STEREO RECORDINGS can deliver a realistic soundstage and convey impressions of size, detail, and depth that cannot be captured through other recording methods. Last month in our Recording column, we explored basic stereo recording techniques. This month, we explore some advanced stereo mic configurations.
Mid-Side Recording Originally developed by Alan Blumlein, Mid-Side or “M/S” stereo is a powerful coincident recording technique, providing very good localization, a strong middle image, and excellent mono-compatibility. M/S enables you to control the balance between direct sound from the source and reflected sound from the room boundaries. Mic placement is simple, but implementation requires a bit of deeper understanding.
M/S stereo is the only coincident technique that allows use of two different microphones, generally one cardioid and one figure-eight (bidirectional). The cardioid “Mid” microphone is placed directly in front of, and on-axis to, the source. The bidirectional “Side” mic is placed directly above or below the cardioid mic, 90 degrees off-axis to the source (see Figure 1). The pickup lobes of this mic face the side walls of the room, and one of the rejection points of this mic faces the source; the other rejection point of the bidirectional mic faces the rear wall. Sound captured by the Side mic is almost completely reflected; the mic receives little to no direct sound from the source.
Fig. 1. This photo shows two microphones placed for Mid/Side Stereo. The bottom microphone is a Lawson L47 MP set to figure-eight and facing the side walls of the studio. The top mic is a Lawson L47 (cardioid-only) and is directly facing the source. Your wheels should be turning at this point: We now have a direct “dry” signal plus a separate signal containing room ambience. Each mic will be recorded to a separate track, giving you the ability to balance direct sound with room sound, but this is not a stereo signal. They are two very different mono signals. If you adjust the balance of these tracks, you’ll hear the relationship between direct and room sound change, but you’ll get no sense of left-to-right imaging. Panning the tracks—for example, placing the cardioid on the left and the bidirectional on the right—might prove interesting, but would not be stereo.
The “trick” with M/S relies upon the relationship of the side mic to the source. In a manner of speaking, each side of the bidirectional mic captures a different signal: one set of reflections bouncing off the left side of the room, and the other set of reflections bouncing off the right side of the room. They just happen to be captured by the same diaphragm. These signals are by nature out of phase, due to the fact that they are pushing the diaphragm in opposite directions at the same time. If this sounds confusing, it may help to think about what happens when you speak into a figure-eight mic at exactly 90 degrees off axis: At most frequencies, sound bends equally around the sides of the diaphragm, pushing it in two opposite directions at the same time. Air pressure on one side of the diaphragm cancels air pressure on the other side, so the diaphragm doesn’t move (see Figure 2). This is how the nulls at 90 and 270 degrees off-axis are created on a figure-eight microphone. (Some mics do this better than others, and the null is not always perfect. Performance can vary with frequency and angle of the source to the microphone.)
Fig. 2. This illustration shows how sound bends around the sides of a bidirectional microphone. Air pressure on one side of the diaphragm cancels air pressure on the other side, so the diaphragm doesn’t move. We need a way to duplicate this “out-of-phase-ness” and treat sound from the left side of the bidirectional mic separately from the right side of the mic. There are several ways to do this. The easiest way (and in my opinion, the way to achieve the most control over the process) is to split the signal from the bidirectional mic, record it onto two separate Side tracks, and pan them hard left and hard right. Again, I have to hurt your brain: Think about what happens when you have identical sounds of equal level panned hard left and hard right in a stereo mix. They add, and the sound appears in the center. This is identical to having one track of the bidirectional mic panned center. The trick comes when you reverse the phase of one of the Side tracks, thus mimicking the phase relationship between your ears. Most DAWs include several plug-ins that feature a phase reverse button. I suggest using a trim plug-in (as opposed to an EQ plug-in, for example) because trim plug-ins use minimal processing power and minimize the risk of introducing unwanted changes to the signal. The trim plug-in is inserted on one of the bidirectional Side tracks, and the phase is reversed (see Figure 3).
When you listen to these three mics, you’ll have a mono signal until the moment you reverse the phase of one of the Side tracks. When you hit that phase button, the stereo image will open up like the parting of the Red Sea. The two Side tracks give you independent control over the level of the room ambience, and of course you can EQ and/or compress them for special effects. (More on that in a second.) If you pan the Side tracks to the center, or check the mix in mono, the ambient information is summed and disappears, leaving only the signal from the Mid mic and making M/S totally mono-compatible.
Fig. 3. Three tracks used for M/S shown in Pro Tools. The first track is the Mid track (yellow). The other two tracks are the Sides, left and right (blue). Note that the right Side track has a trim plug-in with the phase reversed. M/S Execution Most of us will use our DAW to create this M/S “matrix.” The Mid mic is recorded to a mono track and panned center. The Side mic can be recorded to a mono track and duplicated later, in theory yielding two identical Side tracks—though I have experienced one or two DAWs that for some inexplicable reason don’t maintain precise timing between duplicated tracks. This part is critical, as is ensuring that neither Side track is ever shifted along the time line. You can split the output of the mic pre and record the Side signal to a stereo track if the plug-in provides separate phase reverse for each channel (see Figure 4). Remember you need to reverse the phase of only one of the Side tracks. If both Side tracks are phase-reversed, the effect is nullified. Another option is simply recording the Side mic to two mono tracks simultaneously by setting their inputs the same.
M/S can also be created using a hardware matrix. This is an outboard processor usually providing an input for the Mid mic and an input for the Side mic. Internal circuitry handles the splitting of the Side mic and phase reversal of one of the resulting signals, and allows adjustment of the balance between Mid and Side levels. The output of the processor is stereo L/R, so you won’t have control over the relationship between the Mid and Side signals later. Some M/S encoders feature built-in mic preamps while others accept line-level input. Often the matrix gives the user the ability to change the output to XY, since audio from the mics can be electronically manipulated to create the same signal that would result from using two cardioid mics in XY.
Variations Depending upon the output polarity of the two microphones, the M/S stereo image might be reversed when you listen back. If so, rotate the bidirectional mic 180 degrees. The cool thing about M/S is that you don’t need a matched pair of mics. For example, you can create great M/S recordings using a Shure SM81 for the middle and a KSM44 for the sides. In fact, you don’t even need two similar mics. You can get excellent results with a cardioid condenser for the Mid and a bidirectional ribbon for the Sides. You can also experiment with using hypercardioid or supercardioid for the Mid, or even an omni. M/S is wonderful for capturing a drum kit or small ensemble in a medium-size room.
Since the M/S pair is coincident, it provides level differences but no timing differences, but since the side microphone rejects direct sound from the source and favors the reflections from the side walls, sound entering the Side mic is “delayed.” Since these reflections are so random in relation to the direct sound, phase becomes less of an issue because the Side sound bears little relationship to the mid signal.
Fig. 4. M/S in Digital performer using one mono track for the Mid and one stereo track for the Sides. Note that the trim plug-in has separate phase reverse for each channel (red arrow). I Ain’t No Dummy Binaural recording is a method that uses two microphones and a dummy to mimic the behavior of the human head. True binaural recording can be startlingly realistic and three-dimensional when played back via headphones because there is no crosstalk between channels, but the effect is diluted over speakers because sound from each speaker inevitably reaches the listener’s opposite ear. A proper binaural recording requires use of a dummy head that approximates the shape, weight, and texture of a human head—including the contours of the outer ear. Thus the reason you’ll practically need a second mortgage to afford a high-quality, commercially manufactured binaural dummy. You can however, get pretty good results by inserting a pair of very small condenser mics into a modified mannequin head or wig holder. These mics should be omnidirectional since the outer ear makes your hearing directional, not the eardrum. It is also important that the diaphragms are set flush with the surface of the head, since recessing them into the dummy’s ear canal actually creates the effect of sound having passed through the ear canal twice: once when it is recorded, and again when it is played back into the listener’s ear.
A variety of manufacturers produce in-ear microphones that can be worn like ear buds (Roland CS-10EM, MMAudio MM-BSM-8), or clipped to an eyeglass frame (Core Sound Binaural Microphone Set) for binaural recording. For increased realism, 3Dio offers The Free Space, a pair of life-size silicon ears with no head in between them. (Sounds like a few people I know. . . . ) With prices starting below $100 street for some of these options, experimentation is affordable. If you’re lucky enough to have recently won the lottery, you can check out the King of All Dummies: the Neumann KU-100 binaural head, which sells for a smart $8,000.
Jecklin Disk Pioneered by Jürg Jecklin of the Swiss Broadcasting Corporation, the OSS (Optimal Stereo Signal) technique, a.k.a. ‘Jecklin Disk,’ is a sort of middle ground between binaural and spaced pair. The configuration uses two omnidirectional mics placed 6.5 inches apart—the approximate distance between our ears. Centered between the two mics is a round, acoustically absorptive disk with a diameter of 11 inches. The purpose of the disk is to enhance channel separation down to approximately 200Hz, below which the wavelengths are long enough to bend around the disk. (This characteristic simulates the behavior of sound bending around one’s head.) Jecklin disks—as well as a variation known as the Schneider disk—are available complete with mounting hardware for the mics and stand. The nice thing about OSS is that it’s relatively easy and inexpensive to DIY a Jecklin disk from a thin (approximately 1/8-inch thick) board with one-inch foam mounted on both sides. Recordings made using OSS provide much of the realism of binaural recording, but maintain their “strength” when heard over loudspeakers.
Fig. 5. These Audio-Technica AT4050 microphones are placed in a Faulkner Phased Array, set to figure-eight and spaced 7.8 inches apart. Faulkner Stereo With a career spanning more than 30 years and thousands of recordings, Tony Faulkner is one of the most successful orchestral recording engineers in the world. Faulkner’s Phased Array was developed as an alternative and a remedy for some of the limitations of the Blumlein configuration. When used for orchestral recording, Blumlein captures a lot of room ambience, since the rear lobe of a figure-eight mic is equally as sensitive as the front lobe. Reducing the amount of ambient pickup requires the mics to be moved closer to the source, but this tactic manifests additional problems: The stereo image becomes unnaturally wide, and the musical balance of the ensemble is skewed since the mics are now closer to some instruments than others. Faulkner’s array employs two figure-eight mics placed side by side, at 0 degrees (on-axis) to the source, with their patterns parallel. The capsule spacing is defined as 20 centimeters (approximately 7.8 inches), and height is set to that of the listener (see Figure 5).
Faulkner’s spacing mimics the spacing between the ears, thus representing the phase and amplitude differences a listener might experience if seated at the position of the mics. Theoretically, the Faulkner Phased Array (as with most non-coincident arrangements) should present some phase issues when summed to mono. But in an article published in the July 1981 issue of HiFi News and Record Review, Faulkner pointed out that human hearing is not coincident. Regarding the microphones in the Faulkner Phased Array, he stated, “. . . the plane of maximum path difference coincides with the null in the directional polar diagram of the microphones.” In other words, sounds producing the largest arrival time differences between the two mics tend to come from positions where the mics exhibit maximum rejection. The result is that phase issues are actually minimal. Though intended for use on orchestra and chamber ensembles, I’ve used the Faulkner Array successfully in small studios for drums and percussion.
Are You Outta Yer Tree? Originally developed in the 1950s by engineers at Decca Records in the U.K., the Decca Tree has traditionally been used for orchestral recording. The Decca Tree employs three microphones to provide a spacious but stable stereo image that maintains integrity even when incorporated into a variety of surround sound formats. The Decca Tree requires a specialized mounting bar that holds the left and right mics two meters apart, with the center mic a meter and a half in front of the left and right, forming a “T”. (AEA manufactures several Decca Tree mounting systems.) The distance between the two mics provides the impression of space, while the center mic adds clarity and detail to the center image. Since the center mic is actually closer to the source than the left and right mics, the center image is very strong—preventing a “hole in the middle” that might otherwise occur with other spaced-pair techniques. Each mic is recorded to a separate track and panned to its designated location.
At its time of development, the Decca Tree was typically used with three Neumann M50 omnidirectional microphones; if you aspire to that standard, you’ll definitely need that second mortgage, because the M50 was discontinued long ago and fetches high prices on the used market. Whether you use vintage Neumann or newer microphones, the Tree is suspended on a mic stand a few feet behind, and approximately eight feet above, a conductor’s head. In situations where reduced ambience is desired, the pickup patterns of the left and right mics can be changed to something more directional, such as cardioid or sub-cardioid. Some engineers have had success using the Decca Tree in conjunction with close mics for recording drums, but you’ll need a room large enough to move the thing around, and a mic stand sturdy enough to hold a considerable amount of weight.
Suggestions for All Stereo Miking Techniques Any time a sound is recorded with more than one microphone, there is potential for phase cancellation and comb filtering, so it’s important to check mic placement with a few test recordings. Monitor the recording while switching between stereo and mono. Ideally you’ll hear only that the image collapses when you switch to mono. Listen for instruments or frequencies that sound hollow, swishy, or (at worst) completely disappear. Cymbal recordings are helpful indicators of when overhead mics are out of phase, because crashes will “swish” and sound flanged. If you hear any of these symptoms, move the microphones. You are not going to be able to fix such problems in the mix. It’s also helpful to have microphones that are very closely matched in frequency response and output level. Some manufacturers specifically offer “matched” pairs of mics (sometimes for extra cost, sometimes not), while other manufacturers maintain that their quality control is tight enough that any two of a particular model are close enough to form a pair. You’ll have to decide how critical “close enough” is, but when a pair of mics matches, stereo imaging is way more accurate than when they don’t. That said, don’t be afraid to experiment and do some research!
Steve La Cerra is an independent audio engineer based in New York. In addition to being an Electronic Musician contributor, he mixes front-of-house for Blue Öyster Cult and teaches audio at Mercy College, Dobbs Ferry campus.