Energy balance: rigid-backed constructions are solved as one-port networks (T = 0). Freestanding constructions are solved as two-port networks and report reflection, transmission, and intrinsic absorption from A = 1 − R − T. Temperature, pressure and humidity update the ambient air density, sound speed and viscosity used by every layer.
Porous layers: generalized Miki, five-parameter JCA, or six-parameter JCAL equivalent-fluid propagation is driven by airflow resistivity/transport properties. Generalized Miki now uses porosity and tortuosity explicitly; optional Bies–Hansen fiberglass estimation can couple bulk density + fiber diameter back into flow resistivity. Low-density flexible batts can use Panneton’s limp-frame effective-density correction; rigid boards default to rigid-frame behavior.
Limp membranes: MLV is treated as a surface mass with explicit surface resistance, tension and a frequency-dependent viscoelastic mass-loss term. Preset MLV can derive surface mass from bulk density × sheet thickness, so thickness changes can alter the acoustic inertance when that coupling is enabled. A nominally lossless membrane still changes impedance and creates coupled cavity/porous resonances even when its own direct dissipation is small.
Ports: small microperforations use a Maa-style viscothermal impedance approximation. Larger holes use a macro-perforation/Helmholtz neck model with open ratio, inertance, viscous loss, and end correction.
Layer interaction: the energy inspector reports direct dissipated power, Δα if each layer is removed, and a +5% parameter sensitivity audit. Impedance discontinuities are not assigned fictional loss: they change reflection, standing-wave phase and particle velocity so neighboring lossy layers dissipate differently. Air cavities additionally support an optional thermoviscous wall-loss model; for full-size panel cavities this direct loss is normally much smaller than the cavity’s impedance-transforming effect.
Finite panels: intrinsic absorption remains bounded by 1. The optional finite-size correction is shown as reverberation-room equivalent absorption area / sabins, not as a second coefficient curve. It estimates object area effects from panel perimeter, wavelength, porous thickness and density. Open porous edges add a separately flagged side-entry engineering estimate. Hybrid/membrane edge correction is extrapolative.
Corner mounting: wall-parallel corner mode keeps the rigid-backed stack and adds an adjacent-boundary field term. The 45° bridge mode discretizes the panel width into strips and solves the triangular rear wedge as a distribution of rigidly terminated cavity depths from near-zero at the wall contacts to half the panel span at the center. A user pressure-zone gain can represent a known local modal antinode, but it affects apparent/equivalent area only because room-modal amplitude cannot be inferred from panel geometry alone.
Design assistant: the first-round generator uses a constrained beam search over practical porous thicknesses, air gaps and optional facings/membranes. It performs a fast normal-incidence screening pass, then re-scores finalists with the full mounted diffuse/corner model. Suggested designs are candidates, not global mathematical proofs of optimality.
Limits: this is an advanced transfer-matrix engineering model, not a 3-D Biot/FE/BEM solver. Structural frame elasticity, detailed edge framing, exact modal radiation, nonlocal reaction, flanking paths and mounting hardware can require measured data or numerical multiphysics analysis.