Working with MS theory if you had a Single Point Source Pulse

[OzGizmo]

Please note: OzGizmo did not start this thread. The first 37 posts in this thread came from another thread.
You may continue this topic, but please be careful where you folks take this...
If there is no (real) value in the discussion it just might end up in the trash.
I know you will do the right thing!


So working with MS theory if you had a single point source pulse (lets call it a gunshot as we are recording in the real world)

If that shot occurred at 90Deg L, the output of the MS matrix would be mainly [fig8 (+) and small amount of mid mic (+)]

If that shot occurred at zero Deg which is straight in front, the output of the MS matrix would be mainly [mid mic (+) and fig 8 (+/-) which would cancel out]

If that shot occurred at 90Deg R the output of the MS matrix would be mainly [fig8 (-) and small amount of mid mic (+)]

If that shot occurred at 45Deg L, the output of the MS matrix would be [fig8 (+) and equal of mid mic (+)]

If that shot occurred at 45Deg R, the output of the MS matrix would be [fig8 (-) and equal of mid mic (+)]

So with that in mind the gun shot at 45Deg R would in fact cancel out as you are combining a (+) signal from the mid mic with a (-) signal from the fig 8 mic. [if combined in equal parts]

[apple-q]

Resulting patterns are here

http://www.wesdooley.com/pdf/technique.pdf

[OzGizmo]

Originally Posted by apple-q

Resulting patterns are here

http://www.wesdooley.com/pdf/technique.pdf

Thank you for the link to the theoretical description of MS and its use. BUT this is based on pure mathematics rather than reality. [It actually mentions that polar patterns in the article were created by mathematical calculations]

I find it interesting that manufacturers of MS mics will often give an overlay of the M and S patterns rather than an "actual" usable polar pattern.
We may have probably seen the graph sheets of the individual mic including the serial number. (I have them for my Sennheiser 416/ 816)

Talking with some colleagues today about MS mics and the possibly of a "dead" spot in the right channel, this would not normally be a problem UNLESS there is a sound on that EXACT point of 45 deg. on the polar pickup, like a door slam, hand clap, drum or cymbal hit.

If anyone knows of a real life polar pattern for an MS mic I would love to see it.

[0VU]

Originally Posted by OzGizmo

Thank you for the link to the theoretical description of MS and its use. BUT this is based on pure mathematics rather than reality. [It actually mentions that polar patterns in the article were created by mathematical calculations]

....

Talking with some colleagues today about MS mics and the possibly of a "dead" spot in the right channel, this would not normally be a problem UNLESS there is a sound on that EXACT point of 45 deg. on the polar pickup, like a door slam, hand clap, drum or cymbal hit.

....

Since you're keen on real world results over theory, bear in mind that the only way your particular circumstance could exist in the real world is with a perfect point source in a perfectly anechoic space. Both of which are distinctly un-real world. (Though I guess you could get the same result in a non-anechoic space if the specific reflection pattern cancelled itself out at the mic - but I suspect that's even less likely to occur in the real-world.) Otherwise a less than perfect point source and/or reflections from elsewhere in the acoustic space will prevent the phenomenon which you describe. It's certainly not something I've ever noticed as a problem when using MS - but I've only used it in the real world.

[John Willett]

Originally Posted by OzGizmo

I find it interesting that manufacturers of MS mics will often give an overlay of the M and S patterns rather than an "actual" usable polar pattern.

Of course - the polar-pattern will vary as you adjust the relative levels of the two capsules.

There is no way a manufacturer can do such a plot as it would very rarely be what you are actually using.

[apple-q]

Originally Posted by OzGizmo

So with that in mind the gun shot at 45Deg R would in fact cancel out as you are combining a (+) signal from the mid mic with a (-) signal from the fig 8 mic. [if combined in equal parts]

Don´t understand this. There is no "hole" in the pattern. It will cancel out in the sum but not in the difference which is how the directional info is maintained.

Check out the sum/difference patterns in the paper I´ve posted. Neither in theory nor in practice there is a hole at 45° right.

The ONLY difference between theory and practice is that a practical mic will have a frequency dependent pattern but this doesn´t result in "dead spots".

[OzGizmo]

Originally Posted by John Willett

There is no way a manufacturer can do such a plot as it would very rarely be what you are actually using.

Not at all... for example the Shure VP88 has a 3 position switch for L,M,H stereo. And that's about the same number of settings and patterns as many top end studio voice mics and they do it. So why not the maker of a MS mic?

0VU, this means every thing on that 45deg R ch line would be down in level, it might mean that group of the choir, or those instruments in the orchestra

[OzGizmo]

The green line is where the maximum cancellation will occur because M+ and S- one with a positive amplitude and the other with a negative amplitude will cancel each other. (the cancelation will be a dip where the M+ & S- overlap rather than a fine line)

apple-q the link you posted was to a theoretical paper only where the calculations were done with mathematics and NOT with a real audio signal.

[apple-q]

Originally Posted by OzGizmo

The green line is where the maximum cancellation will occur because M+ and S- one with a positive amplitude and the other with a negative amplitude will cancel each other. (the cancelation will be a dip where the M+ & S- overlap rather than a fine line)

apple-q the link you posted was to a theoretical paper only where the calculations were done with mathematics and NOT with a real audio signal.

I´m afraid you are misunderstanding MS

In what way is your light-beam model not theoretical or mathematical? I´ve never heard of sound travel like a laser beam. In fact your model is exactly the same that is used in that paper assuming that sound travels like a laser beam and mic patterns are independent of frequency. It also doesn´t matter if it´s a gun shot or a sine-tone coming from a speaker at 45°

Check out fig 2d second row in that paper. You can see your exact "dead spot" in the pattern of the left lobe of the resulting super-cardioid figure. So at 45° your laser-sound from the right will cancel out yes. But only on the left channel of the resulting XY setup. Which is perfectly intentional. Sounds coming from 45° right will only be heard on the right channel not the left. What I´m saying is that there is no "dead spot" in a way that a sound won´t be heard at all from the mic. Not even in your theoretical special case. This is how the directional decoding works in the first place.

And again the only difference between theory and practice is that the perfect patterns that are used for the calculations (and your example) are in fact frequency dependent in reality. But that´s about it. As a result there will be some frequency dependent "cross-talk" when the 45°-sound is picked up so the higher the frequency the better theory and practice will match.

Also: Your green line represents the maximum outer limit the mic can reproduce on stereo speakers with correct localization. By steering the level of the S signal you are changing the pickup angle.

To sum things up: There are no "dead spots" across the stereo panorama.

[John Willett]

Originally Posted by OzGizmo

The green line is where the maximum cancellation will occur because M+ and S- one with a positive amplitude and the other with a negative amplitude will cancel each other. (the cancellation will be a dip where the M+ & S- overlap rather than a fine line)

apple-q the link you posted was to a theoretical paper only where the calculations were done with mathematics and NOT with a real audio signal.

This is totally unimportant.

What you display would be the null point in the left resultant microphone.

When you polarity reverse the fig-8, this becomes the front of the right microphone and there will be a corresponding null on the other side.

It sounds to me like you don't understand how MS works.

[OzGizmo]

I do understand MS VERY WELL, may I make the suggestion that next time you have a MS mic set up "Open your mind.. and listen", listen to just the right channel O/P and move the mic a little to a sound coming from 45deg off axis.

When you listen to the O/P of R then switch across to the O/P of L ... why does L sound a little brighter than R?

Why does the image of a MS mic as DCaswellUK says "leaning to the left" previously in this thread, others have also noticed this as well. Infact have a read of the opening post on this thread....

As I said "Open your mind.. and listen"

I know this may be hard for some to grasp, "perhaps its a bit like convincing people that the earth is round rather than flat"

"There's none so blind as those who will not see" (or in an audio fourm..Listen!)

[panatrope]

Originally Posted by OzGizmo

I do understand MS VERY WELL, may I make the suggestion that next time you have a MS mic set up "Open your mind.. and listen", listen to just the right channel O/P and move the mic a little to a sound coming from 45deg off axis.

When you listen to the O/P of R then switch across to the O/P of L ... why does L sound a little brighter than R?

Why does the image of a MS mic as DCaswellUK says "leaning to the left" previously in this thread, others have also noticed this as well. Infact have a read of the opening post on this thread....

As I said "Open your mind.. and listen"

I know this may be hard for some to grasp, "perhaps its a bit like convincing people that the earth is round rather than flat"

"There's none so blind as those who will not see" (or in an audio fourm..Listen!)

As I have heard others say: "I like to keep my mind open ... but not so open that my brains fall out."

I, and others, accept that you may hear what you say you hear. However, please accept that others do listen to the results of M-S recording and do not hear what you describe. It may be your rig, it may be in your hearing, it may be in your environment. It is still a valid comment, but one with which others disagree.

Likewise, the irrelevance of mathematics/theory to "real audio signals". The same mathematical principles are used in the design of the microphone, and describe the fundamentals of microphone behaviour either singly or in groups (arrays). "Real audio signals" can also be described mathematically and behaviour analysed, but the underlying results are described by the same processes as analysing simple signals. This strikes me as a harking back to old amplifier evaluation, and using synthesised square waves instead of "real ones". In that world also, improvements came only when mathematical analysis tools could model the problem and predict how improvements could be made.

Mathematical theory and observations may not always coincide, but they must co-exist. For the most part, the mathematical description of M-S co-incides with behaviour observed in "the real world". The challenge is to find out why when they don't. Merely maligning mathematics is not helpful.

[apple-q]

Originally Posted by OzGizmo

I do understand MS VERY WELL, may I make the suggestion that next time you have a MS mic set up "Open your mind.. and listen", listen to just the right channel O/P and move the mic a little to a sound coming from 45deg off axis.

When you listen to the O/P of R then switch across to the O/P of L ... why does L sound a little brighter than R?

Why does the image of a MS mic as DCaswellUK says "leaning to the left" previously in this thread, others have also noticed this as well. Infact have a read of the opening post on this thread....

As I said "Open your mind.. and listen"

I know this may be hard for some to grasp, "perhaps its a bit like convincing people that the earth is round rather than flat"

"There's none so blind as those who will not see" (or in an audio fourm..Listen!)

First of all your 45deg-theory is plain wrong no matter how open your mind migt be. I take your post that you agree on that since you simply moved to us not being open minded enough to follow the theory of the ever lurking dead spots in MS stereo. How about presenting some real facts instead of claiming ownership of the "open mind" for yourself. Just because you can not see the round shape when you look out the window, doesn't mean the earth is flat.

Secondly the o/p used two different mic pres for m and s and two mics of two completely different brands. His s mic is not even a "true" 8 but a vintage dual diaphragm mic.

There can be lots of reasons why his XY signal is leaning to the left. Without some examples of the original discrete MS signals it wil be rather difficult to "open ones mind".

none of my MS setups is leaning to the left (neither in theory nor in practice). If the did I would sent in the mics for servicing. Also left doesn't sound brighter when I move sources to the right. One reason for this could be that the frequency response of one of the 8 sides is worse than on the other (maybe his m269 is having issues on one of the cardioids). Maybe the output of the neumann is not balanced between front and rear cardioid. Could be anything.
If you took two completely different mics for a real XY setup your result might as well be leaning to one side but this has nothing to do with the "math" behind XY-stereo.

but again this has nothing to do with the theory behind MS or minds not being open enough. If it was a general problem of MS it would happen in every MS setup out there.

[OzGizmo]

apple-q, this thread was started with a MS user NOT an XY user!!!! please get it right.

It DOES occur to some degree on ALL MS mic setups, the 45 deg null point on the right is when the mid mic and side mics are of equal level, by reducing the side level/s the null point moves further back, increasing side levels moves the point forward.

I find it interesting that you so flippantly dismiss my findings yet you are actually haven't bothered to try it yourself.
As a fellow professional this is the least you could have done.

[apple-q]

Originally Posted by OzGizmo

apple-q, this thread was started with a MS user NOT an XY user!!!! please get it right.

It DOES occur to some degree on ALL MS mic setups, the 45 deg null point on the right is when the mid mic and side mics are of equal level, by reducing the side level/s the null point moves further back, increasing side levels moves the point forward.

I find it interesting that you so flippantly dismiss my findings yet you are actually haven't bothered to try it yourself.
As a fellow professional this is the least you could have done.

Jesus, the XY example was to illustrate what I think could be the reason for the O/P´s issue because he was using 2 VERY different mics and there are too many variables to find the problem from his original posting.

You also seem to miss that the de-matrixed MS signal is XY-Stereo resulting in 2 "virtual supercardioids" as you can see in that paper. You can also see what you call "null point". Again there is no contradiction between what you say and that paper. Only your conclusion is wrong because you think that "null-point" is only present on the right side.

Did you actually read what I posted about that "dead point"? That "null-point" has nothing to do with left/right balance.
The crossing point (your green line) is nothing else but the border of the stereo opening angle that you´re steering with the level balance between M and S. Of course is moves the angle
These points are on both sides of the mic not only the right side. The "null-point" is at 45° on the LEFT channel of the decoded signal AND at 45° from the left direction of the RIGHT channel. Which means that if a sound source is at 45° on the right it will only be heard on the right channel. If it´s 45° on the left, it will sound on the left channel of the decoded signal. What a surprise! It´s called coincident stereo.

By steering the S-signal you are changing the angle of the virtual XY pair represented by your green line. ANY sound coming from further back is outside the stereo image such a mic can capture.

Having said all that it has absolutely nothing to do with the initial issue. If his signal is leaning to the left it has nothing to do with your null-point because it´s on both sides!!!

I am going to stop here because it´s pretty obvious that you do not understand this or I am incapable to explaining it.

You also seem to ignore that I wrote that this does NOT happen with my MS setup an that I did try to reproduce the issue.

[OzGizmo]

OK thanks for your interest but I dare say you haven't wanted to hear it?
But like all things in life people do have different opinions.

"There's none so blind as those who will not see" (or in an audio fourm..Listen!)

[apple-q]

Originally Posted by OzGizmo

OK thanks for your interest but I dare say you haven't wanted to hear it?
But like all things in life people do have different opinions.

"There's none so blind as those who will not see" (or in an audio fourm..Listen!)

ahhrgh....now you sound like the black knight in "monty python and the holy grail"

[John Willett]

Originally Posted by OzGizmo

I do understand MS VERY WELL, may I make the suggestion that next time you have a MS mic set up "Open your mind.. and listen", listen to just the right channel O/P and move the mic a little to a sound coming from 45deg off axis.

When you listen to the O/P of R then switch across to the O/P of L ... why does L sound a little brighter than R?

Why does the image of a MS mic as DCaswellUK says "leaning to the left" previously in this thread, others have also noticed this as well. Infact have a read of the opening post on this thread....

As I said "Open your mind.. and listen"

I know this may be hard for some to grasp, "perhaps its a bit like convincing people that the earth is round rather than flat"

"There's none so blind as those who will not see" (or in an audio fourm..Listen!)

No, what you describe does not happen at all.

My fig-8 has identical patterns both sides of the single diaphragm so there will be no difference at all.

The only way you will get a difference is if the rear lobe of the fig-8 is different from the front.

This can also happen if you use a dual-diaphragm fig-8 that has been used for vocal work and one of the diaphragms is covered with sh*t.

So your image will only "lean to the left" if there is a problem with the rear lobe of the fig-8.

All my MS recordings have been done with am MKH 30 fig-8 which is a single diaphragm and truly symmetrical in every respect and I have never heard what you are describing. It just does not happen.

[panatrope]

You may wish to bear in mind that the null at 45 deg R occurs in the LEFT channel. By symmetry, there is a null at 45 deg L in the RIGHT channel. Well actually it occurs around about 60 degrees as the cardioid response is not correctly drawn. It occurs at the angle of the synthesised hypercardioid null.

And what would you make of the actually perfect null at 45 degrees right in the left channel of a Blumlein pair, and vice versa. Are we going to have complaints about the same problem in Blumlein micing?

Perhaps a rigorous analysis of what the mathematical analysis might tell you is a good start. And ask yourself if your constructed analysis has possibly resulted in auto-suggestion to your auditory perception about your observed result. Did you hear the effect and then seek the mathematical explanation, or the other way round?

By my observation, this forum equally values subjective and objective contributions. Neither has dominance. But it does value rationality. So, you have an opportunity to give us real reasons.

[panatrope]

Originally Posted by panatrope

By my observation, this forum equally values subjective and objective contributions. Neither has dominance. But it does value rationality.

@Remoteness

Hey, did I just make policy on the run? Did I usurp your functions? If so, my humble apologies for being so presumptuous!

[apple-q]

Originally Posted by panatrope

You may wish to bear in mind that the null at 45 deg R occurs in the LEFT channel. By symmetry, there is a null at 45 deg L in the RIGHT channel. Well actually it occurs around about 60 degrees as the cardioid response is not correctly drawn. It occurs at the angle of the synthesised hypercardioid null.

Not sure who you are referring to when you wrote "you" but yes, this is what I have said earlier. The position of the "null" depends on both the polar pattern of the mid-mic and the level balance between M and S. And yes it is symmetrical on both sides of the stereo image.
But since you can not change the laws of physics one reason for a signal leaning toward one direction could be the imperfection of the polar pattern of the 8 on either side. Another could be a de-calibrated mixer channel which leads to the S signal not being mixed to left and right with the same level and all sorts of other things.
Example: I have a very old U87 in my studio that has been abused over the years in bass drums etc. and it sounds very different from the back than from the front when switched to figure of 8. I would never use this for an MS-setup because it will most likely give me an very unbalanced stereo image. But that has nothing to do with MS-Stereo. It´s simply a bad mic.

The MS setup per se does not lean to the left just like that.

[panatrope]

@apple-q

My remarks were of course directed at Oz Gizmo's responses.

Is that Open-Apple Q or Closed Apple-Q?

[OzGizmo]

Please can we stop the forum ping pong.

As most of you use or have access to a MS microphone try and listen to just the right channel O/P and move the mic a little to a sound coming from 45deg off axis on the right side of the mic and you should notice it.
Perhaps use a point sound source like a speaker with tone or music coming from it or a person talking etc.

PLEASE try it, before making further comments.

[panatrope]

There is no forum ping pong here.

The onus is back on you to justify the (flawed, in my opinion) basis for your conjecture.

As M-S is a symmetrical configuration, please explain why we should hear this only on the right, and not on the left as well?

Please be rational.

[apple-q]

Originally Posted by OzGizmo

As most of you use or have access to a MS microphone try and listen to just the right channel O/P and move the mic a little to a sound coming from 45deg off axis on the right side of the mic and you should notice it.

Notice what?
What will we hear when listening ONLY to the right channel?
Earlier you wrote that the left signal will sound brighter when you move the source to the right. How will we hear that when we only listen to the right channel? Should we not listen to the left channel then? I don´t get it.

But I shall try it this evening. Just wanted to understand what to look out for and how to listen.

[Yannick]

Originally Posted by OzGizmo

The green line is where the maximum cancellation will occur because M+ and S- one with a positive amplitude and the other with a negative amplitude will cancel each other. (the cancelation will be a dip where the M+ & S- overlap rather than a fine line)

That's really funny. Especially because on the R channel the green line would be where M+ is added with -(S-)=S+ so there is no cancellation. It's the null point on the LEFT channel.

Now, what happens with a hypercardioid M mic ? Or a omni M mic with a lower level than the S mic ? Can you give more drawings with red and green lines and comment ?

I'll be in the backyard in the meantime - it's too sunny here


PS I've heard LOTS of recordings with a brighter left ch - about 90% of all orchestral recordings. Maybe that prooves that on all AB omni spaced pairs, there are blind spots on the right side. Or were the violins sitting left ? right ! no left !
"non shall pass !"

[Bennett Williams]

Originally Posted by panatrope

It may be easier to get your head around it if you think of the M-S process as sum and difference. With normal stereo, mono is achieved by adding/summing the L and R signals. If you change the polarity of the R signal and then add, it is the same as the difference. In other words the result is what you took away from the L and R signals to make mono.

In M-S micing, the forward facing (normally cardioid) mic is producing a mono signal. The fig-8 mic at right angles to the M(ono) mic is a pressure difference mic. If it arranged so that sounds from the left of the mic give positive voltage out for positive (increasing) pressure, then sounds from the opposite (R) direction give negative voltage for increasing pressure. And because the null of the fig-8 pattern is facing forward, then there is no output for sounds in front.

So summing the M and the S means that sounds coming from the left will reinforce each other and sounds coming from the R will mean that the S voltage will tend to cancel the M voltage - so the sum of the two produces a left-oriented signal. Reversing the polarity of the signal from the S, which is the same as taking the difference M minus S, means that when M and -S are added you will get a right-oriented signal. Of course the ratio of the strength of the signal from the M and the S will depend on how far left or right of centre the sound source is ... just like panning a mono signal.

It is simple, but don't think about it - try it. The results will help make it perfectly clear.

I've tried something and I got interesting results. In the mid channel I put a mono signal. In the side channel I put a polarity reversed mono signal. Just from a theoretical perspective, if I did use a figure-8 microphone, the mono signal of the mid channel is in phase with the + of the side channel, so the positive numbers just add up and not much changes. The significant part of the equation comes from the out of phase part of the side channel. In theory then, maybe all that is needed in the side channel is a polarity reversed signal..... or maybe I'm a moron, but drums will do that to ya.

[panatrope]

The S-microphone is also a 'mono' microphone, but the essential ingredient for M-S is that a fig-8 microphone is a pure pressure-difference mic. So that the polarity of the output (+ve voltage for +ve increase in pressure) will be the opposite if you address it from the 'rear' side compared to the front. This is where the 'phase reversal' comes from.

This also means that if you reverse the polarity of the S- mic input, the effect is to merely swap L and R in the resulting stereo image.

Maybe this could be a step to be included in assessing the impact of the "brighter" side effect. Making no other changes to the setup, merely reverse the polarity of the S- mic input. Is the L side (now on the R speaker) still brighter or is it now the R side (on the L speaker)?

For even more rigorous testing, to remove the effect of the reversed image, physically swap the position of the L and R speakers so the image is back to the original L to R orientation. Now do this double-blind on a few observers and see what they think.

[OzGizmo]

mmmm interesting that the proponents of the MS technique are now speechless... I thought so, at least they now know a little more about MS recording than they did a few days ago.

I regard MS as "Maybe Stereo" rather than "Mid Side"

If you are serious about using MS be VERY sure to check the outputs of Left and Right as separate components NOT just in stereo.

While MS has an interesting pseudo stereo effect it should NEVER be regarded as a professional recording technique. And probably best suited for domestic DV cameras (and theoretical boffins).

[Corran]

If anyone is speechless it's because there is no need to further explain one of the oldest and most useful stereo techniques to someone who doesn't seem to want to understand.

What you describe is simply the "null" created by each "virtual microphone," which is consistent with the math. Please read again the posts refuting your "findings" and also maybe the original Blumlein tech papers if you can?

To humor you I listened to each side of an MS recording I did last week. Guess what, it sounds like two hyper-cardioid mics facing L and R, just like it is supposed to.

Sorry, you are wrong. M-S is an extremely useful stereo technique, and is the only stereo technique where you have 100% control over the "directions" of the microphones in post other than a Soundfield mic or the rare mics with dual outputs from both sides of a mic. The only time there will be a discrepancy is due to uneven pickup from each side of a dual-diaphragm mic or a broken mic or when one side of the recording is actually, you know, louder/brighter?

Also, my favorite use for M-S is when I am recording a concert with things both in front and to the side of the stage about 45 degrees, so I can use Blumlein in one direction and M-S in the other direction, or vice-versa depending on where the mics are pointed.