ROOMPRINT

IN DEVELOPMENT — Features described here include the current design and planned extensions. Final specifications and release details will be confirmed before launch.

VUHWEX AUDIO / EFFECT VST / IN DEVELOPMENT

Build the room.
Put sound inside it.

ROOMPRINT is an editable architectural reverb: choose a space, change its proportions, replace its surfaces, move the source and listener, then shape the reflections and decay. Turn a tiled shower into a stone chamber—or design an impossible room of your own.

Geometry-driven early reflections. Frequency-dependent late decay.

• The planned hybrid architecture combines geometrically calculated early reflections with a frequency-dependent feedback delay network (FDN) for the late tail. Room edits are intended to change the acoustic configuration, not merely relabel a stock reverb’s Size knob.
• Rectangular rooms use image-source reflection paths with fractional delays and band-dependent surface losses. Path delay is distance divided by sound speed, approximately 343 m/s under default conditions.
• For an opaque boundary, pressure reflection magnitude derives from sqrt(1 − absorption), not directly from the energy coefficient. Scattering must be accounted for consistently rather than counted twice.
• A stable FDN mixing network and passive damping filters translate estimated band decays into a diffuse late field. The proposed feedback magnitude for delay d and amplitude T60 is 10^(−3d/T60).

Dimensions, proportions, and acoustic volume

• Independent Length, Width, and Height coexist with proportional Size scaling, Keep Proportions, and Lock World Positions. Shower Stall starts at 1 × 1 × 2.3 m. Internal SI units and a meters / feet display are planned.
• The panel exposes volume and area correction factors, unmodeled loss, and coupled-volume controls including volume ratio, aperture area, and secondary absorption. Acoustic volume is distinct from a building’s bounding envelope.
• Changing dimensions should alter reflection arrivals, room scale, density, and decay; adding absorption at fixed geometry should primarily alter losses and decay rather than move wall paths. These acoustic responses require measurement to verify.
• The design calls for bounded dimensions, volume, memory, and source/listener wall separation. Resizing may preserve normalized positions or clamp locked world coordinates.

Six surfaces, each with its own material

• North, South, East, West, Floor, and Ceiling are independently selectable. Link Walls, All Walls, Apply to Selected, Reset Surface, and Custom material controls.
• Each surface combines a base material with fractional overlays. Coverage replaces the relevant area rather than adding a second full wall; overlay totals must not exceed 100%.
• The design uses explicit octave bands at 125, 250, 500, 1000, 2000, and 4000 Hz, with frequency-dependent absorption, separate scattering, construction notes, provenance, and confidence labels.
• Shower Stall uses glazed ceramic tile. Small Country Church uses plaster with a 12% glass overlay; the rotunda uses smooth stone with 20% rough stone. These are modeling recipes, not measured finish maps.

A material palette with acoustic differences

• Material families: ceramic tile, glass, concrete, brick, smooth marble / granite / stone, rough stone, plaster on masonry, drywall on framing, thin timber panels, solid timber, plywood, corrugated steel, carpet on concrete, carpet with underlay, heavy / thin curtains, upholstered seating, exposed absorber panels, book shelving, and custom materials.
• Mounting and construction matter: timber paneling is not acoustically identical to solid timber, and carpet over concrete differs from carpet with underlay. Bookshelves need scattering as well as absorption.
• Published-reference, engineering-estimate, and creative data must remain distinguishable. No coefficients are presented here as measured venue data or independently validated material measurements.

Furnishings and people are part of the model

• Carpet, curtains, wall panels, upholstery, seating capacity, and occupancy. Coverage has geometric meaning; furniture and people require equivalent absorption areas rather than arbitrary percentages of reverb strength.
• Zero occupancy must retain the empty seats. Occupied-seat absorption should replace / adjust empty-seat values rather than add a complete second seat. A room without seating shows “No audience seating.”
• Open doors, corridor ends, and an oculus require energy loss. Water is a boundary condition rather than a generic soft absorber; underwater acoustics are outside the design scope.

Estimated decay, with assumptions exposed

• Sabine T60 = 0.161 V / A is the baseline diffuse-field estimate. Eyring uses mean absorption where appropriate: −0.161 V / [S ln(1 − mean absorption)]. Objects, air, and openings must be added without double-counting.
• Band-dependent losses should produce meaningfully different low- and high-frequency decay. Small highly absorbing rooms and irregular spaces exceed the reliability of simple statistical estimates.
• The development design proposes a 60-second natural-mode decay cap and a bounded creative extension up to 120 seconds if stable. These are design limits, not verified current ranges.
• A statistical late tail is not full low-frequency wave simulation. Echo duration, EDT, T20, T30, and RT60 are different measures; a famous echo figure does not establish broadband decay.

Shapes with honest approximation labels

• The interface shows an exact-shoebox image-source label for the rectangular baseline, and “Facet approximation (2nd-order planes)” for the barrel vault and domed rotunda. Those labels concern the geometry method, not an exact reproduction of real acoustics.
• The planned architecture includes shoeboxes, vaults, domes / circles, tiered auditoriums, corridors, and coupled spaces. Shape-specific rendering should only be claimed where implemented; simplified-envelope approaches need an explicit badge.
• A dome’s volume differs from its bounding box, arena seating reduces open air volume, and a cathedral envelope is not one uniformly filled shoebox. Room metadata and correction factors are intended to preserve those distinctions.

Place the source and listener

• Top-down and side views show source/listener positions, height, and reflection-path lines. Stereo source spacing is explicit in the panel. The design calls for draggable positions and bounded distance attenuation.
• Stereo input should retain channel identity through two nearby source positions or another documented excitation strategy. Directional panning is not the same as HRTF / binaural rendering.
• Insert mode keeps dry audio immediate and excludes a duplicate direct path from wet reflections. Relative-to-direct timing, additional artistic Pre-delay, and optional physical propagation must remain distinct from processing latency.

Natural rooms and creative architecture

• Natural mode derives decay from geometry and materials; Creative mode adds exaggerated size, modulation, low / high decay multipliers, and artistic overrides. The interface exposes Size ×, Low Decay ×, High Decay ×, and A/B Morph.
• Continuous morphing is only physically meaningful for compatible configurations. Incompatible geometry or topology requires safe engine crossfades rather than a promise of smoothly interpolating any two buildings.
• Wet/Dry, Output, Pre-delay, Early, Late, Width, Distance, Diffusion, Decay ×, low / high cuts, Freeze, Quality, Mix Mode, and Timing. Separate early and late levels let you balance spatial definition against the diffuse tail.

The Stone Shower workflow

• Start with Shower Stall. Select only the walls and replace ceramic tile with rough stone, leaving the glass door and ceiling unchanged. Resize to 4 × 2 × 3 m, then save the result as Stone Shower.
• The defining requirement is an audible update to reflection loss, scattering, timing, volume, and decay—not just a changed picture. Full model state should return when the project reopens.
• Save As, Revert, Undo / Redo, A/B, A → B, Reset Room, and Reset Surface support editable working copies while preserving factory originals. Modified named-space presets should remain labeled as modified approximations.

Fifty editable starting spaces — factory-bank target

The development target is exactly 50 starting models. Named venues are architectural approximations, not measured impulse responses, surveyed digital twins, endorsements, or independently verified acoustic matches. Published reference dimensions, assumed dimensions, and original generic spaces are separate data categories.

• 1. Sistine Chapel
• 2. Sydney Opera House — Concert Hall
• 3. Madison Square Garden — Empty
• 4. Abbey Road — Studio One
• 5. Abbey Road — Studio Two
• 6. Vienna Musikverein — Golden Hall
• 7. Boston Symphony Hall
• 8. Amsterdam Concertgebouw — Main Hall
• 9. Versailles — Hall of Mirrors
• 10. Pantheon, Rome
• 11. Great Pyramid — King’s Chamber
• 12. Notre-Dame de Paris
• 13. Grand Central — Main Concourse
• 14. Hamilton Mausoleum — Chapel
• 15. Inchindown — Empty Oil Tank
• 16. Shower Stall
• 17. Tiled Bathroom
• 18. Public Restroom
• 19. Walk-In Closet
• 20. Furnished Bedroom
• 21. Empty Apartment
• 22. Living Room
• 23. Kitchen
• 24. Wooden Sauna
• 25. Elevator Cabin
• 26. Vocal Booth
• 27. Garage Rehearsal
• 28. Wood-Paneled Live Room
• 29. School Classroom
• 30. School Gymnasium
• 31. Small Country Church
• 32. Brick Basement Club
• 33. Cinema Auditorium
• 34. Old Wooden Theater
• 35. Hotel Ballroom
• 36. Library Reading Room
• 37. Museum Gallery
• 38. Shipping Container
• 39. Concrete Stairwell
• 40. Parking Garage
• 41. Brick Warehouse
• 42. Aircraft Hangar
• 43. Subway Platform
• 44. Pedestrian Tunnel
• 45. Storm Drain
• 46. Underground Cistern
• 47. Indoor Swimming Pool
• 48. Limestone Cavern
• 49. Stone Crypt
• 50. Abandoned Asylum Corridor

Randomization for rooms, not just knobs

• Separate Room, Materials, Furnishings, Positions, and Random All actions are, with intensity, seed, Repeat Seed, and New Seed.
• Locks cover dimensions, selected surfaces, occupancy, positions, creative controls, and Mix / Out. Random operations should be one undo transaction and preserve gain, bypass, and safety limits by default.
• Natural randomization should create sensible combinations. Creative Wild can explore strange proportions and materials while remaining bounded.

Acoustic state and engine transitions

• Project recall is designed to retain complete geometry, materials and custom coefficients, coverages, furnishings, capacity, positions, creative controls, seed, quality, provenance, approximation metadata, and preset identity in host state.
• Material and geometry changes should build off the audio thread, then crossfade engines with a defined tail policy. Smoothing a slider alone is insufficient when delay networks change.
• Automation, mono-to-stereo / stereo-to-stereo, offline rendering, Freeze, tail reporting, sample-rate changes, and deterministic updates are development requirements. CPU, latency, stability, and actual decay measurements are not claimed as tested.

Release information

• ROOMPRINT is designed as a fully offline desktop effect. Development prioritizes VST3 and AU, with Reason an important validation host. Final supported formats and systems will be confirmed before release.
• Audio demonstrations, model-bank completion, material provenance, host results, pricing, and downloads are still to be confirmed. This is a musical architectural approximation, not a certified room-acoustics prediction tool.