Tapped, front-loaded, back-loaded, offset-driver, stubbed, compound, multiple-entry and Paraflex horns. See how fast the air moves inside the path, add stuffing span by span, and fold the result into a cabinet.
Tapped Horn · B&C 21SW152Half Space · 1.0 m · 100 W · dashed outside 25–120 Hz
Path length
2.50m
−3 dB band
39–113Hz
Peak at 100 W
123dB
Excursion / Xmax, 25–120 Hz
6.0 / 15.0mm
Every graph, schematic and fold on this page is output from the simulator’s own solver, computed when the page was built.
Horn types
8 ways to load a driver with a horn.
Choose where each face of the diaphragm meets the path, and the simulator builds the acoustic network for it. Every type can share a workspace with your boxes, so you can compare them directly.
Front-loaded horn
The driver’s front face feeds the main horn and its mouth; the rear face loads the rear chamber.
Drawn with a sealed rear chamber. Choose Port to vent it like a bass-reflex box; the port radiates beside the mouth.
Drawn by the simulator from an example Paraflex horn.
Build the path
Shape the horn span by span.
A horn path is a chain of spans. Each span runs between two stations, with a length, an area at each end and a flare, and neighbouring spans share a station, so the path stays continuous while you edit. Change it in the Horn Editor tile, as a drawing or a table, and the graphs follow.
Flare profiles
Conical, exponential or parabolic, chosen per span
Spans per path
1 to 8
Span length
2 cm to 6 m
Cross-section
5 cm² to 2 m² at every station
Taps and entries
Driver, tap, stub and entry positions along the path
Chambers
Rear and throat chambers where the horn type has them
Multiple-entry
Extra drivers on their own channels, with their own power, polarity, delay and chambers
Paraflex
PH1 to PH4, with four main and three auxiliary spans
Diagnostics
See what the air is doing inside the horn.
Particle velocity, pressure and acoustic impedance read the path at any station you pick, and a fourth graph shows the output of each mouth or directly radiating driver. Find the narrow section where air moves fastest before you cut a single board.
Particle velocity, peakm/s
Constriction
Peak 5.8 m/s at 38 Hz0.95 m along the path · 600 cm²
Mouth
Peak 3.3 m/s at 42 Hz2.50 m along the path · 2,400 cm²
Pressure, RMSPa
Closed end
Peak 1,485 Pa at 40 Hz0.00 m along the path · 400 cm²
Constriction
Peak 1,131 Pa at 41 Hz0.95 m along the path · 600 cm²
Tapped Horn example at 100 W in half space, 20–300 Hz. The same data drives the horn particle velocity and horn pressure tiles in the app.
Stuffing
Damp a horn one span at a time.
Stuff any span: set the material’s airflow resistivity, from 100 to 100,000 Pa·s/m², how much of the span it fills, and which end it fills from. Absorption follows the Miki model for porous materials.
Largest change, 25–120 Hz
1.0dB
Highest peak above 200 Hz
119 → 116dB
Native comparisons check exported stuffing placement, length and volume. Supported layouts export with setup notes, but Miki and Hornresp use different material models, so their responses can differ. Shared multiple-entry horn stuffing and Paraflex auxiliary-path stuffing remain unsupported for export.
Tapped Horn, SPLHalf Space · 1.0 m · 100 W
EmptyH1 50 % filled from the closed end, 10,000 Pa·s/m²Working band
Side sectionSerpentine · 5 runs · 530 mm wide
Path length
2,500 mmtarget 2,500 mm · +0.0 %
Modelled air volume
363 Ltarget 409 L · −11.4 %
The tapped horn example, folded with Fold into cabinet when this page was built.
Fold into cabinet
Fold the path into the box you can build.
Pick a pattern and lock the dimension your room decides, or draw the passage cell by cell, and the Fold tile packs the horn into the box. The horn the fold builds is drawn dashed on every graph beside the one you designed, with a 3D cutaway and a cut list for the build.
Works with front-loaded, back-loaded, tapped and offset-driver horns.
Shows what the fold does to the sound, and lets the horn follow the fold with one click.
Bends are not modelled as losses, and it does not produce fabrication drawings.
The horn model is regression-tested against native Hornresp captures, and horns that fit Hornresp’s layouts export as native Hornresp records, so you can check a design in both.
Supported active filters and stuffing export with setup notes. The export identifies omitted settings, such as other EQ and listening distance, and when several drivers are represented by one equivalent driver.
Unsupported geometry, too many spans, multiple-entry driver delay and unsupported stuffing layouts are refused with a reason. Stuffing placement is checked, but the material models can give different responses.
These are comparisons between simulators, not measurements of built cabinets.
Native Hornresp comparisons
122
122 comparison cases across 8 horn families: 114 complete curve sets and 8 sampled diagnostic checks. 109 of 109 agreement checks meet their stated tolerances. The other 13 cases track stuffing-model differences, without claiming agreement.
Front-loaded26
Tapped horn22
Offset-driver16
Back-loaded13
Stubbed9
Compound13
Paraflex13
Multiple-entry horn10
The comparisons cover 20 Hz–5 kHz, with stuffing curves through 2 kHz and sampled diagnostics at three frequencies. Directivity, thermal power compression and bend losses remain outside this model.
Front-loaded, back-loaded, tapped, offset-driver, stubbed, compound and multiple-entry horns, and Paraflex horns in all four variants, PH1 to PH4. Front-loaded, offset-driver and stubbed horns can have a sealed or ported rear chamber. Max SPL and headroom work for every type, and each can sit alongside sealed, vented, passive-radiator and bandpass boxes in the same workspace for comparison.
How accurate is it?
The regression suite compares 122 native Hornresp cases across 8 horn families, covering response, electrical outputs, pressure, particle velocity, filters and maximum output. 109 of 109 agreement checks meet their stated tolerances; 13 further cases track stuffing-model differences without claiming agreement. Full-curve comparisons cover 20 Hz–5 kHz, with stuffing through 2 kHz; sampled diagnostics use three frequencies. These are comparisons between simulators, not measurements of built cabinets.
Can I export a design to Hornresp?
Yes, for horns that fit Hornresp’s layouts; a multiple-entry horn exports as a set of records. Supported active filters and stuffing transfer with setup notes. The export identifies omitted settings and equivalent-driver representations. Unsupported geometry, too many spans, multiple-entry driver delay, shared multiple-entry horn stuffing and Paraflex auxiliary-path stuffing are refused with a reason. Stuffing placement is checked, but the two material models can produce different responses.
Is horn simulation free?
Yes, like the rest of the simulator. You can build, simulate and fold horns without an account.
How are horns modelled?
Each horn path is a one-dimensional wave model built from conical, exponential or parabolic spans, coupled to the driver’s Thiele/Small model. The model is linear. It does not include bend losses, directivity or thermal power compression.
Can it fold a horn into a cabinet?
Yes, for front-loaded, back-loaded, tapped and offset-driver horns. The Fold tile sits beside your graphs: pick a pattern and lock the dimension your room decides, or draw the passage cell by cell, and drag the driver or a board to tweak it. The horn the fold builds is drawn dashed on every graph, and a 3D cutaway, cut list and Hornresp values are one click away. A saved fold also exports to Boundary Lab to solve it in 3D. Bends are not modelled as losses, and it does not produce fabrication drawings.
Is horn simulation finished?
Not yet. Horns are in alpha: the model is regression-tested against Hornresp, but the Horn Editor and the Fold tile are still changing.
Can an AI assistant work on horns?
Yes. Connected over MCP, Claude, ChatGPT or another client can read, edit and simulate every horn type in your workspaces with the same solver.