Help

Using the site

Finding enclosures

Catalog: All enclosures as cards. Filter by category (sub · kick · mid · top) to narrow the list. Each card links to the full design page with spec table, curves, plans, and driver info.
Find by constraints: Filter on hard numbers: F3, max SPL, internal volume, driver size and count, topology, and tags. Sort by any metric including efficiency densities (dB/L, dB/kg) to find the best design for a given constraint.
Design space: Scatter plot of any two metrics across the full catalog, with Pareto frontier highlighting. Good for spotting which designs are genuinely best on two axes at once.

Comparing builds

Compare enclosures: Overlay response curves (SPL, phase, impedance…) from multiple enclosures. Measurements are prioritised over sims when both exist. Optional peak normalisation lets you compare shapes regardless of absolute level.
Compare drivers: T/S parameters side by side plus a radar chart, up to 4 drivers. Links are shareable via ?ids= in the URL.

Drivers & horns

Driver catalog: Cone · compression · horn tabs with filter and sort. Each driver page shows full specs, derived hints (EBP, sealed Vb), and every enclosure in the catalog that uses it.
Substitute candidates: Every driver and enclosure page lists drop-in alternatives (same size for cones, same exit for compression drivers), ranked by parametric similarity and tiered close · usable · risky. Flags warn when impedance, power handling, or Xmax regress.
Horn ↔ compression driver matching: Horn pages list compatible compression drivers and vice-versa, matched on throat exit diameter. No stored link, derived live from the specs.

Planning a rig

Stack builder: Assemble a full rig from catalog boxes with per-line quantities. The state is encoded in the URL so builds are shareable. The summary covers cabinet count, weight, transport volume, plywood sheets, bandwidth, and power-summed max SPL with per-category array gain.
Amplifier matching: For each line, every series/parallel wiring of the cabs is shown with the resulting load rated ok · caution (below 4 Ω) · danger (below 2 Ω) · inefficient (above 16 Ω), plus recommended amp channel sizing.
Crossover suggestions: One crossover point per adjacent band pair, found from the boxes' own SPL curves rather than their stated specs: the frequency where the two bands' composite curves actually cross at equal SPL, searched on a 1/12-octave log grid and clamped up to any compression driver protection floor. The search may reach one octave past a curve's own last data point using that edge's slope. Where no true crossing exists, the closest approach is used and flagged as approximated; beyond 20 dB apart a gap is reported instead. Corners are starting points for a real alignment, not DSP presets: the model is magnitude-only, with no phase term.
Predicted system response: Per-band curves plus a composite. The LR4 toggle applies ideal Linkwitz-Riley 4th-order filters at the suggested crossover points and shows the coherent sum.

Data & provenance

Measured vs sim badges: Every enclosure is marked measured or sim. Measurements come from REW or Klippel, sims from Hornresp or AkAbak. The two are never blended on the same curve.
Where driver specs come from: Driver parameters are transcribed, and their sources are uneven: most entries link to third-party parameter aggregators rather than manufacturer datasheets, and some carry no source link at all. Treat every figure as manufacturer-nominal unless you check the link. Published Thiele-Small parameters are typically single-sample and vary 10-20% unit to unit, so a value shown to three decimals is a transcription, not a tolerance.
What the predictions assume: Predictions are arithmetic over stored curves, not acoustic simulations. The main assumptions: array gain is unbounded and frequency-independent, distance is point-source inverse-square with no air absorption, the crossover model is magnitude-only with no phase term (so it cannot show lobing or crossover cancellation), and no prediction derates for power or port compression. Time alignment, directivity, rigging and limiter settings are out of scope entirely. Full detail, with every constant quoted from source, is in docs/methodology.md in the repository.
Exports: Compare and stack views export CSV or JSON. The stack also has a print-friendly sheet. Raw data is available via JSON APIs: /api/manifest.json, /api/drivers.json, /api/horns.json, /api/curves/<slug>.json.

Glossary

The vocabulary behind the specs: enough to read a datasheet and size a rig.

Driver parameters

Fs (Hz): Free-air resonance of the driver: the frequency where the moving mass and suspension naturally resonate. A driver makes little useful output below Fs in most boxes, a sub driver wants Fs under ~40 Hz.
Qts / Qes / Qms: How damped the driver's resonance is: Qes is electrical damping (motor strength), Qms mechanical (suspension losses), Qts the combination. Qts is the single most important matching parameter when swapping drivers, a box designed around Qts 0.35 behaves differently with a 0.5 driver.
Vas (L): Equivalent compliance volume: the volume of air with the same springiness as the driver's suspension. Together with Qts and Fs it sets the box volume and tuning a driver wants.
Xmax (mm): How far the cone can travel (one way) while the voice coil stays in the magnetic gap. Beyond it output stops rising and distortion climbs fast. Xmax bounds a sub's maximum clean SPL at low frequencies.
Sd (cm²): Effective radiating area of the cone. SPL capability at low frequencies is set by Sd × Xmax (displaced volume): two 15″ drivers move roughly as much air as one 21″.
Sensitivity (dB): SPL at 1 m for 1 W (2.83 V into 8 Ω) input, measured free-field. Applies to cone drivers. Compression drivers use a separate figure measured on a reference horn (sensitivityHornDb), which is not comparable to a cone's free-field number. 3 dB more sensitivity = same loudness for half the amp power.
CD sensitivity (dB): SPL of a compression driver at 1 m, measured on the manufacturer's reference horn, not free-field. Typical range: 105–115 dB, versus 95–100 dB for a cone driver. The two figures measure different things (horn-loaded vs free-field) and cannot be compared directly.
EBP: Efficiency Bandwidth Product: Fs ÷ Qes. A rule of thumb for what loading suits a driver: under 50 favours sealed, over 100 favours ported/horn, in between works either way.
Min crossover (Hz): The lowest frequency a compression driver may be crossed at (with the stated slope). Crossing lower sends excursion the diaphragm can't take, this is how tweeters and CDs die at parties. When in doubt, cross higher.

Horn parameters

Throat exit (in): The diameter of the horn's mouth at the driver side, in inches. A compression driver mates a horn when their exit diameters match (1", 1.4", and 2" are the common sizes). Matching is geometric only: verify bolt pattern and crossover range before assuming full compatibility.
Coverage angle (H×V): Horizontal and vertical dispersion of a horn at its nominal frequency, measured at the −6 dB points. A 90°×60° horn is wider than tall: useful in wide rooms or horizontal arrays. Coverage narrows with rising frequency.
Flare cutoff (Hz): The frequency below which a horn stops providing useful loading and directivity control. Output rolls off steeply below the cutoff. Sets the practical lower limit of the crossover range a horn can handle.
Flare profile: The mathematical curve of the horn's expanding walls. Conical: straight-walled, widest pattern, simplest to build. Exponential: classical flare, lower cutoff for a given mouth size. Tractrix: gentler impedance transition, common in PA horns. Oblate spheroidal: optimised for constant directivity across frequency. Radial: sector-shaped walls, good pattern control in one plane. Multicell: array of smaller horns, wide coverage, vintage PA look. Diffraction: shallow waveguide plate, very wide coverage from minimal depth. Hypex: hyperbolic-exponential, a family sitting between conical and exponential, the flare constant trades loading against midband ripple. Le Cléac'h: expands so the wavefront stays spherical along the axis, minimising higher-order modes, favoured for horn hi-fi and full-range builds.

Box & system figures

F3 / F6 (Hz): The frequency where the box's response has fallen 3 dB (or 6 dB) below its average level: the practical lower limit of the cabinet. Lower F3 = deeper bass, usually paid for in box size or sensitivity.
Fb (Hz): Port/vent tuning frequency of a ported or horn box. Below Fb the driver unloads, excursion rises fast and Xmax limits output. A subsonic highpass just below Fb protects the drivers.
Max SPL (dB): Loudest output at 1 m before a limit is hit. Two independent ceilings: excursion-limited below Fb (cone travel runs out) and thermal-limited above (voice coil heat). The headline figure is whichever applies.
Crest factor (dB): Gap between a signal's peaks and its average (RMS) level. Compressed dance music sits around 6 dB; a sine test tone is 3 dB. Your amp must pass the peaks cleanly while the cabs see the average power.
Array gain / coupling: Stacking cabinets adds output: subs close together couple coherently (+6 dB per doubling, four subs play 12 dB louder than one), while mid/top cabs sum as power (+3 dB per doubling). This is why sub walls work.
Distance loss: A point source loses 6 dB per doubling of distance: 130 dB at 1 m is ~104 dB at 20 m. Big stacks and walls behave better than this in their coupled range, but it's the safe planning assumption.
Crossover / LR4: The filter splitting the signal between sub, kick, mid and top. LR4 (Linkwitz-Riley 4th order, 24 dB/oct) is the live-sound default: each side is −6 dB at the crossover point and they sum flat when time-aligned.
Measured vs sim: Curves here are either simulated (Hornresp/AkAbak prediction) or measured on a real build (REW, Klippel). Sims are honest starting points; measurements are reality. The catalog never blends the two.

Power & electrical

AES / program / peak power: AES: continuous power the driver survives (2 h noise test). Program: ~2× AES, the music rating. Peak: ~4× AES, instantaneous. Match amps on continuous power into the real load, not the peak number on the sticker.
Impedance (Ω): The electrical load the cabinet presents. Nominal (4/8/16 Ω) is an average, real impedance dips lower with frequency. Halving the load roughly doubles amp power drawn, until the amp runs out of current.
Series / parallel wiring: Parallel divides impedance (two 8 Ω cabs = 4 Ω), series adds it (= 16 Ω). Most amps are happy at 4 Ω, good ones at 2 Ω; below that you cook the amp. Underpowering with a clipping amp kills more drivers than honest overpowering.

Topologies

Sealed: A closed box: tight, predictable, compact, forgiving of driver substitutions, but the least efficient way to make bass. Rare in freeparty rigs except for infra duty.
Bass reflex (ported): A box with a tuned port that boosts output around Fb. Best extension per litre, easy to build and sim, the default for kicks and tops, common for subs. Watch port velocity (chuffing) at high level.
Bandpass: The driver fires into a chamber and all output leaves through ports: high output over a narrow band, with the box acting as a filter. Loud but one-note if pushed beyond its band; harder to verify without measurements.
Tapped horn: A folded horn where both sides of the cone feed the same path. The best SPL per litre and per watt in its passband, but very driver-specific: a tapped horn designed for one driver rarely works with another. Sim before substituting anything.
Front-loaded horn: A horn in front of the cone (often with a sealed or vented rear chamber): big sensitivity gains and controlled directivity. Needs a large mouth to load low, which is why bass horns are huge or stacked in multiples.
Folded horn: A horn path folded to fit a transportable box (W-bins and friends). Classic punchy festival kick/bass sound; the fold compromises the top of its range slightly.
Scoop: A rear-loaded horn open at the front: the freeparty classic. Cheap to build, forgiving of driver choice and build tolerances, very loud per watt, at the cost of muddier, less controlled bass than a tapped horn or reflex. Walls of scoops are a sound system rite of passage.
Transmission line: A long damped duct behind the driver, ~quarter-wavelength at tuning: deep, clean extension from a moderate box. More a hi-fi/studio choice than a rig workhorse, big for the SPL they make.
Cardioid: A single cabinet built to radiate forward and cancel rearward, either with opposed drivers or a rear driver fed delayed and inverted. Buys roughly 10-15 dB of front-to-back rejection over the low end, which keeps stage and neighbours out of the sub field, at the cost of some forward output and extra amplifier channels. Distinct from end-fire, which achieves a similar pattern by spacing several ordinary cabinets and is a property of the arrangement, not of the box: Boxdex records topology per cabinet, so end-fire has no taxonomy value.

Availability

Free download: Plans are published here or linked freely: download and build.
Paid: The plans are sold. Buy them from the designer or vendor through the linked source or a listed contact.
Contact for plans: Not a public download: reach the designer through a listed contact channel to obtain the plans.
Commission only: Built to order by the designer or shop rather than distributed as plans. Contact them to commission a build.
Out of print: The design is no longer distributed. Listed for reference, a contact may still help you track down the plans.

Going further

Boxdex crossover suggestions and predicted responses are derived estimates, not substitutes for real acoustic simulation and measurement. These resources cover the tools and theory to go further.

Simulation

Akabak: Free Windows electroacoustic simulation environment (R&D Team / Joerg Panzer). Models horns, waveguides, vented and bandpass boxes. The standard tool for validating crossover points and horn geometries before a build.
Hornresp: Free lightweight horn and vented/bandpass box simulator (David McBean). Fast T-S parameter workflow, good for early-stage horn geometry exploration and quick SPL/impedance curves.
VituixCAD: Free Windows engineering tool for multi-way loudspeakers (Kimmo Saunisto). Merges driver measurements into full-system responses, designs passive and active crossovers, and models directivity. Often used alongside REW or ARTA measurement data.
WinISD: Free Windows box simulator for sealed, vented, and bandpass alignments (Linearteam). Narrower scope than Hornresp, T-S parameters straight to box tuning curves, a common first stop for subwoofer alignment.
Unibox: Free spreadsheet-based box modeling tool (Kristian Ougaard) built on Microsoft Excel or OpenOffice. Same T-S math as the dedicated simulators, useful when you want to see and edit the equations directly.

Measurement

REW · Room EQ Wizard: Free cross-platform measurement suite. Captures SPL, impulse response, and transfer functions via a calibrated mic. The standard for validating DSP alignment and crossover settings in the field.
Open Sound Meter: Free, open-source measurement app (desktop + iOS/Android). Spectrum, SPL, magnitude, phase, impulse, coherence. Good for quick on-site checks without a full laptop rig.
ARTA: Free measurement suite (impulse response, frequency response, distortion, impedance) via a standard sound card. Bundles STEPS for stepped-sine distortion analysis and LIMP for T-S parameter extraction, a European standard alongside REW.
Smaart: Commercial dual-channel real-time analysis platform (Rational Acoustics), the touring and install industry standard for system alignment, delay setting, and transfer-function comparison against a reference. Not free, but the tool most professional system techs actually carry.

Reading

Sound Systems: Design and Optimization · Bob McCarthy: The standard professional audio reference for multi-way system alignment, crossover strategy, delay, and equalisation. Written by the director of system optimisation at Meyer Sound. Goes well beyond what any calculator can tell you.
Loudspeaker Design Cookbook · Vance Dickason: The other standard reference alongside McCarthy, in print since 1978. Covers closed and vented box theory, driver and crossover design, and cabinet construction in depth, the book most T-S parameters and box-alignment formulas ultimately trace back to.
Introduction to Loudspeaker Design · John L. Murphy: Shorter and more approachable than the two references above. Strong on diffraction, cavity effects, and horn/waveguide loading specifically, useful once the basic T-S math is familiar and the questions turn to why a horn measures the way it does.

Communities

diyAudio: Large English-speaking DIY audio forum with dedicated subforums for horn speakers, subwoofers, simulation, and driver measurements. Active Hornresp and Akabak discussion threads.
HornPlans: French-speaking forum specialised in horn loudspeaker design and DIY builds. Discussion threads on Akabak, Hornresp, crossover alignment, and horn geometry.
Speakerplans.com: Forum built around free scoop, tapped-horn, and folded-horn bass bin plans for outdoor PA and sound system rigs, the lineage most horn-loaded subs on this site descend from. Less active than it once was, but the archive of threads runs deep.
The LAB, ProSoundWeb: Live sound and touring professionals discussing rig design, system alignment, and gear choices. Less build-focused than diyAudio, more useful for how a system actually gets deployed, tuned, and run at a gig.
Audiofanzine, forum Sonorisation: Active French-language sound reinforcement forum. DIY subwoofer builds, gear comparisons, and PA troubleshooting alongside HornPlans' horn-specific focus.