Topic: Technical Mic Questions

[DSatz]

> Looking at the lobes of the patterns, it almost seems as if the subcard has a smaller lobe than the card pattern.

[reply rewritten] Neumann's Web page for the TLM 170 R has the patterns displayed in an odd sequence, with cardioid stuck in between omni and wide cardioid. If you click them in their real order { omni, wide cardioid, cardioid, hypercardioid, figure-8 }, you can see the pattern become progressively narrower.


> The figure 8, is the narrowest of all. [ ... ]

> If you are saying that the lobes for the figure 8 and standard card are identical, then I get it ...

No, you just said that the figure-8 pattern's front lobe is "the narrowest of all," and you were right! Consider the 1 kHz curves for the front of the microphone. The cardioid's response is down 3 dB at around 65 degrees off axis, and down 6 dB at 90 degrees. A figure-8 pattern, being narrower, has greater attenuation than the cardioid does as you get away from 0 degrees. Its response is down 3 dB already at 45 degrees off axis, and at 90 degrees it has (theoretically) no response at all.

Or think of it this way: The TLM 170's figure-8 pattern is the result of subtracting the output of a backward-facing cardioid from the output of a forward-facing cardioid. (Technically the signals are added, but the rear-facing diaphragm is reverse-polarized so the voltages subtract.) A cardioid still has some response at 90 degrees, so two back-to-back cardioids will both put out signals when sounds arrive at that angle. But due to their opposite polarization, the net result is zero. That creates the null of the figure-8 pattern.

Now, what if you move the sound source slightly forward from the 90-degree null? The output of the front-facing cardioid will increase, while that of the rear-facing cardioid will decrease. The result is an increasing signal of positive polarity. The farther you move toward 0 degrees, the more the front-facing cardioid's output will increase while the rear-facing cardioid's output will decrease. When you reach 0 degrees, the sound source is in the null of the rear-facing cardioid, so nothing much is subtracted from the output of the front-facing cardioid--which is fully on axis at that point.

Move the sound source back to 90 degrees and this time work backwards; you'll get the same progression with the signal polarity reversed. As you approach 180 degrees, the front cardioid puts out less of a signal while the rear cardioid puts out progressively more. But the rear cardioid's signal is being subtracted from the front cardioid's signal--so while the net sensitivity is increasing, the output is inverted in polarity relative to the sound source.


> I am still trying to figure out whether there is some way to isolate the signals from the two figure 8 mics, so that I am only getting what appears to be the smaller, narrower frontal lobes of the figure 8, and canceling out the rears.

The answer to that is still no. What makes the figure-8's front lobe narrow is the subtracting of signals of the rear-facing cardioid within your TLM 170. If you cancel that rear lobe, you also cancel its narrowing effect on the front lobe--and then you've got yourself a cardioid all over again.

Whether you do this with an M/S matrix or a mixing board with a polarity switch or in any other way, that's what you'll get. I'm sorry; it's pretty much a 2 + 2 kind of thing, and you're going for 2 - (-2) as a variation, but it still adds up to 4.

--best regards

P.S.: I was out walking and thinking about your initial question; I have the feeling that we're still not quite addressing it. I think this may be because your point of reference is a Sony stereo microphone with three cardioid capsules, one of which points forward (the "M" signal) and the other two of which are used to synthesize a sideways-facing figure-8 (the "S" signal). Maybe you didn't realize that the signals from the two side-facing elements are combined in opposite polarity--not simply added or mixed together. And maybe you wanted to subtract the signals of those left- and right-facing elements from the "M" microphone so as to narrow the front pickup pattern of "M".

The signals of the left- and right-facing elements aren't available separately--but if they were (as with some quad microphones having a "clover-leaf" pattern of cardioid elements, e.g. the Neumann QM 69 from the 1970s), this would indeed narrow the front pattern of the "M" microphone. At the same time, however, the overlap between the two side-facing elements (centering precisely behind "M") would create a rear lobe in opposite polarity to "M"'s front lobe. So again, it's the same result as what you get by simply dialing in the hyper- or supercardioid pattern of your TLM 170, or very nearly so.

[gratefulphish]

P.S.: I was out walking and thinking about your initial question; I have the feeling that we're still not quite addressing it. I think this may be because your point of reference is a Sony stereo microphone with three cardioid capsules, one of which points forward (the "M" signal) and the other two of which are used to synthesize a sideways-facing figure-8 (the "S" signal). Maybe you didn't realize that the signals from the two side-facing elements are combined in opposite polarity--not simply added or mixed together. And maybe you wanted to subtract the signals of those left- and right-facing elements from the "M" microphone so as to narrow the front pickup pattern of "M".

The signals of the left- and right-facing elements aren't available separately--but if they were (as with some quad microphones having a "clover-leaf" pattern of cardioid elements, e.g. the Neumann QM 69 from the 1970s), this would indeed narrow the front pattern of the "M" microphone. At the same time, however, the overlap between the two side-facing elements (centering precisely behind "M") would create a rear lobe in opposite polarity to "M"'s front lobe. So again, it's the same result as what you get by simply dialing in the hyper- or supercardioid pattern of your TLM 170, or very nearly so.

Now, you are sort of getting to what I was thinking, but not quite.  My point of reference had nothing to do with any Sony mic.  I was thinking classic M/S, with a TLM-170, or other similar mic perpindicular to the stage, and a center mic (TLM 170, or any mic with the polar pattern of your choice, most likely card, as the center.)  As I always understood it, M/S processing split the signal of the figure 8 mic in two, for lack of better terms, the left side, and the right side.  In doing the mix, you then started with one side, and then mixed in a portion of the center mic.  Same with the other channel. But I always, (and maybe well incorrectly) understood the left and right to be discrete signals, and that when you were mixing one, you were for practical purposes, only "getting the signal from that side," plus any center channel that you mixed in.

Hence my question, if these two lobes of the figure 8 are really discrete in M/S mode, then why could you not separate them, whether you call it left/ right, when the mic is perpindicular, or front/back, if the mic is facing forward.   As I understood your last answer, the effect of completely removing the opposite signal, whether in M/S or not, effectively "changed" the polar pattern of the original figure 8, to something closely resembling, or equal to, a cardiod.

So, after all of this, the question is, for processing, are those signals separate, and if so, can they be separated from one another, without causing an effective change in the polar pattern.  I know this sounds confusing, but my brain keeps telling me that if you can do M/S, then you can do this.

[DSatz]

Ah, the core misunderstanding has finally revealed itself, I think. In M/S matrixing, the front and back halves ("lobes") of the figure-8 signal are not processed separately! They can't be separated (as everyone's been trying to tell you), and fortunately, M/S doesn't need them to be separated.

One key fact which I think may have eluded you (or perhaps you just haven't quite realized its implications yet) is that the two lobes of the figure-8 microphone have opposite signal polarity from each other. If you fire a pistol in front of an omni mike--shooting blanks, I hope, and wearing earplugs--the output signal voltage should initially swing in a positive direction. The same is true if you fire the pistol in front of a figure-8. But if you fire the pistol in back of the figure-8, the signal from that same microphone will initially swing negative. Are we clear about that?

If so, OK--and M/S works like this: You set up a forward-facing "M" microphone of any type or pattern that you like, and you place a sideways-facing figure-8 "S" microphone at (ideally) the same point in space--at least in the horizontal plane which conventional stereo playback systems are concerned with. The "M" microphone's output, by itself, is a mono pickup of whatever you're recording, and should be usable as such directly. And now that you have a Ph.D. in mono compatibility with M/S, let's see how stereo works with it.

A sound source that is perfectly in front of the M/S microphone pair in a completely dead room will be picked up by the "M" microphone on axis--but it will fall right into the 90-degree null of the "S" microphone and generate no signal there at all. A sound source that is 45 degrees to the left of center (as the microphones "see" the world) will still be picked up by the "M" microphone, but now it is no longer in the null of the "S" microphone, so it will be picked up by both "M" and "S" at the same time. The two signals will have the same polarity, since by convention, the front (= positive polarity) lobe of the figure-8 is turned to the left for M/S.

A sound source that is 45 degrees to the right of center will be picked up by both "M" and "S" microphones, too. But in this case, the sound will reach the rear (right-facing) lobe of the figure-8, causing the "S" signal to have inverse polarity (it will be 180 degrees "out of phase") in relation to its "M" counterpart.

You can see, therefore, that if you simply add the M and S signals together, all sound sources that were in the center or on the left will be included in the result--but all the sound sources that were to the right of center will tend to cancel out because of the polarity confict. On the other hand if we invert S before adding the signals, the reverse will occur: All sound sources that were in the center or on the right will now be included in the result, while all sound sources that were to the left of center will tend to cancel out because of the polarity conflict.

If this is unclear, please stay with it and see whether it makes more sense after you mull it over a while. It is the key point in understanding M/S--because all that an M/S matrix does is simply to add M + S to derive the stereo L signal, while to derive the stereo R signal it inverts S and adds that to M (or you could say that it subtracts S from M; same thing).

The above explanation is simplified since, unless you're on location at Wimbledon, there usually are sound sources between the extreme left and right (which may well be farther apart than +/- 45 degrees, depending on the pattern of the "M" microphone). There is also reverberant sound to consider, and some nifty tricks that can be done by processing the M and S signals separately before combining them in the matrix. (By the way, I liked the first movie a lot better than the second one; did you?)

But the main point is that an M/S matrix does nothing more than (M + S -> L) and (M - S -> R); it never does any separate handling of signals from the front and rear lobes of the S microphone, which are only theoretical constructs anyway--not signals that could exist as separate real entities. When I go out walking in the daytime, I have a shadow; it is measurable and everyone agrees that it is there, but I can't weigh it or eat it.

Does this help? If not, please keep asking. It's well worth developing a solid, gut-level feel for how this all fits together; that can be very useful to you in future situations.

--best regards

[gratefulphish]

+T again.  I want to thank you again for your patience, knowledge, and abitlity to communicate it, so that someone like me can actually begin to grasp the concept.  (BTW, so it wasn't just me, the Neumann site had the card and subcard patterns out of order?)  I am trying to study this whole subject, in-depth, and you have been greatly helpful. If there are any written treatises that you would suggest, not involving too much advanced math or physics, I would love to read one.  I will reread your explanation again tomorrow ( I have to head off to a show) and respond with any further questions.  Thanks again, you are a credit to this board.