Topic: Quick omni question

[Tom McCreadie]

..., so I guess I'm calling horsefeathers on that part too.

Me too. I think it's a winning recipe for audio 'pea soup'.
Ach, pea soup, opaque as it is, can be tasty.  :-)

[aaronji]

I'm going to call horsefeathers on pointing an omni regardless of whether you align them vertically or horizontally. If it's a proper omni, then the only thing that matters is its spacing from its mates.

Omnis definitely have directional characteristics at higher frequencies.  The frequency at which this starts occurring is related to the diameter of the microphones.  Just look at the polar pattern graph for a DPA 4006 (or an MK2, if you prefer).  Pretty obvious and can certainly be used to your advantage...Omnis with very small diameters (like a 4060) are much closer to being perfectly omni, but even they exhibit this tendency at very high frequencies.

[Gordon]

I ended up just taping them to the nos bar maybe 5 or 6 inches apart.  I about forgot to even do it    my 3 yo was begging for ice cream the entire time I was setting up and thus had me a bit distracted.  anyway while there is not the "spit omni sound" it sounds very good to my ears.  a bit more crowd than the neumann's but that is to be expected with omnis.  these little mics really impress me!  going to figure out a way to do a proper 2-3 foot split at yonder next weekend.

[capnhook]

I'm going to call horsefeathers on pointing an omni regardless of whether you align them vertically or horizontally. If it's a proper omni, then the only thing that matters is its spacing from its mates. Distance creates the difference in signal arrival that creates the stereo image. If they're vertical instead of horizontal, then you have eliminated the space needed to create the image, so I guess I'm calling horsefeathers on that part too.

Most omnis' frequency response curves differ on-axis vs. off-axis.  You might actually point them to help get the response you want at a particular time.

[Gutbucket]

I'm going to call horsefeathers on pointing an omni regardless of whether you align them vertically or horizontally. If it's a proper omni, then the only thing that matters is its spacing from its mates. Distance creates the difference in signal arrival that creates the stereo image. If they're vertical instead of horizontal, then you have eliminated the space needed to create the image, so I guess I'm calling horsefeathers on that part too.

If the omnis you are using are larger than DPA 406x, they will be noticeably directional at the upper end of their frequency range.  You needn't take my word or that of other TS'ers for it.  Just look at the manufacturer's published polar patterns, or better yet, play around with the microphones and listen for yourself.   

The larger the size of the microphone housing, the more directional they will be up top and the lower in frequency that directionality will be noticeable.  That's why microphone manufacturers like Schopes, Neumann, Naiant and others offer optional attachments for their pressure-omnis which change the size and shape of the microphone housing around the capsule which allow users to modify the pickup pattern and frequency response of their microphones in predictable ways.

On the second part, you are correct that when the microphones are spaced one above the other, there will be minimal time-of-arrival differences for sounds arriving along the horizontal plane which is the same height above the ground as the mid-point between the microphones.  Those sounds will image in the center between the speakers.  Sounds arriving from above or below that plane will image toward one speaker or the other. 

More importantly I think, there is a lot more going on in stereo than directional imaging.  The diffuse sound, the reverberant ambience, comes from all directions and the horizontal plane only makes up a small slice of that. Spacing the microphones along ANY axis creates differences between channels for that stuff. 

The obvious experiment is to make a recording with the omni microphones positioned coincidentally, and then with them vertically spaced 3' or so, and listen to see if those recordings sound identical or not.  Even if you aren't inclined to actually do the experiment, just thinking about it probably makes it clear they will not sound identical, even though determining the nature of the differences probably requires actually trying it.

Besides, all two channel stereo (with the exception of binaural recording and playback) is something of a feathered horse anyway.  It's just an illusion and we are the magicians doing the slight-of-hand tricks.

[Gordon]

so as small as the church audios are does it make a difference if they are facing up or straight at the sound source?  Obviously I know nothing about omnis.

[Gutbucket]

A sound source usually sounds a little brighter up top when the microphone is pointed at it.  Somewhat similar in frequency range and degree of difference to the effect of using a foam windscreen or not.  Pointing omnis either up or forward is mostly a treble-range EQ tweak.

With a microphone the size of the miniature DPAs, that effect is minimal and way up above 15kHz.  Monitoring on headphones in a quiet room, and moving around the microphone while rubbing my fingers together a few inches from it and listening to the HF part of the 'swishsa' sound I can hear a slight difference, on-axis verses far off-axis, but it's subtle.  Subtle enough that I don't really worry about the orientation of those when recording music.
 
With a small but not miniature omnis, the frequency range is not really low enough to make a big difference in the range which strongly effects imaging, but an inter-channel difference in that range can improve the sense of 'airiness and openness' in a non-imaging stereo sense when the two microphones are pointed in different directions.

Since I can make pretty much the same subtle overall treble EQ adjustment afterwards as I would get from pointing the mics up or forward (and I tend to EQ things later anyway), I'm thinking primarily about ways of maximizing that subtle inter-channel difference which I can only get from the microphone arrangement.  So with small omnis I'd probably point them directly to each side, opposing each other.  If I wasn't going to EQ things and wanted all the high-frequency brightness I can get for the main sound, I'd angle them less far apart and point each directly at the stacks.

[Church-Audio]

Omni mics need a boundary between them IMO. So I doubt it will be very good unless you can manage that. You would be better off wearing them on a hat than clamping them to a bar with out a boundary between them.

[Gutbucket]

No hat, this was all about a quick last-minute setup of a secondary recording on the same stand without other stuff to assist. 

But Chis has a good point.  A baffle (if you can rig one) IS the best idea when the omnis need to be mounted close together, even if it's only a cardstock music flier or piece of pizza box or something taped to the vertical mic-stand between the two mics.  The baffle makes the microphones directional down through the midrange frequencies determined by the size of the baffle.  It's more or less a further extension of the same omni directional aspects we've been talking about.  That arrangement would definitely be the better choice for Left/Right imaging, and probably what I would choose if I was just running the omni pair and couldn't space them enough horizontally, but it won't get that open, spaced-omni ambience sensation stuff. 

The best choice of what configuration to use is always determined by what you are trying to achieve.  Different priorities suggest different answers, as much as different constraints dictate which are doable.


As Chris is very familiar, measurement microphones are usually small diameter omnis, mounted at the end of a thin tube, and both of those things help to reduce these directional effects.  When used for measuring room acoustics they are often pointed directly upwards, so the response is as even as possible for sound arriving from any direction in the horizontal plane, even though it is pretty much the same for all directions.

[DSatz]

bryonosos wrote:

> I'm going to call horsefeathers on pointing an omni regardless of whether you align them vertically or horizontally. If it's a proper omni, then the only thing that matters is its spacing from its mates.

With all due respect, this is why people who make recordings REALLY need to learn how to read polar diagrams. You would then know right away that this isn't so.

Perhaps when you say "a proper omni" you mean "a microphone that is fully omnidirectional at all audible frequencies." But essentially no professional-quality recording microphones are that way. And it's not because it's hard to build a microphone with this behavior; you can in fact buy suitable capsules ready-made for less than a dollar each.

The crux is that sound waves are reflected, either partially or completely, whenever they strike any rigid object whose dimensions are about 1/4 of a sound wavelength or larger. Sound wavelengths are inversely proportional to frequency, and in air, range from about 55 feet (for 20 Hz) to about 2/3 of an inch (for 20 kHz). One inch (the diameter of many large-diaphragm condenser capsules) corresponds to a sound wavelength at about 13 kHz, and 1/4 of that is just above 3 kHz. For a typical small-diaphragm capsule with about 1/2" diameter (the housing dimensions matter, too, but the capsule dimensions matter even more), the critical frequency is about twice as high. But in the top audible octave, even what we normally call "small" diaphragms aren't acoustically "small"--their size causes them to interact with the sound waves at those frequencies, in a way that alters their directional response (because the degree of reflection and diffraction depend on the angle of sound incidence).

Below I've posted a scan of the frequency response and polar response diagrams for an excellent, small omnidirectional microphone that I used a lot in the 1970s, and that also was used by WGBH-FM for live broadcasts of the Boston Symphony Orchestra back then. It has essentially flat high-frequency response for random-incident sound (i.e. when no particular direction of sound arrival predominates--an acoustically "diffuse" sound field). But as the graph shows, it has a pronounced high-frequency rise on axis, with the 3 dB point at around the 6 kHz that you would predict from the capsule's ~1/2" diameter (1/4 wavelength). So you damn well better aim it if you're going to use it at moderate miking distances! Sometimes you even need to aim it away from the sound sources (as WGBH did).

A capsule of about 1/4" diameter or smaller can avoid the bulk of these problems, and such capsules are favored for this reason in certain acoustical measuring applications. But the sensitivity of a condenser microphone is proportional to the capacitance of its capsule, and smaller capsules have lower capacitance (since capacitance depends on area, it's proportional to the square of the diameter). So sensitivity and noise become critical issues with very small capsules, and that's why they're not generally used for professional music recording.

--best regards

[voltronic]

bryonosos wrote:

> I'm going to call horsefeathers on pointing an omni regardless of whether you align them vertically or horizontally. If it's a proper omni, then the only thing that matters is its spacing from its mates.

With all due respect, this is why people who make recordings REALLY need to learn how to read polar diagrams. You would then know right away that this isn't so.

Perhaps when you say "a proper omni" you mean "a microphone that is fully omnidirectional at all audible frequencies." But essentially no professional-quality recording microphones are that way.

I think the closest you can get are Earthworks QTC mics.
http://www.earthworksaudio.com/microphones/qtc-series-2/qtc30/

[DSatz]

Maybe so. But for stereo recording, why would you want reflected sound to be picked up with full-strength high frequencies? That only smears and confuses the imaging, unless the venue is very dry. If anything, in spaces with typical reverberation, I think it's desirable for omni microphones to de-emphasize the upper-mid and high frequencies even more than a small-diaphragm capsule naturally does, when those frequencies aren't arriving on axis.

The classic example is the Neumann M 50 microphone, which had its capsule (similar in basic response to the KM 83 shown above) embedded in the surface of a 4 cm acoustically opaque sphere--see attached photos, with the screens of the microphones removed. That's the microphone that Decca used most often for their "Decca tree" recordings, although sometimes they used Schoeps M 201s (presumably in the omni setting, where they operated as pure pressure transducers). You can get a very similar effect by mounting a sphere accessory onto a small pressure transducer capsule--both Neumann and Schoeps sell such accessories. The Schoeps MK 2H and MK 2S work particularly well with them.

[Church-Audio]

Maybe so. But for stereo recording, why would you want reflected sound to be picked up with full-strength high frequencies? That only smears and confuses the imaging, unless the venue is very dry. If anything, in spaces with typical reverberation, I think it's desirable for omni microphones to de-emphasize the upper-mid and high frequencies even more than a small-diaphragm capsule naturally does, when those frequencies aren't arriving on axis.

The classic example is the Neumann M 50 microphone, which had its capsule (similar in basic response to the KM 83 shown above) embedded in the surface of a 4 cm acoustically opaque sphere--see attached photos, with the screens of the microphones removed. That's the microphone that Decca used most often for their "Decca tree" recordings, although sometimes they used Schoeps M 201s (presumably in the omni setting, where they operated as pure pressure transducers). You can get a very similar effect by mounting a sphere accessory onto a small pressure transducer capsule--both Neumann and Schoeps sell such accessories. The Schoeps MK 2H and MK 2S work particularly well with them.

What a thing of beauty. This sphere is to control the polar pattern and make off axis response smoother?

[DSatz]

The sphere causes a slight increase in upper-midrange response in the forward direction--basically a "presence" boost--while smoothly rolling off the high frequencies at the sides and back. The graphs on http://www.schoeps.de/en/products/ka40/graphics show the on-axis part of this effect. It's a subtle effect which allows the microphones to focus well when placed at greater than usual distances.

I think some people here might benefit from knowing about this technique, since I get the impression that people often have to record from greater distances than they would otherwise prefer to use.

Originally this approach was used by German broadcasters in the mono era to pick up an entire orchestra with chorus and soloists from a single miking position in a reverberant hall. You have to back the microphone away from the sound sources by quite a distance if you're going to get everything in balance without favoring one group of performers or another.

Schoeps' accessory sphere is called the KA 40; Neumann's is called the SBK 130. Neumann also builds a sphere of this kind directly into their "M 150 Tube" (so you can see the polar diagrams by going to http://www.neumann.com/?lang=en&id=current_microphones&cid=m150_data and clicking on the "Interactive view of diagrams" link), and also as part of the "KK 133 SBK" capsule of their KM A/KM D series.

--best regards

[Gutbucket]

I've never seen such good detailed photos of the acrylic sphere of M50 before.  What is the grey "chewing gum" blob atop the sphere? Some sort of cushion against the mesh basket housing?

Ratio of the pressure on the surface of a cylinder, cube and sphere, of the normal axis to the pressure in the incident sound wave-

This is from Stan Linkwitz's site. He credits- G. G. Muller, R. Black, T. E. Davis, "The Diffraction Produced by Cylindrical and Cubical Obstacles and by Circular and Square Plates",  J. Acoust, Soc. Am., Vol. 10, July (1938)

My own experimentation with similar sized DIY spherical attachments on the miniature DPA 4060-