"The audiophile lexicon describes what a listener hears. This one describes what an engineer decided, months or decades earlier, that made it possible to hear it at all."
Every stereo recording is a set of engineering decisions frozen in time. A microphone type, a spacing, an angle, chosen before a single fader was ever touched. This lexicon is a companion to The Audiophile Lexicon: where that page names what you hear, this one names what was decided, in the room, before the music ever reached a wire.
The Microphone Itself
The physical device where every recording begins.
Transducer. Any device that converts one form of energy into another. A microphone is an acoustic transducer: it converts variations in air pressure into a corresponding electrical signal. It does not record sound itself, only the pressure changes arriving at its diaphragm.
Diaphragm. The thin, moving membrane inside a microphone that responds to air pressure changes. Its size and mass define much of a microphone's character: light diaphragms track transients faithfully, heavier ones trade some of that speed for mechanical stability and lower noise.
Large-Diaphragm Microphone. A microphone built around a comparatively large capsule, prized for low self-noise and strong sensitivity on quiet sources. The trade-off is a diaphragm that grows more directional at high frequencies and can behave less predictably off-axis.
Small-Diaphragm Microphone. A microphone built around a compact, low-mass capsule. It typically offers more uniform directionality and faster transient response across the audible range, which is why it is so often chosen for stereo arrays where consistent off-axis behavior matters more than raw sensitivity.
Proximity Effect. The bass boost that occurs when a directional microphone is placed very close to a sound source. It is a predictable consequence of the microphone's directional design rather than a flaw, and experienced engineers use it deliberately as often as they compensate for it.
Self-Noise. The electrical noise a microphone generates on its own, independent of anything in the room. It sets the practical floor for how quiet a source can be before the microphone's own hiss becomes audible in the recording.
Maximum SPL. The loudest sound pressure level a microphone can handle before its output becomes audibly distorted. It matters far less for a solo voice than it does close to a snare drum or a trumpet bell.
Polar Patterns
The shape of a microphone's sensitivity to sound arriving from different directions. No pattern is more accurate than another, each simply preserves a different slice of the acoustic event.
Polar Pattern. A plot of how sensitive a microphone is to sound depending on the direction it arrives from. It is the single most consequential design choice in any microphone, because it determines how much of the room the microphone will capture alongside the performer.
Omnidirectional. A polar pattern with equal sensitivity in every direction. An omnidirectional capsule has no proximity effect and an exceptionally uniform low-frequency response, which makes it valued for capturing the natural balance of a room, at the cost of capturing everything else in that room too.
Cardioid. A heart-shaped polar pattern, most sensitive to the front and largely rejecting sound from directly behind. Cardioid capsules give the engineer control over how much of the room enters the recording relative to the performer, which is why they dominate close and moderate-distance work.
Figure-Eight (Bidirectional). A polar pattern that is equally sensitive to the front and the rear while rejecting sound from the sides almost entirely. Because it captures both the performer and the reflections returning from behind the microphone, a figure-eight capsule is unusually good at preserving the acoustic character of a room, not just the source facing it.
How We Hear Space
The mechanisms the ear uses to locate sound. Every stereo technique below exists to preserve some combination of these.
Interaural Time Difference (ITD). The tiny difference in the moment a sound arrives at each ear, caused by a source sitting off-center. Never larger than roughly 700 microseconds, it is one of the two primary cues the brain uses to judge horizontal direction, particularly for low and mid frequencies.
Interaural Level Difference (ILD). The difference in loudness a sound produces at each ear, caused by the head partially blocking higher frequencies from the far ear. It works alongside ITD and grows more significant as a sound's wavelength approaches the size of the head.
Head-Related Transfer Function (HRTF). The frequency-dependent filtering the head, torso and outer ear apply to every sound before it reaches the eardrum. It is what lets the brain judge whether a sound is above, below, in front of or behind the listener, on top of simple left-right position.
Precedence Effect. The mechanism that lets the auditory system assign a sound's location to the first wavefront that arrives, folding later reflections into a sense of loudness and room character rather than perceiving them as separate events. It is the reason two loudspeakers, each reaching both ears with a slight delay, still produce a single stable phantom image instead of acoustic confusion.
Binaural Recording. A technique using two microphones mounted at ear position on a head-shaped structure, intended to capture sound exactly as a human listener would hear it. It is remarkably convincing over headphones, where each ear hears only its intended channel, but loses much of its precision over loudspeakers because of acoustic crosstalk between the ears.
Coincident and Near-Coincident Arrays
Techniques where two microphone capsules sit at, or very near, the same point in space, so the stereo image is built mainly from level differences rather than timing.
Coincident Stereo. Any technique in which two microphone capsules occupy virtually the same point in space, so a sound wave reaches both at essentially the same instant. Because there is no meaningful arrival-time difference between channels, the stereo image is created almost entirely by level differences, which gives coincident arrays excellent phase coherence and mono compatibility.
The Blumlein Pair. A coincident technique using two figure-eight capsules crossed at 90 degrees, one facing left, one facing right. Because figure-eight capsules capture both the front and the rear, a Blumlein Pair preserves not only the performers but the natural reflections of the venue, which is why it demands, and rewards, a genuinely good-sounding room.
XY Configuration. A coincident technique using two cardioid capsules, their forward patterns crossed at an angle typically between 90 and 120 degrees. By using cardioids rather than figure-eights, XY captures proportionally more direct sound and less of the room, producing a more focused, controlled image that suits less forgiving spaces.
Mid-Side (M/S). A coincident technique built from two different signals rather than a matched pair: a forward-facing Mid microphone, usually cardioid, and a Side microphone using a figure-eight pattern rotated 90 degrees to measure the difference between the left and right halves of the sound field. The stereo image is created afterward through matrix decoding rather than through the physical geometry alone.
Matrix Decoding. The simple arithmetic that turns a Mid and Side signal into a conventional left and right pair: Left equals Mid plus Side, Right equals Mid minus Side. Because the decoding happens after recording, the engineer can widen or narrow the stereo image later simply by adjusting the level of the Side signal.
ORTF. A near-coincident technique developed by French national broadcasting in the early 1960s, using two cardioid capsules 17 centimeters apart and angled outward by 110 degrees. The spacing approximates the distance between human ears, so ORTF captures both level and timing differences at once, the same two cues the ear itself relies on.
NOS. A near-coincident technique, developed by Dutch national broadcasting, closely related to ORTF: two cardioids spaced roughly 30 centimeters apart and angled at 90 degrees. It trades slightly more timing information for slightly less angle than ORTF, producing a comparably natural but distinctly voiced stereo image.
Spaced and Orchestral Arrays
Techniques that deliberately separate the microphones to capture more of the room itself, at the cost of perfect phase coincidence.
Spaced Array (A/B). A technique using two omnidirectional microphones placed well apart, rather than close together. Because each capsule samples a genuinely different position in the sound field, spaced arrays capture reverberation, arrival time and hall character with a sense of scale that coincident techniques cannot reproduce, at the cost of reduced mono compatibility.
Hole in the Middle. The audible weakening of the center image that can occur when a spaced pair is set too far apart for the size of the ensemble. Centrally positioned performers lose focus and seem to drift out of the stereo picture even though nothing about their playing has changed.
The Decca Tree. A three-microphone orchestral array developed at Decca Records in the 1950s: two outer omnidirectional capsules for left and right, and a center capsule slightly forward of the pair to anchor the middle of the image. It resolves the central tension of spaced recording, width without losing focus, by distributing spatial information across three points instead of two.
Supporting the Main Array
Vocabulary describing how additional microphones are used once a main stereo technique is in place.
Main Array. The primary microphone technique that defines a recording's perspective, stereo image and sense of space. Every other microphone in the session exists to support what the main array has already established, not to replace it.
Spot Microphone. A microphone placed close to an individual instrument to provide subtle reinforcement, most often used when a performer risks being masked within a larger ensemble. Used correctly its contribution is only a few decibels, enough to restore clarity without the listener ever noticing it is there.
Minimalism. A recording philosophy that uses the fewest microphones capable of preserving the musical event, relying on musician placement and room acoustics to do work that would otherwise fall to extra microphones and processing. It succeeds under favorable conditions: a well-balanced ensemble in a genuinely good room.
Multimiking. A recording approach using many individually placed microphones, common where isolation, independent balance control, or acoustically difficult venues make a single array impractical. It trades some of minimalism's natural coherence for far greater control during mixing.
Problems, Rules and Philosophy
The trade-offs and guidelines that follow naturally once two or more microphones are combined.
Comb Filtering. The pattern of frequency peaks and notches that appears when two correlated signals, carrying the same sound, combine after arriving at slightly different times. Named for its resemblance to the teeth of a comb when plotted, it is a consequence of geometry, not a fault in any microphone or cable.
The 3:1 Rule. A practical guideline stating that the distance between two adjacent microphones should be at least three times the distance from either microphone to its intended source. It does not eliminate comb filtering, but it reduces unwanted sound reaching neighboring microphones enough to make the remaining interference far less audible.
Phase Coherence. The degree to which multiple microphones preserve consistent timing relationships between channels. Coincident arrays have excellent phase coherence by design; spaced arrays trade some of it away in exchange for a greater sense of room and scale.
Mono Compatibility. How well a stereo recording holds together when its two channels are summed to one. Techniques with strong phase coherence, coincident arrays and Mid-Side in particular, tend to sum cleanly; widely spaced arrays are more prone to frequency cancellations when folded to mono.
Stereo Geometry. The complete set of distances and angles between microphones and sound sources that a stereo technique establishes. It is described as geometry because that is, quite literally, what determines the recording: the level and timing information a listener's ear will later decode as space.
Every technique here trades one thing for another, width for focus, room for control, timing for level. None of them is the correct answer. Each is simply the honest record of what an engineer chose to preserve, and what they chose to let go.
Questions About Recording Technique
Is one stereo technique simply better than the others? +
No. Each preserves certain acoustic information well while accepting limitations elsewhere. A solo instrument in a small room and a full orchestra in a concert hall call for entirely different geometries, and an engineer's judgement of the room and the ensemble matters more than any single technique's reputation.
Why do so many of these techniques use exact angles and distances? +
Because the numbers are not arbitrary. The 17 centimeter spacing in ORTF approximates the distance between human ears; the 90 degree crossing in a Blumlein Pair is what makes two figure-eight capsules divide a sound field exactly in half. The geometry is chosen to match the way human hearing actually works.
Can a recording be fixed later if the microphone placement was wrong? +
Only partially. Levels, equalization and dynamics can all be adjusted after the fact, but the spatial relationships encoded at the moment of recording, the timing and level differences between channels, cannot be reconstructed once they are lost. This is why engineers treat microphone placement as one of the most consequential decisions in the entire process.
Does more microphones always mean a more detailed recording? +
No. Every additional microphone adds both opportunity and risk: it can reinforce a passage that would otherwise be masked, but it can also introduce comb filtering and disturb the temporal relationships the main array already captured. Experienced engineers ask how many microphones are necessary, not how many are available.
Listen for the geometry, and you start to hear the decisions behind the recording, not only the recording itself.