioid and a fig-8 at the same level. You can increase or decrease stereo separation by increasing or decreasing the level of the fig-8 (and, similar, by widening the angle of two mics in XY). Depth, on the other hand, has to do with decorrelation: L and R have very little amount of same-phase signals. The more decorrelated a signal appears on the speakers the farther away it is perceived. One can easily imagine this by looking at a room with a sound source in it: the farther away the source is from the listener, the larger are the differences between what your right and your left ear get. You can get decorrelation ONLY by spacing the microphones. Coincident techniques don't have decorrelation (that's why they are mono-compatible!), spaced mics have most decorrelation (and the wider they are spaced the more decorrelated they are). ORTF and similar "equivalent" techniques have a little decorrelation, but also level differences.

[mosrite]

You can get decorrelation ONLY by spacing the microphones.

In that case could you feasibly add a touch of decolouration by spacing the Side mic a bit further back from the Mid? Would this be madness or would it simply add decolouration / depth?

[tINY]

You'd certainly add a time delay by moving the figure 8 away form the cardioid. But. you would then sum the signals.

Transients would not arrive at each playback speaker separated in time.

So, "decorrelated" (time-difference) spacial queues would not be present in the recording.

You would get some comb filtering, though......





-tINY

[pkautzsch]

De-correlation is not the same as a small delay.
You would need to add a spaced pair and leave the MS setup as it is.

Originally Posted by tINY

Transients would not arrive at each playback speaker separated in time.

So, "decorrelated" (time-difference) spacial queues would not be present in the recording.

You would get some comb filtering, though......
-tINY

Sorry not correct. You need to try thid and see how it sounds. The ceneter holds, the side and thus the image shifts in time and becomes more vague. The transients would not arrive at the SIDES (both the same amount) at the same the same time as the CENTER...

[tINY]

Fair enough.

The main reason that moving the figure 8 mic away from the cardioid won't result in time-based spacial queues is that you are summing the two together and your two ouput signals are comprised of the same two input signals. Thus, the transients in the left and right speaker occur at the same time.

pkautzsch : please explain what you mean by "de-correlation"

Lou : I think you may be on to something if you are using a 3-speaker playback system.




-tINY

Lou : I think you may be on to something if you are using a 3-speaker playback system.

Big Grin! The only three speaker system I've used was in a 61 Galaxy... But it might be worth a try!

<L>

[pkautzsch]

"de-correlation" is sort of the opposite of correlation
ie. signals that are not very similar to each other phasewise.
With a coincident mic setup the signal coming from the Left mic (and later from the left speaker) is perfectly in phase with the signal from the Right mic - since they are at the same spot. Same applies to panned mono signals. Where a sound source will appear in the stereo image is dependent ONLY on level differences in this case.
As soon as there is a time difference between left and right channel, the two channels aren't in phase anymore. As long as there is ONLY direct sound involved, this simply means that it's coming out of one speaker earlier than from the other.
Now imagine a room with lots of reflections, each being another sound source, and thus sound arriving at the microphone from all directions. When the mics are at the same spot, the (diffuse) reverb will sound the same on both channels, and be completely phase-coherent - which is not what we experience in a real room. The reflections from one wall are different from those from another wall (except in a "perfect" room with a very small sound source exactly in the centre of the room!), and this results in phase differences (of course!). The more different the phase of the reverb is on both channels, the wider and larger is the illusion of the room, and the farther away does the sound source appear.
Summing a decorrelated stereo pair to mono will result in phase cancellations.

[tINY]

So - essentially - decorrelated signals from mics in an anechoic chamber with a single source are just time delayed (ignoring losses of the air).

But, when you have sound reflections, the delays from each reflection point are different than from the source to each mic. So, you loose this effect with any coincident mic technique in a reverberant space.

I guess that I always consider the final receiver the ears of the listner. So, spaced pair decorrelation gets convolved with the listening room decorrelation anyway. It's hard to take out the playback environment when not using headphones......



-tINY

[Matti]

I could go even for Decca three but with MS in the midle, it debends of how
it sounds at the moment of deciding... Or 3 MS rigs or...
Matti
Edit - Decca tree

[pkautzsch]

You got it, tINY. It's basically about reverb differences.
It's an interesting question if decorrelation can occur in anechoic environments though - time delay only doesn't seem decorrelation to me, since it it basically the same signal. You'd be able to tell where the sound source is, but it would of course not sound spacious...*ponder*
As to the influence of the consumer's room: also the control room where we record the stuff has a room that convolves what comes out of our speakers. We know our room, and the consumer knows his. Especially in the classical field, there are a few consumers sitting in their 60° stereo array quite similar to us in the CR, and these people are the ones that review our recordings in magazines.

This discussion is getting completely off-topic. Initially it was just a "quick MS question"...

Originally Posted by MATTI

I could go even for Decca three but with MS in the midle, it debends of how
it sounds at the moment of deciding... Or 3 MS rigs or...
Matti

Matti, did you really think the term is "Decca three" or was that a typo? It is a "Decca Tree" and if you do something else in it it is a modified DT...

<http://www.wesdooley.com/pdf/Surroun...ree-urtext.pdf > gives the specs.

A Decca Tree is already a stereo mic array - why would you want to add a stereo mic as one of the omnis? Seems uneccessary and confusing image-wise to me...

[Matti]

Sorry it was a typo or I just invented something. And this MS thing,
maybe I was thinking of MS rig with outriggers, I don`t remember
what was going on in my head. - You sometimes are too tired to think
clearly and there is this lanquage thing.Matti

"These might help to demonstrate MS (ST250/SP442 DEMO) and Soundfield methods http://www.soundfield.com/downloads/software.php
Matti"

Okay, just checking. There coul dbe a Decca three and a Decca five for all I know - Wes Dooley has one with rear mics too so that could be the Decca Five!