LDC and SDC (nominally cardioid) polar patterns varying by frequency

[Paul W]

I was looking at this page from the Neumann website about sonic differences between large and small diameter condensers.

https://www.neumann.com/en-us/knowl...between-large-and-small-diaphragm-microphones

I understand why LDCs and SDCs will differ in their off-axis sound at high frequencies, with wavelengths comparable to the diaphragm diameter and shorter. (Any wavelength short enough that you have peaks and troughs diagonally across the diaphragm will mostly cancel out, because there will be stripes of high pressure across the diaphragm and stripes of low pressure in between. At long wavelengths, you'll have near-peak values across the whole diaphragm as the peaks go across, and likewise for troughs.)

To me, that would seem to mean that the effect would be to narrow the polar pattern a whole lot across a couple of octaves (?), dependent on the diaphragm diameter, and that the pattern for a small diameter (say 12 or 13 mm) diaphragm would be the about the same as for a large diameter (say 25 mm) diaphragm, just one octave higher.

But that is not what they show for examples. Here are the polar patterns by octave for a Neumann U87A LDC and a Neumann KM184 SDC

If you look at the plots for the U87A, the polar patterns vary significantly from octave to octave, being very subcardioid at 125 Hz, then somewhat less subcardioid at 250 and 500, then classically cardioid only at 1000 KHz, and supercardioid at 2000, and so on.

The plots for the KM184 are quite different. It is subcardioid only at 125 Hz, but looks classically cardioid at 250, and 500, and 1000, and 2000, and 4000. Above that it narrows a lot, as I'd expect, but there's a four-octave range (interestingly where human hearing is most acute) where the polar pattern hardly changes at all, rather than noticeably narrowing at each octave.

Is that normal?

I'm guessing this has something to do with the relationship between the diaphragm diameter and the front-to-back delay due to the capsule thickness and the acoustic delay network, but I don't understand it.

 

[soliloqueen]

cardioid isn't really a real pattern that can be physically constructed like omni and figure 8. the capsule relies on the delay network to construct a cardioid pattern artificially. delay cannot be set in a way such that every frequency from the rear is offset by exactly one half wavelength, as would be required for full cancellation. the capsule can only be constructed with a set of considered compromises so that this is the overall trend. the way this is accomplished is up to the opinion of the designer and varies from unit to unit. when you think of it this way, of course polar response is all over the place.

 

[Marik]

Besides the diaphragm size the pattern also depends on type of the capsule design, specifically, dual diaphragm vs. vented design. So the U87 vs. KM84 is not quite fair comparison, if anything for that reason.

I'd suggest to study a Shure paper (which I believe is a must for anybody interested in the microphones and capsules) on this phenomenon and which should give you a very good explanation:

https://www.shure.com/damfiles/defa...r_ea.pdf-ff7e23c323c72be3deac4ffd742c1d26.pdf

Best, Mark

 

[soliloqueen]

Besides the diaphragm size the pattern also depends on type of the capsule design, specifically, dual diaphragm vs. vented design. So the U87 vs. KM84 is not quite fair comparison, if anything for that reason.

I'd suggest to study a Shure paper (which I believe is a must for anybody interested in the microphones and capsules) on this phenomenon and which should give you a very good explanation:

https://www.shure.com/damfiles/defa...r_ea.pdf-ff7e23c323c72be3deac4ffd742c1d26.pdf

Best, Mark

right! mark is exactly right. one has vents and the other has a diaphragm, this is one of the many design considerations i was meaning, paul. with cardioid capsules, it's up to the designer to decide which of many compromises to take based on what the capsule has to do. sorry i wasn't more specific. this is probably what i should have started with given that one of the big things you pointed out was the difference in low end pattern performance between the two. admittedly, i was a little drunk, so i jumped to the philosophical. bed time for me.

 

[soliloqueen]

i wonder what you'd need to create a theoretically perfect cardioid capsule over the entire audible band? it'd need to have the effective front to back distance change as a function of the frequency. maybe you could design some kind of crazy metamaterial that somehow separates all the incoming sound and pipes every frequency along the correct delay path and then make the capsule out of it. it could be like a 4 foot cube of metal foam where every single facet is precision designed. what an absurd and stupid idea. how would a proximity-free cardioid bass even work? the sound would be delayed by 1/40th of a second at 20hz from the rear. it would have to have a set delay regardless of the distance, which feels physically impossible. the transient response would be dreadfully bad. every frequency would be a different slice of time. i can't fathom any way to practically do this.

 

[Marik]

cardioid isn't really a real pattern that can be physically constructed like omni and figure 8...

Ari,

To avoid any confusion just wanted to elaborate a little. Both, true pressure omni and velocity ribbon fig8 are NATIVE patterns, where the first is a stiffness controlled system and the later is a mass controlled one. Technically, the cardioid is just a mere combination of both (exactly like what we did in our newest hybrid mic we presented at the last NAMM).

Of course, as you mentioned, the implementation of conventional condenser capsules is much more complex because it is a resistance controlled system, where besides forming a pattern we also need to get a flattest response (basically, damp the diaphragm), eliminate chamber resonances and non linear and frequency dependent nature of delay passages (since we are dealing not with actual wavelengths, but rather emulate those with acoustical resistance, compliance, capacitance, etc.), and phase anomalies. All that makes the system much more unpredictable...

Best, M

 

[Marik]

i wonder what you'd need to create a theoretically perfect cardioid capsule over the entire audible band? it'd need to have the effective front to back distance change as a function of the frequency. maybe you could design some kind of crazy metamaterial that somehow separates all the incoming sound and pipes every frequency along the correct delay path and then make the capsule out of it. it would be like a 4 foot cube of metal foam where every single facet is precision designed. what an absurd and stupid idea. how would a proximity-free cardioid bass even work? the sound would be delayed by 1/40th of a second at 20hz from the rear. the transient response would be dreadfully bad. every frequency would be a different slice of time.

I do have an answer to that. And no, the system is quite different, much more simple and elegant

Best, M

 

[soliloqueen]

I do have an answer to that. And no, the system is quite different, much more simple and elegant

Best, M

well, i was just speaking about the problem as a purely physical one. of course, one could construct the "theoretical" pattern "theoretically" instead of trying to hog tie yourself to the inherently limited approach of trying to design a physical cardioid capsule how is that hybrid mic of yours working?

 

[soliloqueen]

Of course, as you mentioned, the implementation of conventional condenser capsules is much more complex because it is a resistance controlled system, where besides forming a pattern we also need to get a flattest response (basically, damp the diaphragm), eliminate chamber resonances and non linear and frequency dependent nature of delay passages (since we are dealing not with actual wavelengths, but rather emulate those with acoustical resistance, compliance, capacitance, etc.), and phase anomalies. All that makes the system much more unpredictable...

see I'm kind of interested in designed resistive materials lately. capsules that use, for instance, textile metamaterials for acoustic resistance instead of doing it the "traditional way" out of metal. those are neat. i have lots of papers on my "to-read" list for when i eventually design a capsule from scratch. I know you made a single-pattern cardioid capsule a few years ago that I heard great things about, but you retired it for the hybrid approach, which says a lot about how well the hybrid approach works.

 

[Marik]

...how is that hybrid mic of yours working?

Beautifully! By far more natural, stable, free of MAJORITY of conventional capsules problems. Two native patterns combined together--the best of both worlds.

Best, M

 

[k brown]

cardioid isn't really a real pattern that can be physically constructed like omni and figure 8. the capsule relies on the delay network to construct a cardioid pattern artificially. delay cannot be set in a way such that every frequency from the rear is offset by exactly one half wavelength, as would be required for full cancellation. the capsule can only be constructed with a set of considered compromises so that this is the overall trend. the way this is accomplished is up to the opinion of the designer and varies from unit to unit. when you think of it this way, of course polar response is all over the place.

The EV RE15 came pretty close.

Any mic that has response this consistent off axis, is also going to have very consistent pattern with frequency.

 

[soliloqueen]

The EV RE15 came pretty close.

12-15dB rejection in the rear and more at 150 suggests that this is more of a supercardioid, but that IS a very even supercardioid to be fair. a cardioid with this level of consistency would be a feat indeed.

 

[k brown]

12-15dB rejection in the rear and more at 150 suggests that this is more of a supercardioid, but that IS a very even supercardioid to be fair. a cardioid with this level of consistency would be a feat indeed.

Yes it's a supercardioid. Was originally designed to use as a boom mic for TV.

 

[emrr]

RE15/16 is probably the best sounding mic for vocals in a room with drums that I’ve used, outside of going full omni. The bleed is very natural sounding. Most cardioids have a really intrusive sounding bleed that’s tough to work with in that case.

 

[k brown]

I was looking at this page from the Neumann website about sonic differences between large and small diameter condensers.

https://www.neumann.com/en-us/knowl...between-large-and-small-diaphragm-microphones

I understand why LDCs and SDCs will differ in their off-axis sound at high frequencies, with wavelengths comparable to the diaphragm diameter and shorter. (Any wavelength short enough that you have peaks and troughs diagonally across the diaphragm will mostly cancel out, because there will be stripes of high pressure across the diaphragm and stripes of low pressure in between. At long wavelengths, you'll have near-peak values across the whole diaphragm as the peaks go across, and likewise for troughs.)

To me, that would seem to mean that the effect would be to narrow the polar pattern a whole lot across a couple of octaves (?), dependent on the diaphragm diameter, and that the pattern for a small diameter (say 12 or 13 mm) diaphragm would be the about the same as for a large diameter (say 25 mm) diaphragm, just one octave higher.

But that is not what they show for examples. Here are the polar patterns by octave for a Neumann U87A LDC and a Neumann KM184 SDC

If you look at the plots for the U87A, the polar patterns vary significantly from octave to octave, being very subcardioid at 125 Hz, then somewhat less subcardioid at 250 and 500, then classically cardioid only at 1000 KHz, and supercardioid at 2000, and so on.

The plots for the KM184 are quite different. It is subcardioid only at 125 Hz, but looks classically cardioid at 250, and 500, and 1000, and 2000, and 4000. Above that it narrows a lot, as I'd expect, but there's a four-octave range (interestingly where human hearing is most acute) where the polar pattern hardly changes at all, rather than noticeably narrowing at each octave.

Is that normal?

I'm guessing this has something to do with the relationship between the diaphragm diameter and the front-to-back delay due to the capsule thickness and the acoustic delay network, but I don't understand it.

This is the main reason vari-pattern LDCs are a poor choice for coincident stereo techniques; there is almost zero stereo sep at low frequencies due to the widening pattern; in the bottom octaves, they are nearly omni when set to cardioid.

For this reason they are also not great for near-coincident techniques; they only really work as intended if the pattern remains largely cardioid at all frequencies.