Physics of polar pattern interference

[Tom Morrow]

Can someone point me towards a good visual tutorial or graphic explanation of how different microphones achieve particular polar patterns? I understand the generalities, that polar patterns are typically achieved by interference of multiple sound paths. But I want to know the specifics of the physical design:
how exactly does an interference tube work?
How do microphones that have flatter off axis frequency response than others achieve this?
Besides interference tubes there are plastic inserts that modify some microphone polar patterns (e.g. AKG C1000S super<->hyper); how do those work?
In microphones that use active electronics instead of interference (e.g. Sanken CS3e) to achieve polar response what exactly do those electronics do?
In mics like the studio akg 414 which have electrical switches for polar patterns, what do those switches do?
Does a super have a different design/shape interference device from a hyper mic, or just different ratios/specs?
I'm hoping to find a diagram/explanation that will show various microphone designs (omni, cardioid, super, hyper) and how waves constructively/destructively interfere to achieve those polar patterns.

I want to understand tradeoffs between off axis and on axis response, what about the design contributes to a flat off axis frequency response.

In particular I'm wondering if it is usually true that the tighter the polar pattern, the more the off axis response will vary by frequency. I think it is true that cardioid mics will be expected to have less frequency dependency in off axis response than super or hyper, but I don't have a sense of exactly why.

[Tom Morrow]

Just as background reading, Steve's post here:
http://www.dvinfo.net/forum/all-thin...icrophone.html does a good job of explaining the generalities.

[Richard Crowley]

An interference tube works by "adding" together the sound waves along the axis of the microphone because from that direction they sum together as they travel down the length of the labyrinth. But sound waves that arrive from an angle cancel each other as they enter the openings along the side. The amount of cancellation is dependent on many things including the design of the microphone, the angle of incidence, and the frequency of the sound.

Note that lower frequencies have longer wavelengths, and so they need physically longer labyrinths to effectively discriminate direction. Back several decades ago, ElectroVoice made a "bazooka" microphone that was shoulder (or tripod) mounted and something over 2m long! (EV 643) Electro-Voice Model 643 In more modern times, people like Sanken have used multiple capsules and electronic signal processing to attempt that kind of low-frequency performance without the monumental size.

The design of microphones is almost a "black art" and still has as much to do with skill and imagination as it does with pure science. In particular, the design of long "shotgun" interference-tube, and/or line-gradient microphones is an evolving art/science. As you observe, some manufacturers (like Sanken) achieve their goals by using several capsules combined electronically into the final signal.

All directional microphones work either by putting baffles ahead of the diaphragm to change the polar (and/or frequency) response, and/or use openings in the BACK of the capsule to use off-axis sound waves to physically cancel what the diaphragm is "hearing" from the front. For example, I have a pair of classic Sony C-37A microphones. They have a single, large-diameter, "exposed" capsule, and a mechanical "shutter" on the back of the capsule. With the shutter open, sound waves are allowed into the back of the capsule and cause the microphone to have a cardioid directional pattern. But with the shutter closed, they become a classic omnidirectional microphone.

Microphone design continues even after 100 years of technology advances because the perfect microphone has not yet been developed. To be sure many particular designs have achieved cult status for various reasons, not all of which involve flat frequency response or stellar polar response.

Also, because of the things you must do (mechanically) to create the polar pattern, different frequencies will be affected differently simply because of their varying wavelengths. Sometimes this is used to advantage (as when rock disk-jockeys exploit the "proximity-effect" that enhances low frequencies at close range. Some microphones use labyrinths BEHIND the capsule to make the mic LESS sensitive to this phenomenon. For example, the ElectroVoice 664 Electro-Voice Model 664

It might be instructive to read through the descriptions of some of the classic microphones on websites such as those by Prof. Coutant, as you can glean some design philosophy (and practice) details from the descriptions. Note also that virtually all of these designs are patented, and the patents (complete with detailed internal drawings) are public records available online at the US Patent and Trademark Office website at United States Patent and Trademark Office

[Tom Morrow]

Cool links Richard.

So interference tubes really have to be tuned to work at particular frequencies, and their effectiveness falls off for frequencies above and below the tuned frequency. That seems like a pretty blunt tool, although I suppose if it's tuned for the frequencies of the human voice that makes it workable.

Of course what we really want is something that behaves like a snoot on a spotlight, effectively absorbing the light (sound) to the sides with obstructions, irrespective of frequency. I'm getting the sense that the plastic thingies that snap into some microphones to change the polar pattern behave more on this type of principle, less on interference.

To some extent I suppose every piece of material you put near a mic capsule will have both effects (interference and occlusion); an interference tube has the same shape as a snoot. So trying to categorize as one or the other may not be so useful.

[Richard Crowley]

Originally Posted by Tom Morrow

Cool links Richard.
So interference tubes really have to be tuned to work at particular frequencies...

Actually, rather the opposite. One of the BIG design headaches is to make them as broad-band as possible and NOT tuned to particular frequencies. This is where much of the magic happens because it is a bending of the laws of physics.

You can see the basics of this scheme in the interior design of the EV RE664 where the high frequency back-port is right around the back of the element, the mid-frequency port in further back on top of the mic, and the low frequency port is a little hole back by the connector end. The next-generation design used a "Variable-D" scheme where the port was a continuous slot along the microphone with varying mechanical low-pass filters (metal with various size holes) along the port. This is most easily seen in one of my favorite microphones Electro-Voice Model RE15

[Garrett Low]

Hi Tom,

There are some really interesting things that go on when dealing with interference tubes and how sound is affected. Don't know how technical or deep into you want to go. Back in my physics study days (I spent 3 years as a physics major before becoming an engineer) looking at waves and doing various experiments, it becomes pretty evident how to manipulate the energy associated with each wavelength to tune the output for a specific result. Interference tubes are definitely "tuned" to gain a specific result.

If you want to sometime, we can discuss some of the physics involved in how to use constructive and destructive interference to accentuate or attenuate specific sounds. A better way to think about interference tube design is not to think in terms of just frequencies but more importantly in timing.

Garrett

[Tom Morrow]

A better way to think about interference tube design is not to think in terms of just frequencies but more importantly in timing.

Good point Garrett. I can imagine that in the real world interference tubes can be tuned to a wide range of frequencies rather than just a single frequency.

I just examined the interference tube on my ME66 by holding it up to a strong light so I could see through the screens. I saw that it was perforated with large holes (each close to the diameter of the tube), 8 or so holes spaced erratically over the length of the tube. I am sure those holes cause all sorts of nonlinearities in the response, and likely tend to widen the frequency range over which it responds by effectively breaking it up into a series of different sized tubes each with different resonant frequencies.

I would assume that there must be tradeoffs, and the wider the frequency response of your interference tube, the less effective it may be at doing its job.

I still don't feel like I know enough to evaluate whether hyper or super pattern boom mics will have more off axis frequency response anamolies than regular cardioids... I'm getting the sense that the answer is "probably but not worth worrying about too much".

[Jon Fairhurst]

It's good to be able to understand the relationship between frequencies and the addition of time-delayed signals. In the analog world, it was described by Laplace. In the digital world, by Fourier.

Here are the wiki descriptions, which are technically accurate, but thick.
Laplace transform - Wikipedia, the free encyclopedia
Fourier transform - Wikipedia, the free encyclopedia

I'm sure that there are better presentations of these transforms that would give a more intuitive sense of the time-frequency-phase relationships. If nothing else, it's a place to start from the theoretical side of things.

[Tom Morrow]

For the first time I made use of my AES membership and bought an article, AES article number 3249 AES E-Library Using Interference Tube Microphones in the Reel World

It contains a brief description and diagram of how interference tubes work. I learned that the holes I saw in the side of the ME66 interference tube are essentially a replacement for the slit that most interference tubes have. I learned that the slit is in fact integral to the functioning of the tube, not just an afterthought or detuning feature. The explanation didn't even mention resonant chambers but rather focused on the path length differences when sound comes from the side and part takes a longer trip through the tube, versus the same path length when sound comes from forward of the tube. Like Garrett says, it's good to think in terms of timing rather than frequencies.

Perhaps most intriguing was this paragraph, which addresses but does not clarify the issue that caused me to start this thread:

Interference tube microphones do not work well in reverberant sound fields. Hard wall
sets such as hallways, and tiled rooms like bathrooms or locker rooms are very reflective
for sound. Shotgun microphones sound very muddy when used in such locations, and
consequently a Cardioid or Hyper-cardioid pattern microphone should be employed. The
reverberation will not be eliminated; however intelligibility will markedly improve.

What I'm really looking for is an answer as to why super or hyper might give a marked improvement over a shotgun indoors.

[Steve House]

Originally Posted by Tom Morrow

...

...The explanation didn't even mention resonant chambers but rather focused on the path length differences when sound comes from the side and part takes a longer trip through the tube, versus the same path length when sound comes from forward of the tube. Like Garrett says, it's good to think in terms of timing rather than frequencies.
...
What I'm really looking for is an answer as to why super or hyper might give a marked improvement over a shotgun indoors.

Again, think timing and phase. The sound starts out at a single source. There's a direct component that goes straight from the point source to the microphone and there's an indirect component that travels from the source to a wall or the ceiling, bounces, and then goes to the microphone. The path that the indirect wave travels is longer than the path the direct wave travels so it arrives at the mic a fraction of a wave later. Plus the indirect wave's phase gets inverted at the bounce. So the result is there's a complex and unpredictable interaction between different waves of varying phases and timings within the mic's interference tube, upsetting that delicate balance of intentional interactions that gives the mic its directionality. Since cardioids and hypercardioids don't have a tube, there's less opportunity for damaging interactions between the various wave-fronts hitting the mic.

To see more about how cards and hypers get their directionality, look up "pressure-gradient microphones." Omnis are pressure transducers, cards and hypers are pressure-gradient transducers.

As an aside, those ports on the sides of shotguns are vital to its function and should never be taped over or allowed to be obscured by the mic mount, etc. The mic should be sufficiently forward in the mount so the clips themselves go around the mic well aft of the base of the interference tube.

[Jon Fairhurst]

Originally Posted by Tom Morrow

What I'm really looking for is an answer as to why super or hyper might give a marked improvement over a shotgun indoors.

As you go off axis, the mid tones roll off smoothly, the high tones roll off with lobes that come and go, and the low frequencies don't roll off much at all. When the voice reflects in the room, you get a boomy echo due to the lack of bass roll off. A less directional mic has less of this effect. What's worse, the HF lobes can cause a comb filter or phasing effect that changes as you move the mic.

With a super or hyper, you will still have some boominess, but at least you don't get the funny phase problems.

The best shotgun mic I've found for indoor use is the Sanken CS-3e. It uses three capsules rather than one capsule and a tube. This makes the lobes less pronounced and, amazingly, the bass rolls off as you go off axis.

On the polar plot, ignore the 16 kHz curve, which is most "air" for dialog, and you can see how smooth this is for a shotgun:
SANKEN MICROPHONE CO .,LTD. | Products [ CS-3e ]

Here is a more typical pattern.
http://media.soundonsound.com/sos/ma...unmicpat_s.jpg

A shotgun works well outdoors where you want to cut the sound at the sides and those sounds aren't correlated with the main signal source. The problem indoors is that the sound from the sides is an echo of the sound that you are trying to capture.

This pattern is typical of a super or hyper cardioid mic. It's a simpler shape than from a lobar shotgun and isn't as aggressive on the sides. The bass rolls off similarly to the higher frequencies and you don't see the lobe pattern at all.
http://www.sounderpro.com.tw/reviw/m...ages/super.gif

[Tom Morrow]

Good idea to search "pressure gradient". I found this article

Understanding & Using Directional Microphones which I don't have time to read now but which looks very promising...

John, yes part of what inspired this thread was my noticing that when I tried the CS3e I didn't hear many of the artifacts that you so eloquently describe. I suppose one thing I'm trying to tease out in this thread is how much of the CS3e indoor magic can be achieved by simply using non-shotgun mics (e.g. super, hyper, or even cardioid).

I know one experienced soundperson who uses an AKG C1000S (cardioid condenser) on a boom instead of a shotgun and gets pretty good results. I'm debating buying a C1000S to experiment with since I can't currently justify a CS3e which costs almost ten times as much.

[Jon Fairhurst]

I believe that a good hyper can beat the CS-3e indoors, but the hyper won't match the off-axis rejection of the CS-3e when there are outside sounds to be eliminated. But the hyper will have a smoother, more consistent, if less aggressive roll off.

As you move from hyper through super to cardioid, you get less and less rejection and more and more room color. While the echo can be realistic when recording in a marble mausoleum, the trend is to record dry and add color in post. As a friend of mine wrote, "you don't want a shot in a bathroom to sound like a real bathroom. You want it to sound like people imagine a bathroom to sound." So recording dry and adding color in post makes sense.

Frankly, one can probably get better sound by hanging blankets, treating a room, and using a cardioid than one can get using a hyper without the treatments. If a room has an ugly echo, you won't eliminate it by going from cardioid to hyper, you just make the echo a bit quieter. A well placed blanket can pretty much eliminate it.

Another consideration is booming. If you are holding a hyper on a boom, you want it to be light weight, and you're then looking at a ~$500 and up mic. On the other hand, if you do interviews, you can use a fixed boom and hang a heavy studio mic. That can open up a number of possibilities.

[Tom Morrow]

The abovementioned article

Understanding & Using Directional Microphones

is well worth reading to understand how non-interference-tube mics achieve their polar patterns by combining two transducers into one element.

I finally understand what people mean when they say a super can be better than a shotgun indoors. When I built my starter audio kit I went for the ME66 mic because I had heard that a super is better indoors, and the ME66 is advertised by Sennheiser as a super so I figured it would be good for indoors.

But now I understand that the Supers that are favorable indoors are the non-interference-tube designs like the Schoeps CMC641. I now see that the ME66 is actually advertised as super-cardioid/lobar, where I think lobar means interference tube.

I now understand that there's nothing particular about the super pattern for indoors; cardioid and especially hyper might also give the flatter off axis response that we want for indoors.

The Schoeps CMC641 super costs approximately the same as the Sanken cs3e proprietary active shotgun, and is also widely regarded as a good mic for indoors. I wonder what situations might cause one to prefer one or the other based on how they handle off axis rejection, but that might belong in another new thread to keep this focused on the physics.

[Jon Fairhurst]

In general, the CMC641 will beat the CS-3e indoors, due to its smoother pattern. (I say "in general" because in any specific case, one might prefer the Sanken.) That said, the Sanken will provide good results that are better than any typical interference tube shotgun.

The place where the Sanken really shines is in a large, noisy area like the Las Vegas Convention Center at NAB/CES. The off-axis sounds are a constant, loud murmur. The CS-3e should provide better off-axis rejection than the CMC641 and since the off-axis sounds aren't echos of the main sound source, the subtle lobes won't interfere with the sound of the target source. And since the LFs roll off, the murmur doesn't become a loud, LF drone like on a typical shotgun.

For outdoor use, I don't know that the Sanken is as rugged as a high-end Sennheiser, like the 8060.

So, when you want one mic to do it all due to budget or space/weight limitations, I'd get the Sanken. It's not the absolute best indoors, nor is it best in the harshest weather, but as long as it's not overly abused, it can deliver good results everywhere.

When working in smaller rooms where you want the best results and can afford (budget/size/weight) a dedicated indoor mic, I'd get the Schoeps.

When working outdoors, possibly in harsh conditions and can afford a dedicated outdoor mic, I'd go for an 8060 or similar.

When recording in large spaces with loud off-axis sounds, like a tradeshow floor, or a high school gym (where you don't want the mic visible in the frame), I'd go for the CS-3e above other choices.

On a soundstage, the choice between an 8060 and CS-3e would come down to personal preference for tone if not the lowest weight. I really like the Sanken sound; however, the 8060 has really high sensitivity. If the actors will whisper their lines, I'd choose the Sennheiser.

Given enough budget, I'd get them all, as well as a nice case, optimized wind protection, poles, etc. Given a limited budget, I'd start with the do-everything Sanken. Even with multiple mics, the Sanken's flexibility allows it to be used for that second channel, indoors or out. You might have to EQ it to match the other mics, but it would be recording a different voice, and different voices often need their own EQ anyway. We only really need matched mics when recording the identical source.