The Listening Room

Aster

Most of what a system sounds like in a room is decided by where the speakers are standing. Not by cables, not by the last upgrade. Here is the method, with the arithmetic shown, so you can check it rather than take it on trust.

Everything below needs a tape measure, two tracks and a free afternoon. Nothing below needs anything bought.

Step 0

Measure and Mark

Find out what the room does before anything is carried into it. Dimensions set which notes get reinforced, and the midpoint between two walls is where the first one cancels.

Try this

A 12-foot-wide room has a favourite note. It is 47 Hz.

Every pair of parallel walls reinforces one frequency hardest: the speed of sound divided by twice the distance between them. For a 12-foot width that is 1130 divided by 24, which is 47 Hz — roughly the open E on a bass guitar. That note will sound louder in that room than the recording made it, no matter which speakers are in it.

Room modes are set by dimensions, not by equipment. The first axial mode between two parallel walls is c/2L, where c is the speed of sound, about 1130 feet per second at room temperature. A 12-foot width gives 1130/24 = 47 Hz. A 15-foot depth gives 1130/30 = 37.7 Hz. An 8-foot ceiling gives 1130/16 = 70.6 Hz. Each of those frequencies is reinforced by the room itself. The practical consequence is that the same speakers measure differently in two rooms of different sizes, and the fix is placement rather than replacement.

1130 divided by twice the width gives the note: 47 Hz in a 12-foot room.

Try this

The worst seat for bass in a 15-foot room is 7.5 feet back

The first mode between the front and back walls fits half a wavelength into the room, which puts its pressure null exactly at the midpoint. In a 15-foot-deep room that is 7.5 feet — dead centre. A seat there hears the least bass of anywhere along that axis, which is why a sofa pushed to the middle of a room can make a capable system sound thin.

A half-wavelength standing wave has maximum pressure at the boundaries and a node — a pressure minimum — at its centre. For the first front-to-back axial mode, that centre is the middle of the room. In a 15-foot-deep room the null sits at 7.5 feet. At that position the fundamental is cancelled rather than reinforced, so the bass reads as weak and uneven even though the system is producing it normally. Moving the seat a foot or two either way is often the entire fix. This is also why Step 0 of the technique marks the floor before anything is carried in: the null is easier to avoid than to diagnose once a system is already set up.

In a 15-foot-deep room that null sits at 7.5 feet.

Method

Clapping your way around a room finds the dead zones in two minutes

Stand at the front wall and clap once. Within a foot or two of the wall it sounds muffled — reflections are coming back fast enough to tangle with the clap itself. Step out slowly and somewhere around two to three feet it opens up. That edge is the closest a speaker should ever sit. No measurement rig, no app, about two minutes.

Close to a boundary, a reflection returns with a delay short enough to interfere with the direct sound rather than be heard as a separate event. The audible result is a clap that sounds choked and a hummed note that goes uneven. Walking outward from the wall, both clear up, typically between two and three feet out. The same walk repeated with a low hum, a called-out word and a stomp covers different parts of the spectrum: the hum tests low-frequency evenness, the call tests midrange clarity, the stomp tests whether bass is consistent underfoot. The technique treats this as Step 0, before anything is measured with instruments, because it establishes the nearest usable boundary for a speaker in a couple of minutes.

The clap clears somewhere between two and three feet out from the wall.

Try this

A square room is harder than a big room or a small one

Width and depth each reinforce one note, calculated the same way. In a 12-by-12 room both land on 47 Hz, so a single frequency gets two modes' worth of reinforcement instead of one. Change either dimension and the two split apart, spreading the load across the spectrum. Square is the one shape where the room's problems stack instead of spreading.

Each pair of parallel surfaces contributes its own axial mode at c/2L. In a rectangular room of distinct dimensions those modes fall at different frequencies and the reinforcement is spread out. In a 12-by-12 room, width and depth both give 1130/24 = 47 Hz, and their second-order modes both give 94 Hz. The same note is boosted twice. Perceptually that reads as one-note bass: a specific pitch that booms while notes either side of it sound recessed. Non-square rooms are not automatically good, but square rooms are reliably harder, and the difficulty comes from the geometry rather than from anything in the signal chain.

In a 12-by-12 room both axial modes land on 47 Hz, and both second-order modes on 94 Hz.

Scale

A 20 Hz note is 56 feet long. A 10 kHz note is 1.4 inches.

Wavelength is the speed of sound divided by frequency, and the range across the audible band is enormous. At 20 Hz a single wave is 56 feet long, far longer than the room containing it. At 10 kHz it is 1.4 inches. That ratio is why bass behaves like pressure filling a space while treble behaves like a beam you can aim.

At 1130 feet per second, wavelength runs from 56.5 feet at 20 Hz down to 0.113 feet — about 1.4 inches — at 10 kHz. A 65 Hz note is 17.4 feet. A 500 Hz note is 2.3 feet. This single ratio explains why the two halves of the spectrum need different treatment. Wavelengths longer than the room cannot form a free-field wave at all; they load the room as pressure, which is why bass depends on where the speaker sits relative to the walls. Wavelengths of a couple of inches are short enough that moving your head changes what arrives first, which is why imaging depends on angle and timing. The technique splits along that line: Step 1 places for bass, Steps 2 and 3 place for everything above it.

At 1130 feet per second, 20 Hz measures 56.5 feet and 10 kHz measures 1.4 inches.

Method

The gap between the speakers should not exceed twice your distance to them

An equilateral triangle puts you at the front edge of the sound field. Pulling the speakers slightly closer together than your seat is from them moves you back a few rows and widens the stage, which most people prefer. Push the base past twice the sides and the phantom centre stops holding — you hear two speakers instead of one image.

The listening triangle is a taste control with a geometric limit. An 8-8-8 arrangement is equilateral and reads as a close, front-row perspective. Something like 8-8-7, slightly acute, moves the apparent vantage back and widens the stage while keeping the centre stable. The failure mode is at the other extreme: as the base grows relative to the sides, the included angle at the listener widens, and past roughly twice the side length the phantom centre image collapses. Separation increases but the sense of a single coherent event does not survive it. This ratio is a rule of thumb from the technique rather than a measured threshold, and the recommendation is to treat it as an upper bound rather than a target.

The technique states the base should not exceed twice the sides, and gives 8-8-8 and 8-8-7 as worked arrangements.

Method

Mark the floor before you carry anything in

Three pieces of tape do most of the work: one where your ears will be, one for each speaker. Marking first turns placement into a bounded search rather than an open one, and it means the mid-room null and the boundary dead zones are identified while they are still cheap to avoid rather than after a system is standing in one.

Step 0 of the technique produces a workable zone rather than a position. Width and depth are measured, the listening position is fixed, and the boundary dead zones are located by the clap walk, which typically rules out the first two to three feet from a wall. What remains is the region in which everything later happens. The three marks — Center for the listening position and a Left and Right Reference for the speakers — give the later steps a coordinate system, which is what makes millimetre-scale refinement meaningful later on. Without them, each adjustment is relative to an unrecorded previous state, and a position that was better two moves ago cannot be returned to.

The three marks are named Center, Left Reference and Right Reference.

Step 1

Anchor the Angle

Distance from the front wall, judged on bass alone. Three qualities — low, loud, linear — and they do not peak in the same place.

Method

Three words decide where a speaker sits: low, loud, linear

Moving a speaker away from the front wall changes the bass in three separate ways, and they do not peak together. Low is how deep it reaches. Loud is whether the note survives rather than cancelling. Linear is whether every note carries the same weight — the hardest of the three, and the one most setups never reach.

Distance from the front wall controls how the speaker couples to the room's length mode. Too close, within about six inches, reflections return fast enough to blur detail. Too far and the speaker sits near a null where the fundamental cancels. Between those, three qualities vary independently. Low is the deepest frequency the room still supports at that position. Loud is amplitude — whether reinforcement or cancellation dominates. Linear is evenness across notes, and it is the demanding one: a position can be deep and strong while still making one note in a bass run jump out. The technique tunes on a slow ballad with an upright bass because spaced notes with audible decay let all three be judged separately, which a continuous electronic bassline does not.

The technique names three criteria — low, loud and linear — and calls linear the hardest of the three.

Method

You cannot tell you are on the peak until you have gone past it

Sliding a speaker outward, the bass improves, keeps improving, and then thins. The instinct is to stop at the first spot that sounds good. The method is the opposite: deliberately overshoot until it gets worse, then come back. A peak is only identifiable from both sides, which is why the technique treats overshooting as a required move rather than a mistake.

Position-versus-quality is not monotonic. As a speaker moves away from the front wall the bass passes through successive reinforcement and cancellation regions, so quality arrives in waves rather than climbing to a single maximum. A listener who stops at the first improvement has no way to know whether they are on a crest or on the way up to one. Going past it, hearing the bass thin, and returning establishes both edges. The technique escalates resolution as it converges: one-inch steps to find the region, half-inch steps to narrow it, then millimetres. It also caps marked candidates at two, because comparing more than two positions from memory is unreliable — the comparison has to be A against B, not A against a remembered impression.

The technique steps in inches, then half-inches, then millimetres, and marks at most two candidate positions.

Try this

When one speaker is near a corner, move the other one

In an asymmetric room the bass image pulls toward whichever speaker has more boundary reinforcement. The counterintuitive fix is to adjust the far speaker, not the near one. Pull it out until the bass moves about a quarter of the way across the stage, mark it, keep going until it reaches three quarters, mark that. The answer is inside that range.

Boundary loading is asymmetric when one speaker is closer to a corner or a side wall, and the low frequencies image toward the more strongly loaded side. Rather than pulling the near speaker away from its corner and giving up its reinforcement, the technique moves the far speaker to rebalance. Doing so sweeps the perceived bass position across the soundstage, and the two ends of the useful range can be identified by ear: the position where the image has travelled about 25 percent toward the far speaker, and the position where it has travelled about 75 percent. Those two marks bracket the working range. Within it, and only within it, the low-loud-linear judgement is made. The range matters more than any single position, because it converts a vague hunt into a bounded search.

The two marks are taken where the bass image has crossed 25 percent and 75 percent of the soundstage.

Method

An upright bass is a better test tone than a test tone

Test tones are continuous, which hides the thing worth hearing. A plucked upright bass on a slow ballad gives spaced notes with audible decay, and decay is where cancellation shows itself: a note that stops short rather than fading is being cancelled by the room. A steady tone gives a level, and a level cannot show you that.

Steady-state test signals measure amplitude at a frequency, which is useful for instrumentation and poor for the judgement this technique asks for. Room interference is a time-domain phenomenon as much as a level one, and its clearest audible signature is truncated decay — a note whose tail disappears earlier than it should because a reflection is arriving out of phase with it. The technique therefore sets two requirements for a test track, spaced notes and audible decay, and an upright bass on a slow ballad is the common recording that satisfies both. Spaced notes are needed because the tail has to be audible in the gap rather than masked by the note after it. The same applies to the linear criterion: whether every note in a run carries equal weight can only be judged on a passage that plays several different notes, which is why continuous electronic basslines and test tones are ruled out.

The technique sets two requirements for a test track: spaced notes and audible decay.

Step 2

Slide the Stage

Lateral spacing, judged on a voice. Too close and it doubles, too far and it thins. The gap between is narrower than most people expect.

Try this

If the singer's mouth sounds three feet wide, the speakers are too close together

Speakers too close together let their outputs overlap and comb-filter, which spreads and doubles a voice until it stops sounding like one person. Too far apart and the same voice thins and frays at the edges. Between them is a spacing where it snaps into a single solid image. For most rooms that lands somewhere between six and ten feet.

Midrange and treble wavelengths are short — about 2.3 feet at 500 Hz and 2.7 inches at 5 kHz — so small differences in path length from each speaker to the listener translate into audible interference. When the speakers are too close, the two arrivals overlap across a wide band and comb filtering smears the central image, which is perceived as an unnaturally wide or doubled mouth. When they are too far apart, the centre is underfilled and the voice thins and loses its edges. The technique moves one speaker laterally while holding its distance from the front wall fixed, judging only the apparent size and solidity of the voice, and notes that the good positions arrive in waves so the peaks are marked and the dips ignored.

500 Hz measures 2.3 feet and 5 kHz measures 2.7 inches, which is why small lateral moves are audible.

Method

The letter S tells you whether the treble is right

Sibilance is the most diagnostic sound in a vocal recording. Set up correctly, an S is crisp and unremarkable. Set up wrong, it either hisses and spits or goes dull and lisping. Because it sits where the ear is most sensitive and where interference is most audible, it moves before almost anything else does.

Sibilant energy concentrates roughly between 5 and 10 kHz, where wavelengths run from about 2.7 inches down to 1.4 inches. At that scale, changes in speaker position and angle of a few inches alter arrival timing enough to change the perceived character of the consonant. The technique uses S sounds as the check at two separate stages: while sliding speakers laterally, where excessive overlap adds hiss, and while setting toe-in, where too much on-axis energy sharpens it into harshness. The target is neither bright nor soft but simply unremarkable — the consonant should pass without drawing attention, which is what a natural recording of it does.

Sibilant energy sits roughly between 5 and 10 kHz, where a wavelength runs from 2.7 inches down to 1.4 inches.

Step 3

Tie the Toe

Angle, from zero to about thirty degrees. Set by leaning forward and back rather than by sitting still and guessing.

Try this

Lean forward. If it sounds better, angle the speakers outward.

Toe-in is hard to judge by sitting still, because the ear adapts. The technique uses a body test instead. Lean eight to twelve inches forward and listen, then the same distance back. Clearer and less harsh leaning forward means toe out. Fuller and smoother leaning back means toe in. Equal turns on both speakers.

Leaning forward and back changes the listener's position relative to each speaker's on-axis response in a way that mimics changing the toe angle, without touching anything. It converts an absolute judgement, which the ear is poor at, into a comparison, which it is good at. The mapping is direct: if the forward position sounds clearer and less harsh, the speakers are currently aimed too directly and should be angled outward; if the rearward position sounds fuller and smoother, they should be angled inward. The usable toe range runs from zero degrees, with the front baffle parallel to the front wall, to about thirty degrees, where each speaker aims at the listening seat. Adjustments are made in equal increments on both sides, because asymmetric toe twists the stage.

The usable range runs from zero degrees, baffle parallel to the front wall, to about thirty degrees, aimed at the seat.

Method

Pivot on the tweeters, not on the cabinet corner

Rotating a speaker around the wrong point changes its position as well as its angle, which quietly undoes the placement work already done. The technique pivots on the acoustic centre — the plane running through both tweeters — so toe-in changes angle and nothing else. It is a small distinction that keeps each step independent.

The whole method depends on moving one variable at a time. Distance from the front wall is set in Step 1 and lateral spacing in Step 2, and both are judged to millimetre resolution by the end. If Step 3 then rotates each speaker about an arbitrary point such as a rear corner, the tweeter translates as well as turns, and the earlier settings are no longer what they were measured to be. Pivoting about the plane through the tweeters keeps the acoustic origin fixed while the radiation angle changes, so the toe adjustment is isolated. The same reasoning drives the order of the steps: bass placement first because it constrains the most, then lateral spacing, then angle, then height, each one a smaller move than the last.

The pivot is the plane through both tweeters, which Step 3 of the technique calls the acoustic centre.

Try this

A reflection off a wall three feet away arrives 5 milliseconds late

Sound reaching you off a side wall travels further than sound reaching you directly. For a wall three feet to the side that detour is about six feet, and at 1130 feet per second it arrives roughly 5.3 milliseconds after the direct sound. Early enough to fuse with it rather than be heard separately, which is why side walls change imaging instead of producing an audible echo.

The extra path length of a first side-wall reflection is approximately twice the distance from the listening axis to the wall. At three feet that is six feet, giving 6/1130 = 5.3 milliseconds. Delays in this range fall inside the window where the ear fuses the reflection with the direct sound rather than perceiving it as a discrete event, so the effect is not an echo but a change in apparent source width and position. This is why toe-in matters beyond simple aiming: the angle determines how much energy is directed at the side walls, and therefore how strong that fused reflection is. Correctly set, it widens the stage. Badly set, it smears the centre image.

Six feet of extra path at 1130 feet per second gives 5.3 milliseconds.

Step 4

Equalize and Elevate

Height and tilt. This is the step that decides how large the good-sounding area is, rather than how good one fixed seat sounds.

Method

Sight across the tops of the speakers like iron sights

Height and tilt need to match between left and right, and a tape measure is slower and less reliable than an eye. Line your eye up so the top edge of the near speaker overlaps the far one and adjust until the two read as a single plane. Quarter-turns on the front spikes. The eye resolves alignment errors a ruler will miss on an uneven floor.

Rake is the tilt of the baffle relative to the floor, and it changes how the arrival times of the drivers line up at ear height. Left-right symmetry matters more than any absolute value, because a mismatch delays one side relative to the other and twists the stage. Measuring from the floor fails when the floor itself is not level, which is common. Sighting across the top edges compares the two speakers to each other rather than to the floor, which is the comparison that actually matters, and the eye is highly sensitive to a break in a straight line. Adjustment is by quarter-turn increments on the front spikes, with the rear spikes started fully retracted so the full range of tilt remains available.

Adjustment is by quarter-turns of the front spikes, with the rear spikes started fully retracted.

Try this

Slump in your chair. If it sounds fuller, raise the speakers.

The same comparison trick that sets toe-in also sets height. Hover six to twelve inches up and listen, then slump the same distance down. Clearer and tighter up means lower the speakers. Richer and fuller down means raise them. This is the adjustment that widens the sweet spot, so the sound survives you not sitting perfectly still.

Ear height in a normal chair runs about 36 to 40 inches, and driver alignment is referenced to it. Moving the head vertically changes the listener's position relative to the vertical radiation pattern and to floor and ceiling reflections, and the direction of improvement indicates which way the speakers should move. If the raised position is clearer and tighter, the speakers are sitting high relative to the ears and should come down; if the lowered position is richer, they should go up. Beyond tonal balance, this step governs how large the usable listening area is. A well-set rake widens the region over which the balance holds, which is what makes a system tolerant of normal movement rather than demanding a single fixed head position.

Ear height in a typical chair runs 36 to 40 inches, and the test moves the head 6 to 12 inches either side of it.

Step 5

Refine and Repeat

Everything again, smaller. Millimetres on position, half-degrees on angle. The step people skip.

Try this

Moving a speaker one millimetre changes arrival time by three microseconds

It sounds like superstition until the arithmetic is done. A millimetre is 0.0033 feet, and at 1130 feet per second that is about 0.003 milliseconds — three microseconds. Small, and not nothing: it is enough to shift where a cancellation sits relative to your ears, which is why the final step of the technique works at that scale.

Arrival time is distance divided by the speed of sound. One millimetre is 0.003281 feet, so at 1130 feet per second the change is 2.9 microseconds, about 0.003 milliseconds. Whether that is audible depends entirely on what it is added to. In isolation it is far below any threshold. Superimposed on an interference pattern where a null already sits close to the listening position, a shift of that order moves the null's position relative to the ears and changes what is cancelled. This is why the refinement step exists and why it works in millimetres on position, half-degrees on toe and sixteenth-turns on rake, after the earlier steps have already found the right region in inches.

One millimetre is 0.003281 feet, which at 1130 feet per second is 0.003 milliseconds.

Method

Good positions arrive in waves, so mark the crests and ignore the dips

Sliding a speaker, quality does not climb steadily to one best point. It rises and falls repeatedly as reinforcement and cancellation alternate. The method is to mark the crests, skip the troughs, and then compare no more than two candidates directly against each other rather than against a remembered impression.

Because room response at a given position is the sum of a direct arrival and many reflections, moving the source traces a series of alternating maxima and minima rather than a single optimum. Practically, a listener sliding a speaker hears the sound improve, degrade and improve again. Two habits follow. First, mark peaks as they occur instead of trying to hold them in memory, because auditory memory for timbre degrades within seconds. Second, cap the comparison at two candidates and switch directly between them; a three-way comparison from memory is unreliable enough to produce a worse final choice than a two-way one. The technique applies the same discipline at every stage, from inch-scale anchoring through to the millimetre refinement pass.

The technique caps the comparison at two marked candidates at every stage.

Notes on the numbers

Every figure on this page assumes the speed of sound at about 1130 feet per second, which is right at roughly 70 °F. It moves about a foot per second per degree, so a cold room shifts a 47 Hz mode by well under a hertz. Immaterial for placing a speaker, and not exact enough to quote to three decimals.

The triangle rule — that the gap between the speakers should not exceed twice your distance to them — is a rule of thumb rather than a measured threshold. It is a useful upper bound, not a target.

Adapted from William Lloyd Kline’s 2025 room-setup notes, informed by John Hunter’s teaching. Aster is our adaptation; it does not imply John Hunter’s endorsement. One figure in the original notes was corrected in the process: a one-millimetre speaker move changes arrival time by about 0.003 milliseconds, not 0.035. The original divided feet where it meant inches. The conclusion survives the correction — a millimetre is still an audible change — but the number did not, and it is stated correctly here.