The Architecture of Illusion

"A soundstage is not reproduced. It is reconstructed, in the room, from timing cues so small that a few centimetres of movement can create or destroy it."

Soundstage is one of the most argued-over qualities in high-end audio, and also one of the most misunderstood. It gets dismissed as subjective decoration by listeners who trust only a frequency response plot, and it gets treated as a mystical property that some gear simply "has" by listeners who trust only their ears. Neither position survives contact with the physiology and the physics involved.

This article makes the case, from first principles, that soundstage is neither myth nor magic. It is a measurable, mechanism-based consequence of how the auditory system processes timing, and of how a loudspeaker's radiation pattern meets a room at a specific position and angle. Understanding the first half explains why two systems that measure similarly on a frequency response chart can image very differently. Understanding the second half explains what to actually do about it, for dynamic, planar, and horn-loaded loudspeakers alike.

Hearing Is More Than a Tone Test

At love cable / Tonmeister we start from a simple, often overlooked fact: hearing a high-frequency tone is only one narrow aspect of hearing. Each person experiences sound and music differently, and even with limited high-frequency sensitivity, that experience is real and meaningful. So when someone argues "that person has bad hearing, so they can't judge," the claim rests on a category error.

Judging music reproduction is not a pass or fail tone test. It is a comparative process built on timing, phase, and spatial cues that remain fully accessible well beyond the point where extreme highs fall away. High-frequency threshold tests measure absolute sensitivity at one frequency. They do not measure how the brain fuses timing, level, and spectral cues across two ears to produce timing precision, transient attack, and soundstage.

This is why "can you hear 15 kHz?" and "can you hear timing, phase, and image differences?" are different questions, governed by different mechanisms and requiring different standards of proof.

Frequency Audibility Is an Absolute Threshold

Whether a person can hear a given frequency is a question of absolute sensitivity, tested with pure tone audiometry: a single tone, presented at increasing level until it is detected. The result is one number, a threshold in dB SPL, for that frequency, for that ear.

This threshold is well mapped. Sensitivity peaks around 2 to 5 kHz and falls off toward both extremes, and the high-frequency falloff is neither gentle nor optional. It degrades steadily with age through presbycusis, a complex, multifactorial process involving changes in the cochlea, the stria vascularis, and the auditory nerve. This is physiology, not taste, and it follows broadly the same trajectory for anyone of a given age with a normal hearing history.

It also tells you almost nothing about the rest of this article, because none of what follows depends on detecting a tone in isolation. It depends on comparing.

Timing Resolution: The Ear Is a Precision Instrument

The auditory system's sensitivity to time is far finer than most people expect, and it underlies much of what makes reproduced music sound "right" or "wrong" in ways that resist simple description.

Interaural Timing: Microseconds Matter

Interaural timing sensitivity, the comparison of arrival time between the two ears used for localisation, operates down to roughly ten microseconds of difference in trained listeners, and lower still under some experimental conditions. That is not a typo. The auditory system routinely resolves timing differences an order of magnitude smaller than a single sample period at standard digital audio sample rates. This is precisely why even small, frequency-dependent timing errors in a signal chain are not automatically inaudible just because they look negligible on paper.

Digital Timing: Jitter as Temporal Smear

In the digital domain, this timing sensitivity shows up as jitter: small, unwanted variations in the timing of digital samples relative to their intended clock. Jitter does not distort a waveform's amplitude directly. It smears the timing of when each sample is reconstructed, which manifests as sidebands around the original signal and a subtle loss of image focus and transient precision rather than an obvious tonal change. We treat the mechanism, its sources, and its history in detail in Jitter Is the New Wow and Flutter, and its practical remedies in External Master Clocks and Word Clock Impedance.

Decades of AES-published research have established audibility thresholds for jitter that vary by jitter type and frequency content, generally in the range of tens of nanoseconds for random jitter, and considerably less for periodic jitter correlated with the music signal, which the ear is better equipped to detect as a distinct artefact than as generalised noise. This is the physical basis for taking master clock quality and clock cable impedance seriously, rather than treating them as superstition.

Analogue Timing: Group Delay and Transient Response

In the analogue domain, timing shows up as group delay and transient response. A crossover, a feedback network, or a cable's parasitic capacitance and inductance can all introduce frequency-dependent delay, meaning some frequencies arrive at the ear microseconds or milliseconds later than others, even though the amplitude response looks flat.

The ear does not need to consciously identify "delay" to register the consequence. It registers it as smeared transients, a loss of the sharp leading edge on a struck note, or instruments that sound slightly less located and slightly more diffuse than they should. Perceptual studies on crossovers and filters have repeatedly shown that large enough group delay errors produce audible degradation, particularly on transient-rich material, even when steady-state frequency response appears unchanged.

Phase Shift Is Timing, Not Volume

Phase and polarity are frequently confused, and the distinction matters here. Polarity is a simple inversion, positive and negative swapped, with no time component at all. Phase is a shift in a waveform's position in time relative to another waveform, expressed in degrees of a cycle, and it is frequency-dependent: the same absolute time delay produces a different phase shift at different frequencies. We set out the full physical distinction, including why the two are so often conflated and how to tell them apart by ear and by measurement, in Phase vs Polarity.

Phase Coherence and Transient Reconstruction

This matters for two audible reasons. First, non-minimum-phase behaviour in crossovers, drivers, or electronics changes the relative timing of different frequency bands arriving at the ear. A well-designed, phase-coherent monitor keeps the arrival timing of bass, midrange, and treble aligned, so a transient reconstructs as a single coherent event rather than several smeared, staggered ones. This is directly audible as tightness and "rightness" of transients, and it is measurable with a phase or group delay plot, not a matter of opinion. It is also the design principle behind Lipinski Sound's phase-coherent monitoring, engineered by Professor Andrew Lipinski around a minimal-feedback, reactive-component-free signal path specifically to preserve this alignment.

Stereo Imaging as an Interchannel Phase Relationship

Second, phase relationships between the left and right channels are much of what stereo imaging actually is. A sound source's apparent position depends on consistent interaural phase relationships across frequency. Cables, connectors, or components that introduce different phase behaviour on one channel versus the other, even subtly, degrade image specificity and stability in ways a frequency response measurement alone will not reveal. This is a real, physical mechanism, not an audiophile abstraction, and it is one of the more legitimate explanations for why two components measuring similarly in frequency response can still produce a different sense of image lock.

Soundstage: A Comparative Process, Not a Detection Task

Localising sound in space uses interaural time differences at low frequencies, interaural level differences at higher frequencies where the head shadows one ear from the other, and the head-related transfer function, the specific filtering the pinna and head apply depending on arrival angle. Layered on top is the precedence effect, where the first arrival of a sound dominates perceived direction even when a delayed reflection follows a few milliseconds later. This is what lets a listener localise sound accurately in a normal, reflective room instead of hearing a directionless blur.

None of this is a single-number measurement. It is the brain fusing timing, level, and spectral cues across two ears continuously, largely below conscious awareness, using exactly the timing and phase sensitivity described above. Because it depends on an individual's head and ear geometry, and sharpens with musical training and listening experience, real variation between listeners on the same system in the same room is expected, not suspect. Minimum audible angle experiments consistently show measurable, repeatable differences in localisation acuity between individuals.

Why "Bad Hearing" Does Not Disqualify Judgment

Because there is no laboratory-issued "correct" soundstage the way there is a correct answer to a tone-detection test, judging a reproduced image requires a reference. The only genuine reference is unamplified acoustic music heard live, in a real room, with no electronics involved, because that is the one situation where the interaural cues reaching the ears are the actual physical cues generated by real instruments in real air. A system is attempting to reconstruct a plausible version of those cues from two, or a few, channels. Knowing what real spatial cues feel like from direct experience turns soundstage evaluation from a vague aesthetic impression into a comparison against something physically real, even though the final judgment remains personal, because no two listeners' ears or listening histories are identical.

This is precisely why "bad hearing," in the narrow sense of a raised high-frequency threshold, does not disqualify someone from judging soundstage or imaging. Those judgments rely on comparative timing and phase cues, not on detecting extreme tones.

So Yes, Differences in Gear and Cables Are Audible

Put the pieces together and the conclusion follows directly from the physiology rather than from marketing language. If the ear resolves interaural timing to single-digit microseconds, if jitter in the tens of nanoseconds is documented as audible, and if phase and group delay errors measurably shift the arrival timing of different frequency bands, then components and cables that differ in these specific electrical respects (output impedance, capacitance, inductance, clock stability, channel-to-channel phase matching) can produce genuinely audible differences.

This is not a claim that every difference reported in audio is real. Confirmation bias, expectation effects, and level-matching errors are well documented, and they inflate a large share of reported differences, particularly in sighted, uncontrolled comparison, a subject we cover at length in What You Hear, What You Measure. Both facts are true at once: real electrical differences that map onto documented audible mechanisms exist, and a large amount of what gets attributed to them in casual listening does not survive controlled comparison. The way to tell them apart is the same as anywhere else in this field: identify a plausible electrical mechanism, and where possible confirm it against a controlled comparison or a measurement, rather than treating a vivid impression alone as proof.

At love cable / Tonmeister, this is the working assumption: cables should be neutral conduits that preserve timing, phase, and dynamics, not tone-shaping accessories. We design and build with that constraint in mind, controlled impedance, consistent geometry, and careful attention to parasitics, so the cable does not become the weakest link in the timing and phase chain. It is the same reasoning set out in Signal Integrity and, on the recording side of the same coin, in The Hidden Architecture of Music.

Rooms Have a Physical Ceiling Before Equipment Does

Timing, phase, and imaging all assume a room that is not actively working against the signal, and rooms have hard physical limits of their own, including a ceiling on how loud they can be played before things start to fall apart. The foundations of how rooms shape sound are set out in Acoustical Basics.

Room Modes: The Bass Is the Room

Every room supports standing waves, or room modes, at frequencies determined by its dimensions. A simple rectangular room has axial modes at frequencies where a room dimension equals a multiple of a half wavelength, producing peaks and nulls that can vary by ten to twenty decibels or more at different listening positions, entirely independent of the loudspeaker's own performance. Below the room's Schroeder frequency, the point where modal density becomes high enough that these individual peaks blur into a statistically smoother reverberant field, bass response is dominated by the room, not the equipment. In small rooms this crossover point can sit as high as 100 to 200 Hz, meaning a substantial part of the musically important low-frequency range is being shaped primarily by room geometry. We cover the practical consequences of this, and how multiple subwoofers can smooth it, in The Subwoofer Question.

Dynamic Headroom Is Frequency-Specific

This has a direct consequence for how loud a room can be played cleanly. Because modal peaks concentrate acoustic energy at specific frequencies at specific locations, the frequencies landing on a peak reach high sound pressure levels well before the overall programme level would suggest, straining driver excursion and amplifier headroom at exactly those frequencies while the rest of the spectrum still has margin to spare. The audible result is often described as boominess, congestion, or a bass that "gives up" before the rest of the system does, and it is a room effect layered on top of, not separate from, the equipment's own dynamic headroom.

There is also boundary reinforcement, sometimes called room gain, where nearby walls and corners add low-frequency output because they limit the directions sound can radiate into. This raises effective bass output at low frequencies, which sounds like extension but is really the room adding energy the loudspeaker did not generate. Combined with modal peaks, this means the point at which audible distortion or compression first appears at a given playback level is frequency-specific and room-specific, not a single volume knob position where "the room saturates." Some frequencies in a given room will show audible strain at a moderate overall level while others remain clean well beyond it.

Room treatment, careful loudspeaker and listening position placement relative to modal patterns, and matching amplifier headroom to both the loudspeaker's sensitivity and the room's own gain characteristics are what push this ceiling higher. But the ceiling is real, set by the room's physical dimensions and boundary conditions before a single component is chosen, and no amount of equipment quality removes it. Understanding where it sits in a specific room is part of the same disciplined, mechanism-first approach that applies to timing, phase, and soundstage: identify the physical cause, then address it, rather than treating the symptom as an unexplained limitation of the gear.

Our stance: mechanism first, then measurement, then listening. We treat hearing as a multi-dimensional, comparative process. High-frequency tone detection is one narrow slice of that process, not the definition of "good hearing." Timing resolution in the microsecond range, phase coherence across bands and channels, and individualised spatial cues are all central to how music is actually experienced. From that perspective: cables and components that alter timing, phase, or channel matching in measurable ways can be audible, because they act on mechanisms the ear is known to use. Claims that ignore mechanism and rely only on subjective impression are incomplete. Claims that deny any possibility of audibility because "the frequency response is flat" are equally incomplete, because they ignore timing, phase, and spatial processing.

From Physics to Placement

Everything above establishes why soundstage is a timing problem rather than a volume problem, and why the ear is equipped to judge it with a precision that has nothing to do with how high it can hear a test tone. What follows applies that principle to the one variable that costs nothing and changes everything: where the loudspeaker sits in the room.

A loudspeaker does not produce a soundstage on its own. It produces a sound field, which is then shaped, correctly or incorrectly, by its position in the room, its angle relative to the listener, and the reflections the room adds to it. The recording contains the spatial information described in the previous sections. Placement and acoustics decide whether that information survives the trip to the listening chair intact.

This is why a two-centimetre shift in loudspeaker position can matter more than an equipment upgrade costing thousands of euros. As established above, interaural time differences on the order of tens of microseconds are enough to shift a phantom image, and a reflection arriving even a few milliseconds after the direct sound can widen, narrow, or entirely collapse a stereo image, long before it becomes audible as a distinct echo. Correcting this is not a matter of buying differently. It is a matter of placing correctly.

What follows applies this across three fundamentally different ways of generating sound: the dynamic cone and dome driver, the electrostatic or magneto-static (planar) panel, and the horn-loaded compression driver. Each radiates into the room differently, and each therefore demands a different placement discipline to produce a convincing, stable, three-dimensional image.

Four variables recur across all three categories, and understanding which one dominates for a given loudspeaker type tells you where to spend your effort:

• Arrival order. The precedence effect gives priority to the first-arriving wavefront when the auditory system judges direction, treating later arrivals mainly as information about the space rather than about direction. This is why early, strong reflections from a nearby side wall or an equipment rack can compete with the direct sound and blur imaging, while a later, more diffuse reflection can enhance the sense of space without destroying image position.
• Time coherence. A source that reproduces all frequencies with a consistent time relationship, so a transient rich in harmonics arrives as a single coherent event rather than a smeared collection of frequency bands, images with more precision than one that does not. This depends on driver alignment, crossover design, and how the different frequency bands of a multi-way loudspeaker combine in time, not only in amplitude, exactly the mechanism described above for phase coherence. A loudspeaker can measure flat on-axis and still smear a transient if its drivers are not time-aligned.
• Dispersion. How a loudspeaker disperses sound, narrowly or broadly, forward-only or in multiple directions, determines which room surfaces get excited first and by how much energy. This single variable is the reason a dynamic box speaker, an electrostatic panel, and a horn system each demand a distinct placement strategy even when reproducing the identical recording in the identical room.
• Symmetry. The two channels must present the room with matched conditions. Unequal distances to side walls, unequal toe-in, or one loudspeaker nearer to a corner than the other will unbalance the timing and level relationships the stereo illusion depends on, regardless of how accurate either loudspeaker is in isolation.

Dynamic (Cone and Dome) Loudspeakers

The conventional multi-way box loudspeaker, using cone woofers and dome or cone tweeters mounted on a flat or sloped baffle, is the most common design in high-end audio, and it behaves acoustically closer to a point source than either a planar panel or a horn. This has direct consequences for placement.

Driver alignment and the crossover. In a multi-way dynamic loudspeaker, the low-frequency, midrange, and high-frequency drivers each have a different acoustic centre, the effective point in space from which their sound radiates. If these acoustic centres are not aligned, and the crossover network does not preserve time relationships between bands, the arrival of a broadband transient becomes smeared across a few hundred microseconds to a few milliseconds. This is inaudible as a separate event, but audible as a loss of image sharpness and a flattening of depth. Loudspeakers designed with first-order or minimal-order crossovers, physically stepped or sloped baffles to align acoustic centres, and attention to step response, not only frequency response, are specifically addressing this. It is one of the more overlooked reasons why two loudspeakers with near-identical frequency response can image very differently.

The point-source advantage. Because a well-designed dynamic loudspeaker approaches a coherent point source above the bass region, its imaging is comparatively forgiving of listener head movement and produces a stable centre image without requiring an unusually wide or narrow listening triangle. This makes the classic equilateral triangle, listener and both loudspeakers equidistant from one another, a reliable starting point rather than merely a rule of thumb.

Toe-in and the tweeter axis. Because dynamic tweeters, especially dome types, narrow their dispersion above a few kHz, toe-in angle directly controls the balance between direct sound and side-wall reflection. Aiming the tweeter axis directly at the listener's ears typically produces the most precise centre image and the best high-frequency energy balance, but can also expose recordings to a slightly forward, occasionally fatiguing top end in a live room. Aiming the axis to cross just in front of or just behind the listener softens this without materially harming image focus, since on-axis and off-axis response of most well-engineered dome tweeters differs by only a few dB within that narrow window. Differences of two to three degrees, corresponding to a few centimetres of cabinet rotation at typical listening distances, are audible.

Boundary distance. Dynamic loudspeakers, particularly those with rear-facing ports, are sensitive to distance from the front wall behind them. Too close, and the front wall reflection reinforces the upper bass, producing a boomy, ill-defined foundation for the soundstage; the low frequencies that should sit behind and below the image instead smear forward and mask it. Moving the loudspeaker forward, even by ten to twenty centimetres, decouples this reinforcement and typically restores a cleaner, deeper image, at the cost of bass extension that a well-treated room or a subwoofer can restore separately. Distances to the side walls should be unequal to the distance to the front wall, and to each other's proportional distances, to avoid reinforcing the same reflection paths on both channels simultaneously, a direct consequence of the comb-filtering behaviour described under Phase vs Polarity.

Electrostatic and Magneto-Static (Planar) Loudspeakers

Electrostatic panels and magneto-static (planar magnetic, sometimes called ribbon-adjacent) loudspeakers radiate from a large, thin, low-mass diaphragm rather than a small, heavy cone. This changes the physics of dispersion fundamentally, and with it, the placement rules.

Dipole radiation. Most electrostatic and magneto-static panels are dipoles: they radiate sound from both the front and rear of the diaphragm, with the rear wave out of phase with the front. The resulting pattern is figure-eight rather than the roughly hemispherical pattern of a forward-firing dynamic cabinet. This has two consequences. First, almost as much energy is fired directly into the wall behind the loudspeaker as toward the listener, so the front-wall reflection is not a minor contributor to the sound, it is a second, delayed source of comparable strength. Second, because the front and rear waves are out of phase, they cancel along the plane of the panel, to the sides, producing very little energy toward the side walls compared to a dynamic loudspeaker of similar size. Side-wall first reflections, a major concern for dynamic boxes, are consequently less critical for planar dipoles, while the rear-wall reflection becomes the dominant variable.

Distance from the front wall. Because the rear-firing wave must travel to the wall and back before combining with the direct sound, its arrival is delayed by an amount directly proportional to the distance from the panel to the wall. A panel placed close to the wall produces a rear reflection that arrives quickly and combines destructively at specific frequencies, comb filtering, typically audible as a thin, cupped midrange and a soundstage that clings to the plane of the speakers rather than extending behind it. Pulling the panel further into the room, often sixty centimetres to well over a metre depending on the specific design, delays this reflection enough that it is perceived as ambient information rather than as coloration, which is precisely what allows a dipole to produce its characteristic deep, spacious image when properly set up. This is the single most consequential placement variable for this loudspeaker category, more so than toe-in or triangle geometry.

Narrow vertical and horizontal sweet spot. A tall electrostatic or planar panel behaves, in the vertical plane, closer to a line source than a point source over part of its range, and its horizontal dispersion, particularly at higher frequencies, is often narrower and more directional than a dynamic dome tweeter's. The practical result is a smaller listening window: ear height relative to the panel, and precise toe-in angle, matter more, and are less forgiving of a few centimetres of error, than with most dynamic designs. Many panels are deliberately angled inward more sharply than a dynamic loudspeaker would require, sometimes aimed to cross well in front of the listening position, to compensate for their narrower dispersion and to keep the direct sound dominant over the room's contribution.

Room symmetry becomes more critical. Because so much of a dipole's output is directed toward the front wall, and in many room layouts toward whatever furnishing sits behind the panel, differences in the two rear-wall boundary conditions (one loudspeaker near a window, the other near a bookshelf, for example) will produce two channels with different rear-reflection character. This asymmetry is more audible with dipole radiators than with the more front-focused dynamic loudspeaker, since a larger fraction of total output is involved.

Horn-Loaded Loudspeakers

Horn systems, whether full-range multicellular horns, compression-driver horns crossed to a cone woofer, or the tractrix and exponential geometries common in high-efficiency designs, control dispersion deliberately through the geometry of the horn itself, rather than relying on the natural directivity of a bare diaphragm. This is their defining characteristic, and it inverts several of the placement priorities that apply to dynamic and planar designs.

Controlled directivity. A well-designed horn maintains a specified, constant coverage angle, commonly somewhere between 60 and 120 degrees horizontally and a narrower angle vertically, across most of its operating range. Because the horn itself, not room diffraction or dome directivity, is setting the dispersion pattern, the ratio of direct sound to reflected sound at the listening position is far more predictable and, in a well-designed system, far more favourable than with a typical dynamic loudspeaker of similar size. This is a principal reason horns were historically favoured in larger, less acoustically controlled spaces such as cinemas and studios: they concentrate energy on the listening area and correspondingly reduce the energy sent toward the side walls, ceiling, and floor.

Reduced sensitivity to side-wall treatment, increased sensitivity to horn axis. Because a horn's pattern is tightly controlled, small toe-in errors have an outsized effect: rotating a horn a few degrees off-axis can move the listener from the flat, extended portion of the horn's coverage into its rolled-off edge, altering tonal balance far more abruptly than the equivalent rotation on a dome tweeter or a planar panel. Horn placement therefore rewards precise, repeatable toe-in, frequently aimed to cross slightly behind or exactly at the listening position, more than it rewards distance experimentation from the side walls, since those reflections are already comparatively low in energy by design.

Horn-woofer time alignment. In two-part horn systems, where a compression driver is horn-loaded above the crossover point and a cone woofer, often loaded into a bass horn or a sealed or ported enclosure, covers the low end, the compression driver's acoustic centre is frequently set back from the woofer's due to the physical length of the horn and driver mounting. Left uncorrected, this produces the same kind of transient smear discussed for dynamic loudspeakers, and for high-efficiency horn systems, whose entire appeal rests on speed and immediacy, this error is particularly audible. Physical stepping of the baffle, electronic delay in an active crossover, or careful mechanical design to equalise path length are the standard corrections, and the resulting alignment should be verified, not assumed, since it is a common source of underperformance in horn systems assembled from separately sourced components.

Boundary reinforcement, deliberately used. Many bass horns are specifically designed to use a room corner, or the floor and front wall junction, as part of the horn's own loading, extending its effective length acoustically without extending its physical size. This means placement guidance for a horn subsystem can be the opposite of the guidance for a sealed or ported dynamic woofer: rather than moving the enclosure away from boundaries to reduce reinforcement, corner or wall coupling is often required for the horn to load correctly and reach its designed low-frequency extension. Manufacturer documentation for the specific design should always take priority here, since folded and corner-loaded horns vary considerably in their intended boundary relationship.

A Comparison of Radiation Behaviour

The Practical Method: Small Moves, Large Consequences

Across all three categories, the working method is the same, even though the variable that matters most differs.

1. Establish symmetry first. Measure, do not estimate, the distance from each loudspeaker to the nearest side wall, the front wall, and the listening position. Small asymmetries between the two channels undermine every subsequent adjustment.
2. Set the triangle, then refine by type. Begin with an equilateral or near-equilateral listening triangle for a dynamic loudspeaker, a somewhat wider or more steeply toed-in triangle for a planar dipole to compensate for its narrower dispersion, and precise axis-aiming, verified at ear height, for a horn.
3. Move in small steps. Five to ten centimetres at a time for boundary distance, one to two degrees at a time for toe-in. The nonlinearity of reflection timing and dispersion means these small steps can produce disproportionately large changes in image focus, particularly for dipole and horn designs.
4. Change one variable at a time. Distance from the front wall, side wall distance, toe-in, and listener height all interact, but adjusting more than one at once makes it impossible to identify which change produced which result.
5. Use familiar material. A recording you know well, ideally one with a simple, well-defined stereo or mono centre image such as a solo voice or a small acoustic ensemble, will reveal image shift and focus far more reliably than an unfamiliar or heavily processed recording.
6. Let measurement confirm what listening suggests, not replace it. Time-domain and frequency-domain measurement at the listening position can confirm boundary reflection timing, useful in diagnosing the issues described above for dipoles and ported dynamic designs, but the final judgment of a convincing soundstage remains a perceptual one. If something still sounds wrong once symmetry, triangle, and boundary distance are confirmed, the systematic method for isolating the remaining fault is set out in Troubleshooting and Diagnostics.

Measuring What You Hear: A Note on Method

Point six above says to let measurement confirm what listening suggests, not replace it. That guidance only holds if the measurement itself can be trusted, and the single most common reason it cannot is a variable most people never check: the room's own noise floor.

Silence Is Not Silent

A domestic listening room is acoustically busy even with nothing playing. Refrigerator and HVAC compressors cycle on for minutes at a time, adding broadband energy concentrated below 200 Hz, in the same territory as the room modes discussed above. Mains hum at 50 Hz and its harmonics, traffic, and structure-borne building noise all add further low-level content. A frequency-response or impulse-response measurement taken without first characterising this baseline folds all of it into the result indistinguishably from the room's genuine acoustic behaviour. A narrowband spike from a compressor cycling at the wrong moment can look exactly like a room mode on screen. It is not one, and no amount of loudspeaker repositioning will move it.

Before measuring anything else, measure nothing: playback off, doors and windows in their normal listening state, logged as a real-time spectrum over at least one full appliance cycle, using the same microphone, gain, and position that the subsequent measurement will use. This baseline is what makes everything measured afterward interpretable, particularly the boundary-reflection timing and low-frequency behaviour that the placement guidance above depends on.

The Microphone Itself

A consumer or recording-oriented microphone is not a measurement instrument. What is needed is a microphone with a documented, ideally individually calibrated frequency-response correction file, an omnidirectional pattern that holds up well into the low kHz range, and self-noise low enough that it does not become the limiting factor when resolving a quiet noise floor. None of this needs to be expensive. It needs to be correct, and an uncalibrated instrument, however well regarded, is measuring itself as much as it is measuring the room.

Pink Noise, Sweeps, and Signal-to-Noise

Pink noise, with equal energy per octave, is well suited to a quick by-ear tonal balance check. It is a continuous, averaged signal, and any noise present during the measurement window is averaged directly into the result with no way to separate it out afterward. A logarithmic sine sweep, deconvolved against the known reference signal into an impulse response, produces substantial processing gain by comparison, resolving genuine room and boundary-reflection behaviour meaningfully below the level of uncorrelated ambient noise. It is also what makes the time-domain analysis behind the reflection-timing discussion above possible in the first place: gating the impulse response separates the direct sound from the first reflection, which is precisely the distinction that determines whether a given boundary distance is helping or hurting the image. As a working rule, treat measured data as trustworthy only where the signal sits at least 10 to 15 dB above the logged noise floor at that frequency.

Position, Not Just a Point

A single measurement at a single point describes that point only. This is the same symmetry principle behind the practical method above, applied to the microphone rather than the loudspeaker: take readings at the primary seat and at several points spread across the listening area, and treat a feature that appears everywhere as genuine room or boundary behaviour worth acting on, and a feature that appears only at one spot as local interference that a small seat or panel adjustment may resolve more directly than any broader change.

Used this way, measurement does what point six asks of it: it confirms whether a placement change actually altered the reflection timing or image-relevant response it was meant to address, rather than producing a number that merely looks authoritative. A measurement taken without first establishing the noise floor cannot reliably answer that question, no matter how convincing the resulting plot appears on screen.

In Summary

Soundstage is not a fixed property of a loudspeaker, and it is not a matter of taste that measurement has nothing to say about. It begins in the ear, which resolves interaural timing to single-digit microseconds and treats phase as a frequency-dependent timing relationship, not a matter of volume. It continues in the room, where a radiation pattern meets a specific position and angle, and that meeting can be adjusted with a precision measured in centimetres and degrees, not in equipment purchases. A dynamic loudspeaker rewards attention to driver time alignment, boundary distance, and a balanced triangle. A planar dipole rewards distance from the wall behind it above nearly every other variable. A horn rewards precise, repeatable axis aiming and correct time alignment between horn and woofer sections above concern for side-wall treatment. In every case, the underlying physics is the same: the ear assembles a spatial illusion from the first arriving wavefront and the timing of everything that follows, and placement, alongside a signal chain that does not introduce timing and phase errors of its own, is what lets that illusion survive the trip from the recording to the listening chair.

At love cable / Tonmeister, this disciplined, mechanism-first approach, physics before measurement, measurement before conclusion, is the same standard we apply to cable and component design. Our job is to keep the signal path as neutral and mechanism-respecting as possible, so what you hear is the music and the room, not the cable.