Fast Noise Analyzer
High-speed pseudo-acoustic noise propagation engine inspired by ISO 9613-2: geometric spreading, air absorption, ground effect, Maekawa barrier diffraction, and foliage attenuation.
Overview
Fast Noise Analyzer simulates environmental outdoor noise propagation from multiple emitter sources (roads, machinery, trains) onto arbitrary target meshes (building façades, terrain, public realm surfaces) and discrete probe locations.
The engine uses an ISO 9613-2 inspired engineering framework:
📐 Maekawa Barrier Diffraction
Calculates path-length difference (δ) over obstacle rooflines and berms using rubber-band convex hull profiles: Dz = 10 log10(3 + 20N).
🌿 ISO 9613-2 Foliage Attenuation
Evaluates actual ray penetration length through porous tree/vegetation volumes (Soft Context), applying continuous path-length attenuation.
🌱 Ground Absorption Effect
Incorporates ground interference attenuation Agr based on mean propagation height over terrain and acoustic ground factor G (hard vs. porous soil).
⚡ Unified Mesh Collider & Culling
Pre-merges context meshes into single accelerated spatial structures and culls distant emitters (>45 dB below dominant sources) for order-of-magnitude faster solving.
Inputs (11)
- Run (Boolean): Execution toggle switch (displays OFF when disabled to save compute).
- AnalysisMeshes (Mesh list): Building massing or terrain meshes to receive the vertex-colored noise heatmap.
- ContextBuildings (Mesh list, optional): Solid obstacle meshes (buildings, barriers, berms) causing acoustic diffraction.
- SoftContext (Mesh list, optional): Vegetation volumes for foliage path attenuation.
- Emitters (Point list): Source coordinates (e.g. from Noise Emitter From Curves).
- EmitterPower (Number list): Sound pressure level Lp at 1 m in dB(A) (default:
85.0). - AnalysisPoints (Point list, optional): Discrete inspection probe points (e.g. window centers, balconies).
- SensorOffset (Number): Normal offset distance in meters to prevent self-shadowing by the target mesh (default:
0.2 m). - CustomColors (Colour list, optional): Gradient color spectrum (quiet to loud).
- Ground (Mesh, optional): Terrain mesh used to determine propagation height for ground effect calculations.
- GroundFactor (Number, optional): Ground factor G:
0.0= hard paving/water,1.0= soft turf/lawn (default:0.5).
Outputs (7)
- NoiseMeshes (Mesh list): Heatmap meshes with continuous vertex colors.
- MeshPoints (Point list): Evaluated vertex coordinate points.
- MeshValues (Number list): Sound pressure levels in dB(A) corresponding to each mesh point.
- AnalysisPtsOut (Point list): Pass-through of discrete test probe points.
- AnalysisPtsValues (Number list): Evaluated dB(A) levels at the discrete probe locations.
- DashboardData (Text / JSON): Unified JSON data payload with min, max, energetic average, and arithmetic average for the Analysis Dashboard.
- Info (Text): Detailed acoustic disclaimer, calculation metrics, and active attenuation breakdown.
Method, assumptions and simplifications
The same notes appear in the component's Info output.
Method
- For each receiver (every vertex of AnalysisMeshes, moved Offset along its normal, plus the AnalysisPoints) the level from each emitter is Lp = Power - 20 log10(d) - 0.0028 × d, with d the straight distance (min 1 m).
- Emitters more than 45 dB below the loudest free-field contribution at that receiver are added without screening.
- Ground effect (only with Ground): A_gr = G × max(0, 4.8 - (2 h_m / d)(17 + 300 / d)), with h_m the mean height of source and receiver above the Ground mesh.
- If the straight path hits HardCtx, the barrier term comes from a height raster of the obstacle tops (about 500 × 500 cells, at least 0.5 m): the path over the tops is the upper convex hull of the profile, δ = its extra length, N = 2δ / λ with λ = 0.69 m (500 Hz), Dz = 10 log10(3 + 10 N) as in ISO 9613-2 with C2 = 20, C3 = 1 and Kmet = 1, capped at 20 dB (one edge) or 25 dB (several). The larger of Dz and A_gr is used. If no top is found, 5 dB.
- If the path is not screened and crosses SoftCtx, the length through foliage gives 0 dB below 10 m, 1 dB for 10-20 m and 0.05 dB/m up to 200 m.
- Contributions are added as energy: L = 10 log10(Σ 10^(Li/10)).
- Mesh colours span the min to max of the run; Info and Dashboard report max, min, energetic and arithmetic averages over vertices.
Assumptions
- Power is the A-weighted level at 1 m of each point (default 85 dB(A)); roads come as point sets from Noise Emitter From Curves.
- One 500 Hz band stands for broadband traffic noise; air absorption 2.8 dB/km (20 °C, 70 % RH).
- Model units are metres; terrain screens only if it is also in HardCtx.
Simplifications and limits
- No reflections, reverberation or weather effects (wind and temperature gradients).
- Diffraction only over the top of obstacles, never around their sides.
- Ground effect uses one G for the whole site and the 3D distance; foliage is ignored on paths already screened.
Workflow
- Noise Emitter From Curves -> Emitters and Power; buildings -> HardCtx; tree crowns -> SoftCtx; terrain -> Ground.
- Mesh_dB or Pts_dB -> Point Probe or Analysis Dashboard.
References
- ISO 9613-2:1996, Attenuation of sound during propagation outdoors, Part 2: general method of calculation.
- ISO 9613-1:1993, calculation of the absorption of sound by the atmosphere.
- ISO 9613-2:1996, Acoustics: attenuation of sound during propagation outdoors, Part 2 (barrier term, C2 = 20).
- Maekawa Z. (1968), Noise reduction by screens, Applied Acoustics 1(3).