The ultimate loudspeaker-in-room listening test — reflections and directivity v1.24

Walls & ceiling — a low / high

Two absorption coefficients, low band and high band, split at 1.5 kHz – the same split the reflection shelf and the two decay times of the tail use, so one number cannot drift from the other. They are typical published values rather than measurements of any particular wall, which is the same standing as every other number in this model.

The order of the two numbers is the point, not their size. Bare concrete absorbs almost nothing at either end (0.02 / 0.03), so its tail stays bright as it dies: that is the hard, raw decay of an empty concrete room. Curtains absorb the top four times harder than the bottom, so the tail goes dark and the bass outlasts it. Panelled surfaces run the other way round – the panel resonates and absorbs the low end harder than the high one (gypsum board 0.28 / 0.07, plywood panelling 0.22 / 0.09), which is why a stud-wall living room sounds brighter and thinner than a masonry one of the same size, and why it needs no bass trap so much as it is one.

The low figure is the value at 125 Hz, not an average across the band, and that is deliberate. What it drives most audibly is the modal region below roughly 170 Hz, where the mode bank gets its decay and its Q – and a stud wall really does absorb around 0.29 down there, which is exactly why such a room has less bass boom than a concrete box. The cost is that the same figure also colours reflections and the low half of the tail all the way up to 1.5 kHz, where 125 Hz is the wrong reference: a gypsum panel is down to 0.04 by 1 kHz, so its reflections come out about 1 dB darker than published data would give, and a carpet or a curtain rises to 0.5 or more, so theirs come out about 2.5 dB brighter. That is what the charts below show as the gap between the step and the curve. Reading it takes two numbers and no more; fixing it would take a third band.

Until this page had wall materials it had a single absorption slider, and the slider tied the amount of absorption to its distribution: more absorption always meant relatively more of it at the top. That is what soft rooms do and the opposite of what hard ones do, so the range of tail colours the model could reach was both narrow and back to front. It is now about 12 dB wide and it runs the right way.

Both numbers are drawn below, against published octave band data for the same kind of surface. The line with dots is the published data from 125 Hz to 4 kHz; the flat two-step line is what this model actually uses – one value below 1.5 kHz, one above. Where the step sits away from the curve, the curve is the published figure and the step is the simplification, so you can see what the model rounds off as well as whether the numbers are plausible. The two surfaces you are listening to right now are marked in use.

These are typical published values for that kind of construction, not measurements of any particular wall or floor, and handbooks disagree with each other by a good deal – the two standard rows for heavy carpet differ by 0.18 at 250 Hz on their own. Two rows have no standard table entry at all and say so. Read them as a sanity check, not as a datasheet.

Late tail — Reflections n % of the room budget · the tail fills the rest

100 % on the slider is not the loud end of a preference: it is the level that brings the whole room side – these reflections plus the tail – up to what a diffuse field of this room's absorption would be. So the slider is a percentage of a calibrated amount, and the line tells you how much of that amount the early reflections already supply; the tail fills the rest. That is also why the tail gets quieter in a big room where the reflections are weaker: the budget is shared, not stacked.

Two states can appear on the line, and they stay there rather than hiding here, because they are things happening rather than things explained. Capped means this room is live enough that the honest level would clip the output, so you are reading the room as quieter than it would really be – pick a more absorbent wall or floor material. Held at a floor means the room is so small and hard that the reflections alone already reach the diffuse estimate.

Turn — Left / Right / Face speakers

Turning your head is how people actually localise sound. Face away from the speakers and listen for whether the image follows you: front-back confusion is the weakest point of a generic HRTF, and hearing it is the most useful thing this control does.

Lock — Speakers follow · Head follows

Speakers follow keeps the triangle and moves the whole set-up through the room, which separates where you sit from how the triangle is shaped – two things you can otherwise only change together.

Head follows keeps you facing the speakers, so walking never turns into turning. Mute one speaker and it faces the one still playing, which is what you want when you are listening to a single speaker on axis. Worth knowing when you use muting as an A/B: the head turns with it, so the comparison is one speaker on axis against two in front of you, not the same seat twice.

Both are ordinary controls while they are on – the sliders show the real values, and switching off leaves everything where it is.

Speaker library

A VituixCAD polar export becomes an entry in the speaker list, with a name you give it. The files are read in your browser and the catalogue lives here; a measurement reaches the server only when you back it up to your account or tick it into a shared link.

There are two text fields, and they differ in who reads them. The description is part of the speaker: it goes wherever the speaker goes – into an export file, into the copy in your account, and into a shared link along with the measurement. It is the line that tells whoever opens your link what they are listening to, since the speaker’s name itself does not travel. The private memo is yours alone. It is backed up to your account along with the speaker, so it follows you to another machine, but it is left out of every export file and every shared link – the difference is not whether it moves, but whether anyone else can read it. If the measurement is left out of a shared link, the description stays behind with it.

Export it at the distance you listen from. A polar set is measured at one distance and the point sources then carry the real path lengths at that distance – the examples here were exported at three metres, which stops being honest much closer than two. If your seat is nearer than that, exporting at your own distance is the more accurate choice, on one condition: your ears have to stay level with the speaker, which is what the height lock keeps true by default.

The spheres at the bottom of the list are a different kind of example. Every other entry compares a design choice – a crossover, a driver spacing, a cabinet edge, where the drivers sit on the baffle. The spheres have no design choices at all: a point source on the surface of a sphere, and the only thing that changes from one row to the next is how big it is. That makes them the cleanest way to hear what a cabinet's size does on its own, baffle step included, with nothing else moving. Their pattern depends on one quantity – ka = 2πr/λ – so the whole set is one curve slid along the frequency axis, and the step moves down in frequency in inverse proportion to the diameter. Pick two that differ by a factor of two and the step should move by an octave.

One entry is not left–right symmetric, and it has its own heading for that reason. The classic three-way carries its midrange and tweeter off the centre line of a wide baffle, the way a great many real cabinets are built, so it radiates a different response towards one side wall than towards the other. Everything else in the list is symmetric, which makes this the only row where that is audible at all: the data is the left cabinet of the pair, and the right one plays it left–right reversed, which is how such a pair is normally built and what the page does with it. Two things follow that you can check. The horizontal polar draws both halves rather than mirroring one, and the note under the map tells you which of the two cases the loaded speaker is in rather than leaving you to assume it. And the side wall reflections stop being the same measurement read at two angles – which is the whole reason a cabinet like this is built as a mirrored pair in the first place.

It is playable the moment it is imported. There is nothing to solve and nothing to wait for: the measured curve is convolved into each path directly, as a short impulse response. With an account, Back up copies a speaker to your account so another machine gets it.

The same is true of a speaker that arrives with a shared link or a preset file – it plays from the moment the link opens.

Response — Directivity only and three phase models

A polar export contains two different things, and the page normally uses only one of them. Directivity is what the speaker does off axis – every angle in the export is stored as its difference from straight ahead, which is why 0° is exactly flat. The on-axis response is what it does straight ahead, and it is the part that was divided out. This list puts it back: the first position is directivity only, the other three are directivity and the speaker’s own response, each with a different phase.

Minimum phase is what the magnitude alone dictates — the phase any smooth passive filter of that shape would have. Measured phase adds what the export actually measured on top of it: the part a crossover adds in time and which never shows up in the magnitude at all. A Linkwitz–Riley pair sums to an all-pass, so no amount of equalising can find it — it has to be read from the phase column. Linear phase removes both: every frequency arrives at once. A passive crossover cannot do that — its phase follows from its magnitude, and there is no component to buy that separates them — but an FIR filter can, and plenty of people reading this have built one. What the switch gives you is the comparison without the build: the same speaker in the same room, with and without the phase its crossover adds, changed between two notes.

The measuring distance does not come with it. A constant delay has been fitted out of the phase between 200 Hz and 10 kHz, exactly as the average level has been taken out of the curve, so the list changes dispersion and not arrival time. Without that, where the microphone happened to sit would end up in the model.

The phase is read on axis, and off axis that is an approximation. Measured on the ideal two-way: sideways the difference stays under 0.07 ms out to 60°, but vertically it changes sign across the crossover — at 30° it is +0.04 ms below the woofer’s band and −0.10 ms above it, because the drivers are not on the same axis. The direct sound leaves along the axis and is therefore right; the floor and ceiling bounces leave at 32° and 48° in the default room and are an approximation.

So, plainly: the measured phase does not make the reflections more accurate. Because the sign flips, it improves them below the crossover and makes them slightly worse above it. How much depends on how far apart the drivers are: about 0.1 ms on that ideal two-way, or 7 % of the floor bounce’s own delay, but 0.3–0.42 ms on the classic three-way in this list, which is 20–29 % of it. The three-way is the speaker this whole effect exists for, so read the larger figure as the one that applies to a real loudspeaker. What this list is for is hearing what the crossover’s group delay does, and for the direct sound it is exact.

Linear phase is linear on the speaker’s axis, for the same reason and with the same numbers. It also costs 43 ms of latency — all four positions carry it, so switching between them does not move anything in time — and it rings before the transient. That pre-ringing is not a fault to be hidden: it is what linear-phase crossovers do, and hearing it is the reason the position exists. The directivity impulse responses stay minimum-phase in every position, so off axis there is still a phase the page has not linearised.

Where the limit is: comparing two different crossovers on the same speaker is a different question, and this list does not answer it on its own. Their difference shows as vertical lobing, and the lobing’s magnitude is already in the model — two exports of the same speaker already give different power responses and different reflections here. Getting the lobe’s phase right off axis would need a model with one source per driver. The two ideal two-ways in the list, LR2 and LR4, are that comparison as far as it goes today.

It is applied once per speaker, ahead of the EQ and the room, so the walls and the modes reflect it — the same reason the EQ is there. Bypass under the spectrum does not switch it off: that button compares your curve, not the speaker.

The curve is used as a shape, not a level: its average energy between 200 Hz and 10 kHz is removed, so switching it on changes tone and not loudness. All four positions share that curve exactly — only the phase differs, so none of them changes the tone relative to the others.

A position is grey when the speaker cannot do it, and the line underneath says which case it is. An ideal directivity pattern has no measured response at all. A measurement imported before this existed kept only the magnitude, so it gets minimum phase and no more – import its polar files again and the other two appear. The line also gives the number that decides whether any of this is worth hearing: the largest excess group delay between 100 Hz and 1 kHz, after the constant delay has been taken out — so it counts dispersion, not distance. The example two-ways read a fraction of a millisecond, which is small. The three-way reads ten times more, and that is what it is in the list for: it is the only one with a real 40 Hz high-pass under it.

One reading does not follow it: the critical distance estimate assumes a flat speaker, and the page says so on the line below it whenever the response or the EQ is shaping anything.

Speaker EQ — Bypass (e)

The curve is drawn over the spectrum, and you edit it there. Drag a dot sideways to move the band, up and down to change its gain, and scroll on it to change Q. The knobs below the graph are the same numbers, so either one follows the other. The two dots marked HP and LP only move sideways – a pass filter has no gain to drag.

The knobs turn by dragging up and down, not in a circle: two hundred pixels of travel is the whole range, and holding shift makes that five times finer. The wheel steps one per cent of the range over a knob, or one step of the control with shift – the fine end is where you ask for it rather than where it happens to land. The unit is in the label above each knob, so the reading below it is a number: 20.0k under Hz means 20 kHz.

Double-click any control on this page to put it back to its default – the knobs, and every slider from Off centre to Late tail. That is there because a slider is a poor way to hit an exact value: the four metres of Off centre are about two hundred and fifty pixels wide, so one step of two centimetres is under two pixels and dead centre is a one pixel target. Nothing snaps to zero, deliberately – snapping would make the values either side of it hard to pick, which is the place where fine adjustment matters most.

Note that the graph carries two vertical scales at once, deliberately. The orange spectrum is a level in dBFS, labelled on the left; the blue EQ curve is a gain, labelled on the right. They share the frequency axis and nothing else, so the EQ curve sitting above a peak does not mean it is louder than it – it means it is adding that many decibels there.

The EQ sits between the speaker and the room, so the walls, the floor and the room modes all reflect the curve you set – boost 80 Hz and you hear what the mode does with it, not just what your ear does with it. That is the reason for putting an equaliser on this page at all. Both speakers get the same curve; this is a listening test, not a channel-by-channel calibration.

Bypass is the control that matters, on the E key so you can keep your hands off the mouse while you compare. It is a real bypass: the signal goes round the filters rather than through flattened ones, so the two states are identical to within the arithmetic. Flat resets the curve and leaves the bypass where it is.

Four parametric bands, plus a high-pass and a low-pass whose Q is in decibels – that is Web Audio's unit for those two filter types, and Butterworth is −3.0 dB rather than the 0.707 you may expect. A band with 0 dB of gain is exactly transparent, so there is nothing to switch off.

The two pass filters have a slope of 12 or 24 dB/octave, named after the enclosure rather than the filter, because that is what the choice is for: a sealed box rolls off at 12 dB/octave and a reflex box at 24. Set the high-pass to your box's corner and you are listening to that alignment in this room. The low-pass offers the same two slopes with no such pairing – there is no "reflex" low-pass, so it is labelled by the number alone. Either slope is −3.0 dB at its own corner frequency; 24 dB/octave is two cascaded Butterworth sections, and the pair sums back to the same −3.0.

Both reach well past the bass. The high-pass runs to 10 kHz and the low-pass down to 100 Hz, so the pair is also a band-pass: you can hand the room a single octave and hear which reflections belong to it, or listen to a driver’s own passband on its own. Drive them past each other and nothing comes through — the page says so rather than leaving you to wonder.

Preamp is headroom, not tone. Boosts push the output towards clipping, so the button sets the preamp to the peak boost and the reading tells you where you are: Peak is the tallest point of the curve, net is what is left after the preamp. Sustained clipping sounds like a distorting driver, which is exactly the thing a test like this must not invent.

One number does not follow the EQ: the critical distance estimate knows only absorption and directivity, so it still assumes a flat speaker – the page says so on the line below it while the EQ is shaping anything. The measured direct vs room reading does follow it, because that one is measured downstream of the filters.

Ear height — At speaker height

Binaural rendering in the browser stores its head-related filters on a 15° grid of elevations, and picks the nearest one below the source. Crossing one of those steps changes the tone in a single jump – measured here at 6 dB around 6 kHz for a hundredth of a degree of movement – and the step lands exactly where a listener normally sits, level with the speaker. Worse, a reflection in a vertical wall keeps the source's height, so the direct sound and four reflections per speaker all cross it at the same millimetre.

With this box ticked your ears stay level with the speakers, the elevation stays at zero and nothing is ever crossed. Untick it and the page renders every path through two filters from neighbouring elevations and fades between them, which removes the step – at the cost of moving fourteen more sources every time you move. If the audio stutters, tick it back on.

Floor and ceiling reflections still cross their own steps when you walk towards the speakers, because their elevation depends on distance. That is the part the tick box does not fix.

Vertical + — + angle points up

Only appears for a measured speaker, because it is a question the export does not answer. A vertical polar set is usually asymmetric – a woofer below a tweeter does not radiate the same way up as down – so which sign means up decides whether the floor reflection or the ceiling one gets the off-axis response that belongs to it. Getting it backwards swaps the two. Up is the owner's convention here; the file itself only says +30°.

Directivity check — the measurement against what plays

The measurement is the thick line, what you hear is the dashed one. They are the same data: each path convolves the measured curve directly, so the dashed line is the measurement at the nearest cell – 6° across, 5° up – rather than a model of it. Where the two separate, that separation is the rounding.

The curves follow the Response switch, and that changes what they are. The polar data is normalised to the on-axis response, so on Directivity only 0 dB is the speaker’s own axis at every frequency and what you see is directivity alone. Switch the response on — any of the three phase models — and the on-axis curve is added back into At your seat, All angles and the power response, so they show the real response: 0° is no longer flat, and the dips you are looking at are the ones the chain actually produces. The line under the graph says which of the two you are reading.

The maps stay directivity in both cases. The error map is the measurement minus the cell that plays, so the same curve added to both would cancel out of it anyway; the polar map is a map of directivity and is named as one. Towards each surface follows the switch like the curves do: with the response on, the white direct line is the real thing and is not flat, because it is the sound that actually leaves towards you.

Watch for one thing on Directivity only: aim a speaker straight at the seat and its direct line goes flat, because that direction is the reference everything else is measured against. That is the zero level looking at itself rather than a perfect loudspeaker — switch the response on to see the real one.

In All angles the white pair is the power response – the energy average over the whole sphere, which is what a room ends up hearing. It is the same line in both planes, because the integral uses both. Measured data hides the measurement so you can read the playing curve on its own.

Towards each surface is a different question: not what reaches your ears, but what leaves the speaker in the direction of each wall, the floor and the ceiling. That is where placement is decided – a deep off-axis dip aimed at the wall beside you arrives as a reflection with a hole in it, and moving the speaker or changing the toe-in a few degrees moves the hole. Whether it matters is exactly the thing worth testing: mute that surface and see if you can hear the difference at all.

Two things it deliberately leaves out. It is the left speaker only – the right one is the same curve mirrored unless you have moved something, and the asymmetry is what At your seat is for. And it is what leaves the speaker, before the wall absorbs anything: absorption is smooth and would not put a dip anywhere, so including it would only make you wonder whether a hole belongs to the speaker or to the carpet. Each curve follows its own surface switch above, so the graph shows what is playing.

The error map is what the chain does not reproduce. What you hear at any angle is the measurement at the nearest cell, played through a short minimum-phase filter, so the map is the measurement at the exact angle minus the filter you actually get. Flat grey means it cost nothing there. Most of what you see is the rounding – but the map reads the real filter, so it also shows the two limits that filter has: its length, and a floor below which it cannot follow the measurement at all. The chequered patch in the rear top octave is that floor, not rounding, and it is where the largest number in the table comes from. It has its own key beside the colour scale rather than a stronger colour, because it is a different thing rather than more of the same one – there the chain is not approximating the measurement, it is not following it. The note under the map says which is which.

dB rms over every measured band, for the speaker you are listening to right now. It is near zero on axis, because the pattern barely moves there, and shows up wherever the pattern turns fastest – the deep off-axis nulls above 1 kHz, and the rear.

6° spacing was chosen by measurement: finer spacing buys hundredths of a decibel and doubles the number of impulse responses to keep in memory.

Reflection detail — Full · Shorter filters

This is the only load switch on the page, and it buys steady CPU rather than smoothness. Fourteen paths each convolve a directivity filter. Shorter filters cuts the twelve reflection paths to a quarter of their length and leaves the two direct paths untouched, because the direct sound is the one you localise on.

The load it saves is fixed; what it costs is not. Shortening the twelve reflection filters takes roughly half the convolver load off, whatever is playing. The accuracy you give up depends entirely on the loudspeaker, because a filter length is a length of time: on the bookshelves it is at most 1.6 dB in a single band and 0.17 dB rms, which is why you are unlikely to hear it there at all. On a bigger cabinet it is a different trade, and the list measures that rather than argues it. The same 6.5-inch two-way – same crossover, same driver spacing, nothing changed but the box – costs 0.05 dB rms in the bookshelf, 0.28 dB in the 180 mm tower and 0.75 dB in the 500 mm one, and the wide one gives up 6.0 dB in its worst band. Where it hurts moves down with the cabinet, which is the same fact read in frequency: that worst band sits at 2.1 kHz on the bookshelf, 750 Hz on the wide tower and 370 Hz on the 1 m sphere – and structure that low is long in time, so a quarter of the filter blunts it rather than following it. The sphere gives up 1.8 dB rms and 32 dB in its worst band for the same switch – ten times the bookshelf figure in rms and twenty times in the worst band. The ladder climbs with the cabinet: the spheres are still level with the bookshelves at 31 cm and past them at 38 cm, so that is where this switch starts to cost something. Judge it on the speaker you are actually listening to, not on the list.

What it does not fix: the stutter while you walk. That is not convolving, it is building – every time your angle rounds into a new cell, the page builds a new impulse response, and a filter is built at full length and shortened afterwards. Both settings therefore pay the same price for a step. The switch lowers the load between steps, not the cost of a step. If walking stutters, the levers are the room and the paths: mute surfaces you are not judging, turn Modes off, or move with the sliders instead of holding a key.

Where you can see it, and where you deliberately cannot. The polar map and the error map always read the direct path, so they do not move when you switch – the maps are about the measurement, not about the load. The curves under Towards each surface are drawn from the shortened filters, so those move: that graph is the honest window into what this costs, and a switch whose price is invisible everywhere would be a switch you could not audit.

Every figure above is measured across the whole speaker list, both planes and both sides of the horizontal where a speaker is not left–right symmetric, at five filter lengths — from 0.17 dB rms on the bookshelves to 1.8 on the 1 m sphere. The full length is kept because of the second.

Group delay — the crossover against the room

This is the phase of the Response switch drawn as time. The white curve is the group delay you hear on measured phase, the grey curve the one on minimum phase, and linear phase is the zero line itself — every frequency arriving together. The gap between white and grey is the whole of what the switch does, so if the two curves lie on top of each other, that A/B has nothing to find on this speaker.

The dashed blue lines are this room. Each one is how much later a reflection reaches you than the direct sound, from the same geometry the sound uses — move a wall or walk, and they move. They are there because milliseconds on their own do not say whether they are many: a crossover that delays the bass 2 ms is a different proposition in a room whose nearest reflection is 1.5 ms behind than in one where it is 15.

⚠️ They are not the same kind of millisecond, and the graph is not claiming they are. A reflection is a second arrival of the whole signal; group delay is a smear inside one arrival. Equal numbers do not mean equally audible — the comparison is there for the order of magnitude, which is the honest part. Whether either is audible is what the listening test is for, and the square-wave source is the one that makes the smear easiest to hear.

Two scales, because the answer changes which one you want. Room scale fits the nearest reflection into the picture, which is the comparison. Speaker scale drops it and fits the curves, which you need when the answer is “the speaker does almost nothing” and the curve would otherwise be a flat line along the bottom. The line under the graph names the reflections that are above the top of the scale either way.

On axis only — the stored phase is the on-axis phase, so there is no horizontal or vertical choice here and those buttons are hidden. Off axis the page uses the same phase, and how good an approximation that is measured is in the Response section.

The curve stops at the ends of the measurement, around 26 Hz and 19 kHz. Beyond those the model just holds the last value it was given, and a group delay read off that would describe the model rather than any loudspeaker.

A worked example, on the three-way with the 40 Hz Butterworth: the white curve peaks in the mid thirties of hertz and the grey one reaches about half as high — so about half of the bass delay is the magnitude’s own and would be there in any speaker rolling off that way, and the other half is the part only measured phase has. In the midrange the white curve crosses the floor-reflection line at around 110 Hz — above that, the room is doing more than the crossover.

Ears — Binaural (HRTF)

On, the browser's own head-related transfer function does the listening: interaural delay and ear filtering, which is what makes height and front-back audible on headphones. Off is plain amplitude panning – direction still follows where you face, but nothing goes through a head.

Turning it off is the quickest test of how much of what you are hearing is the generic HRTF rather than the room, because the floor and ceiling reflections become level and delay only. It is also the right setting if you are rendering an impulse response for loudspeakers rather than headphones: the same switch decides what the file below records, so the file always matches the binaural switch you listened through.

Expect a tone shift, not only a change of placement. Measured on this browser's HRTF against plain panning, averaged over both ears and referenced to 500 Hz: about +2 to +3 dB below 250 Hz and −4 to −5 dB from 2 to 5 kHz, with the top octave falling further. Binaural therefore sounds thicker and less airy, and that is the head and outer ear doing their job rather than a fault – though on headphones it arrives twice, since headphone target curves already contain an average outer ear. The colouration above roughly 6 kHz depends strongly on which direction the sound comes from, so it is information rather than tint, and no fixed compensation curve can remove it without making some directions wrong.

Critical distance — r_c n m · Q n · R n m² · seat n dB

The distance at which a real room with this absorption would have its reverberant field overtake the direct sound: r_c = √(Q·R/16π), with the room constant R = S·ā/(1−ā). It is theory on this room's absorption, not a reading from the mix you are hearing – the late tail here is a slider and the reflections are first order only, so the model's own ratio at your seat is a different number. That is the one the seat figure gives you.

The line prints R rather than the more familiar RT60 on purpose. RT60 is clamped to 0.15–2.5 s for the tail, so at extreme absorption the printed RT60 and this distance would no longer be the same room. R is exactly the number r_c is computed from, so you can check the arithmetic from the line itself. The familiar rule of thumb 0.057·√(V/RT60) is the same formula without the (1−ā) term, and it reads 10–20 % low in a normally furnished room and further off in a dead one.

Q is closed form where it is closed form: omni 1, cardioid and dipole exactly 3 – both have a power integral of 4π/3, so they differ in where the energy goes, not in how much of it goes off axis. A measured speaker is not one number, which is why the second line gives the extremes band by band. A value below 1 in the bass is not a fault: normalised polar data is slightly above 0 dB off axis down there, so the cabinet radiates a little more to the sides than on axis.

Measured direct vs room — n dB · n dB wetter

This one is read from the running audio, not calculated: two analysers sit in the graph, one on the direct sound and one on the room side (twelve reflections plus the late tail). The number is the ratio of their powers, averaged over about half a second, and it only exists while something is playing – hence press play when it is not.

It is taken above the crossover and without the mode bank, because below the crossover there is no separate direct sound to compare against, and before the HRTF, because the interaural delay would comb filter the sum and the meter would show a dip neither ear hears.

The second line compares it with the seat estimate on the critical distance row. A residual is expected and is not an error. The late tail has the average early energy subtracted from it – averaged over listening positions – so a seat that is closer to the speakers than that average still reads wet. Near them the early field really is stronger than a room mean. It is not an invitation to pull Late tail down. Reading drier means the opposite: the tail is below its calibrated 100 %, or surfaces are muted.

Three short words can appear instead of a number. silent: nothing is coming through above the crossover. direct only: the room side is muted, which is the anechoic reference. room only: the direct sound is muted.

Stereo triangle — n°

The angle at the listening position, which is a different angle from toe-in and the one people mean by the well known "60 degree triangle". A 60° triangle means each speaker sits 30° off your nose; toe-in is how far the speaker itself is turned towards you, and the two are independent – you can have an equilateral triangle with the speakers pointing straight down the room.

Distance to the speakers — Speakers n / n m · Δ n mm (n ms)

Equal distances mean you are on the centre line, and keeping that is the whole reason the move buttons exist. The delta is given in millimetres and in milliseconds because the second one is what you hear: a few centimetres is a fraction of a millisecond, and that is already an image shift rather than a level difference. It is also the reason the walk is slow — at 0.15 m/s the Doppler shift is 0.04 %, inaudible.

A / B — Now A · A and B differ in toe-in, separation

Store A and Store B put the current settings in two slots, and A / B switches between them. The line under the buttons says which one you are listening to, and what the two actually differ in – that list is the comparison you set up, so if it is empty the switch is silent because there is nothing to hear.

What you are listening in is shared; what you are listening to is not. The room and its materials, which of its surfaces are live, where the speakers stand, where you sit, which way you are facing, the movement locks and the walking speed are the setting rather than the comparison – so once A and B agree on one of them, adjusting it is written into both slots as well as into what you hear. Nothing is thrown away for a change you never meant as a comparison.

The speaker itself, its directivity, its own response and phase model, the equaliser, the binaural mode and muting one speaker of the pair stay in the slot you stored them in, even when both slots happen to agree. Those are the comparisons this page exists for, and a shared setting cannot be compared: choosing a different speaker would change both slots at once and leave nothing to switch between. So a change to one of them counts as an unsaved edit, the line says so, and storing it in the other slot is what turns it into a comparison. Only settings that genuinely differ – the ones the line names – revert when you press A / B.

The listening level, the source and the sine frequency never travel with a switch. A / B compares placement and room, and if the test signal changed with it you would be hearing two differences at once.

Paths — RT60 n s (LF n / HF n) · mean free path n m · n paths

RT60 is Sabine on the current room and materials, and it is what the synthetic tail decays with. The mean free path is the average distance between bounces, and it sets the pre-delay of that tail. The count is the direct sound plus the six first order reflections, per speaker – fourteen in a stereo pair – and mute a surface and it stays, because muting removes energy rather than the path.

Room modes — Modes n · Schroeder n Hz · Q n–n · axial n / n / n Hz

Below the Schroeder frequency the image source model is replaced by an analytic mode bank, because one bounce per surface cannot produce a mode – a mode is what is left after infinitely many. Every mode of an ideal rectangular room is a bandpass filter whose level is the mode's pressure at the speaker multiplied by its pressure at your ears, so a zero means you are sitting in a null, and that is the point. The Q values come from a low frequency RT60, which is longer than the broadband one.

Q also sets how high each mode peaks, not just how narrow it is. In the modal expansion a term peaks in proportion to Q/f ², so the bank weights every mode by that — and because Q itself grows with frequency, the two together leave a net 1/f tilt. That is the low end a room adds and a field does not, and it is why taking a speaker outdoors takes the bottom with it. Without the weighting each mode was a band limit and nothing more: the peaks were all the same height, which no room does.

One term of that expansion is not a bandpass and not a mode: the one where all three indices are zero. Every cosine equals one, so it does not vary with position — it is the room's pressure below its first mode, and it is what you hear when you switch the room off and the bottom octave changes. It is filtered to the lowest mode above and flattened at 10 Hz below, because a real room leaks and the pressure rise cannot continue to DC.

The axial trio is the lowest width, length and height mode: c/2 divided by each dimension. They are the ones you can walk into. If the line says the bank is capped, the bank ran out rather than the room – the modes above that frequency are missing from what you hear.

A symmetric pair fed the same signal cannot excite a left–right antisymmetric mode at all. The two speakers sit either side of the room's centre plane as long as Off centre is at zero, and while they do, every odd width index gets equal and opposite force from the two of them – including the lowest width mode. That is a little under half of every bank the room can have. It is what the physics says about a perfectly symmetric setup playing mono, and it is also why real rooms do not sound like this: real placement is never quite symmetric and real recordings are not mono. Toe-in does not change it and neither does where you sit – the mode is silent from every seat. Sliding the pair off centre does change it, and so does a recording whose two channels carry different bass.

Share a link

A shared link carries the settings, not audio: the room, the seat, the speaker choice, the surfaces, the EQ and your A and B snapshots. The reader's own browser renders it, so what they hear is what you heard, computed again on their machine.

Reading needs no account. Making a link needs a free account – an email address and nothing else – and that is what gives the link an owner: it stays yours, and deleting your account takes every link you made with it. A link nobody opens expires; one that keeps being opened does not. Taking a link down stops it working everywhere you have pasted it, including that forum thread.

Your own speaker measurement travels only if you tick the box that says so. Without it, a link that used one of this site's speakers still works – the reader's page fetches the same speaker by name.

Pressing the button twice does not make two links. A link is made from the settings, so as long as those have not changed you get the same address back – and if you edited the message meanwhile, that text is updated on it. Change the room and you get a new link, because it is a different room.

Pro names it. A Pro link has an address you choose (/p/your-own-address), a message of up to 280 characters shown on the page, your nickname and one link back to your own thread or site. The message is the useful part: “listen to the 45° ceiling reflection and tell me if you hear it” is a reason to open a link, a pile of slider values is not. The link back is rel="ugc nofollow" – it is there for people who click it, not for search engines, and this is said out loud so that nobody buys Pro for the wrong reason. Save text edits those afterwards without the address changing, because by then it is already in somebody's post. Pro links do not expire, and they keep working after a Pro period ends: a dead link would punish the reader, not the person who stopped paying.