Environmental noise screening tool Library: offline snapshot
Disclaimer: Results are indicative only. This tool is a preliminary screening aid — outputs should not be used as definitive compliance predictions or relied upon without independent professional verification.
A screening tool to predict noise levels at nearby receivers. Select source and receiver locations on the map, enter noise source data, and the tool will indicate potential compliance. Depending on location, the criteria will be automatically calculated or can be entered manually.

Development information

Residential development?
Source categories

Map

Set the site location (search or press S), then add sources and up to 4 receiver locations. Straight-line distances are calculated automatically. Markers can be dragged to fine-tune positions.
● Sources: not placed ● R1: not placed ● R2: not placed ● R3: not placed ● R4: not placed
Save JPG
Objects
Map data © OpenStreetMap contributors. Distances are straight-line only and do not account for terrain.

Planning & Design Code — Interface between Land Uses

Upload PDF and/or undertake noise modelling to understand noise sources and relevant provisions from the Planning & Design Code under Part 4 Interface between Land Uses.
Performance Outcome DTS/DPF Relevance
Sensitive receiver means: any use for residential purposes or land zoned primarily for residential purposes, childcare facility, educational facility, hospital, supported accommodation, or tourist accommodation.

Receivers & criteria

Assessment criteria selection
Select which criteria apply to this assessment.
LocationZone / SubzoneZoneLand useCriteriaDay LAeq
dB		Eve LAeq
dB		Night LAeq
dB		Night LAmax
dB

—
 Subject site — — — —

—
INL-5

—
INL-5

—
INL-5

—
INL-5
Day = 7 am–10 pm Night = 10 pm–7 am INL INL = Indicative Noise Level as defined in the Environment Protection (Commercial & Industrial) Noise Policy 2023 Clause 5. It is the averaged value of the Indicative Noise Factors for the applicable land use categories used to determine environmental noise criteria. Day = 7 am–6 pm (Mon–Sat) Evening = 6 pm–10 pm Night = 10 pm–7 am (& all day Sun) Criteria = EPA Pub 1826.5 Zone levels from Table 1 (urban) or Annex B (rural) of EPA Publication 1826.5, adjusted for distance to zone boundary and background noise classification. Minimum limits: 45/40/35 dB(A) D/E/N (urban), 45/37/32 dB(A) (rural). Night cap: 55 dB(A). Night Lmax criterion based on NSW EPA Day = 7 am–6 pm (Mon–Sat) / 8 am–6 pm (Sun) Evening = 6 pm–10 pm Night = 10 pm–7 am Residential: min(Intrusiveness, Amenity) | Non-residential: Amenity only Residential: trigger = lower of intrusiveness (RBL + 5 dB) and amenity criterion (Table 2.2 ANL − 5 dB + 3 dB facade correction). Non-residential (incl. hotel/motel): intrusiveness does not apply (NPI §2.1); trigger = amenity criterion only. NSW Noise Policy for Industry (2017).

VIC assessment parameters

Applies −1 dB per 100 m (rounded down, max 9 dB) when source zone ≠ receiver zone.
0 = fully Type 1 (residential), 1 = fully Type 3 (industrial). Per Figure 1 of Pub 1826.5, IF is derived from zones within 140 m / 400 m circles around each noise sensitive area (NSA). Leave blank to auto-derive per NSA, or enter a value to override all.


NSW assessment parameters

Auto-detected from planning zone code. Override if needed.
Intrusiveness: 40
Intrusiveness: 35
Intrusiveness: 35
Minimums: Day 35, Evening 30, Night 30. Intrusiveness = RBL + 5 dB (Table 2.1) — residential receivers only.

Derivation of noise emission criteria

Receiver Source zone INL (dB) Receiver zone INL (dB) Clause applied Day Night
Place source and receivers to see derivation

Assessment cases

Source contribution

Predicted sound pressure level at each receiver from individual sources, using the selected propagation method. Levels include ground effect, atmospheric absorption, and terrain screening where applicable.

Characteristic penalties

Apply a characteristic penaltyThe presence of an annoying characteristic attracts a decibel penalty under clause 13(3) of the Noise Policy. A penalty is applied because noise with such dominant characteristics are more noticeable and annoying when compared to a constant, broadband or steady noise. For a penalty to be applied, the Noise Policy requires the characteristic to be fundamental to the nature and impact of the noise. The characteristic should dominate the overall noise impact, rather than simply be part of it. Noise characteristics include tonal, impulsive, low frequency, intermittent and modulating. Refer to SA EPA Noise Policy Guideline (2023) for details on noise characters. to the predicted Leq levels before comparison with criteria. The penalised level is shown in the results table.
Modifying factor corrections from Fact Sheet C of the Noise Policy for Industry are added to the predicted noise level before comparison with project noise trigger levels. Applicable corrections include tonal (+2 or +5 dB), low-frequency (+2 or +5 dB), and intermittent (+5 dB, night only). Maximum combined correction: 10 dB.
Characteristic adjustments are added to the predicted noise level per EPA Publication 1826.5. Applicable characteristics include tonal (2–5 dB), impulsive (2–5 dB), and intermittent (3–5 dB). Maximum total adjustment: 10 dB.
Enter the applicable characteristic penalty (dB) to add to the predicted level for each period.
Receiver Day Leq penalty Evening Leq penalty Night Leq penalty
Receiver 1
Receiver 2
Receiver 3
Receiver 4

Predicted noise levels

Day LAeq
Evening LAeq
Night LAeq
Δ = Predicted − Criterion (≤ 0 is compliant) Residential/Rural Living caps: LAeq52 day, LAeq45 night, LAmax 60 night Night cap: 55 dB. Min limits: 45/40/35 (urban) or 45/37/32 (rural). LAmax 65 night.

Octave band spectrum

Recommended treatments (indicative)

GIS Export

Export assessment data for use in GIS applications (QGIS, ArcGIS, Google Earth).
GeoJSON — recommended for QGIS and ArcGIS. Tip: to get a Shapefile, open GeoJSON in QGIS → Export → Save Features As → ESRI Shapefile.
KML — use with Google Earth or Google Maps.
CSV — receivers only, for Excel or data joins.

Methodology

v2.0 — March 2026
Sound pressure level prediction
This calculator provides an indication of the sound pressure level at a specified distance due to various noise sources. Three propagation methods are available: a simple inverse square law screening method (no ground or atmospheric corrections), ISO 9613-2:1996 (which includes geometric divergence, ground effect, and atmospheric absorption based on temperature and humidity), and CONCAWE (Report 4/81, Manning 1981) which adds meteorological corrections for six atmospheric stability categories. It is intended as a screening tool to determine whether more accurate noise modelling and advice should be provided by an acoustic consultant.
Data sources
Planning zones
Planning zone classifications are detected automatically based on the placed receiver location. The data source varies by jurisdiction:
  • SA: PlanSA open data (data.sa.gov.au, CC-BY 4.0) — SA Planning and Design Code zones, served via a locally cached GeoJSON layer. Zone identification uses offline point-in-polygon lookup; users should verify classifications against the SA Planning Portal (SAPPA) for authoritative confirmation.
  • VIC: Victorian Planning Authority — Planning Schemes Online
  • NSW: NSW Planning Portal — Environmental Planning Instruments (LEPs and SEPPs)
SA zone detection is performed offline using a locally cached PlanSA GeoJSON layer (updated periodically from data.sa.gov.au). VIC and NSW zone detection query live APIs and require an internet connection. Users should verify zone classifications independently as data may not reflect recent planning amendments or interim provisions.
Noise criteria
Applicable noise criteria are applied based on the detected zone and the selected state. Criteria sources by jurisdiction:
  • SA: Environment Protection (Commercial and Industrial Noise) Policy 2023 (SA EPA)
  • VIC: Environment Protection Act 2017 and Environment Protection Regulations 2021 (effective 1 July 2021). Noise limits and assessment methodology are set out in EPA Publication 1826.5 — Noise limit and assessment protocol for the control of noise from commercial, industrial and trade premises and entertainment venues. Note: the previous SEPP N-1 framework no longer applies from 1 July 2021.
  • NSW: Noise Policy for Industry (NSW EPA, 2017) — the current recommended guideline for new developments. The NSW Industrial Noise Policy (2000) may still apply to some existing developments and premises under transitional arrangements. Users should confirm which policy applies to their assessment with the NSW EPA.
Users should confirm the applicable policy version and any site-specific conditions with the relevant environment protection authority.
Aerial imagery
Aerial imagery is provided by Esri World Imagery. Imagery currency varies by location — recent development or vegetation change may not be reflected.
Terrain elevation
Ground elevations are fetched from two sources: the Geoscience Australia 5 m LiDAR DEM via WCS (primary, covering most of Australia) and the Open-Elevation API serving SRTM data (~30 m resolution) as a fallback. Elevation data is used to determine effective source and receiver heights above sea level and to calculate terrain diffraction loss along propagation paths. The 5 m LiDAR DEM provides substantially better accuracy in areas of complex terrain. Accuracy of both sources may be reduced in areas of dense vegetation or built-up urban areas (where canopy or building returns affect the DEM surface).
State LiDAR/DEM resources for reference:
  • SA: Surveying and Spatial SA (SAILIS)
  • VIC: DataVic (Vicmap Elevation)
  • NSW: NSW Spatial Services (Elvis elevation portal)
Buildings
Building footprints are sourced from OpenStreetMap (OSM) via the Overpass API. Coverage varies by location — major urban areas generally have good coverage; regional and rural areas may be incomplete. Building heights where shown are sourced from OSM tags and may not be available for all structures. For critical assessments, verify building dimensions against survey data or local government records.
Cadastral boundaries
Property boundaries are sourced from the Esri World Boundaries and Places layer. For definitive cadastral information consult the relevant state land registry:
  • SA: Surveying and Spatial SA (SAILIS / Land Services SA)
  • VIC: Land Use Victoria (Landata)
  • NSW: NSW Land Registry Services (NSW LRS)
MBS 010 screening
MBS 010 (Ministerial Building Standard 010 — Construction requirements for the control of external sound) is a South Australian building standard published under the Planning, Development and Infrastructure Act 2016, available on the PlanSA website. It sets deemed-to-satisfy construction requirements for reducing the intrusion of road, rail, and aircraft noise into habitable rooms of residential buildings (Class 1, 2, 3, Class 4 and Class 9c aged care buildings).
Application of MBS 010 is triggered by the Noise and Emissions Overlay in the SA Planning and Design Code. The current version is Version 3 (1 May 2023). The tool determines overlay applicability using offline point-in-polygon lookup against a locally cached PlanSA GeoJSON layer (updated periodically from data.sa.gov.au); the aircraft noise (ANEF) contours are also sourced from the same dataset. Users should confirm overlay applicability and the current version of MBS 010 with the relevant consent authority.
User-imported site plans
The Site plan overlay tool lets the user import a PDF, PNG, or JPEG site plan and position it on the map as a semi-transparent Leaflet overlay with adjustable opacity, position, scale, and rotation. Aspect ratio is preserved from the image’s native dimensions on initial placement (with a cos(lat) compensation for 1° lng ≠ 1° lat in metres) and on corner-drag resizing by default; hold Shift to free-resize. PDF imports attempt to extract text from the title block to pre-fill the project name and, where present, a street address. Site plan overlays are purely visual — they do not participate in any calculation — and are serialised in saved assessments.
Modelling methodology
Zone detection
Receiver zones are detected automatically based on the receiver map location. SA zones are identified using offline point-in-polygon lookup against a locally cached PlanSA open-data GeoJSON layer (data.sa.gov.au); the SAPPA portal remains the authoritative reference for formal verification. VIC uses the Victorian Planning Authority API and NSW uses the NSW Planning Portal — both require an internet connection. Users should verify zone classifications independently, as data may not reflect recent planning amendments. Zone information is used to apply the applicable noise criteria for that receiver.
State noise criteria
Detected zones are mapped to applicable noise criteria using independent, state-specific logic in the Receivers & criteria panel. Changes to one state’s derivation rules do not affect another. The applied criterion and pass/fail outcome are shown in the Predicted noise levels table and the compliance strip.
  • SA — Indicative Noise Levels (INL) from the Environment Protection (Commercial and Industrial Noise) Policy 2023, derived from the land-use category of the host zone. Clauses 5(4), 5(5), and 5(6) cover special cases — noise emerging into another zone type, receivers within non-residential zones, and receivers in mixed-use areas. Character penalty policy (clause 13(3)) is applied via the Characteristic penalties panel.
  • VIC — Influencing Factor (IF) method from EPA Publication 1826.5, derived from zone type, distance to influencing zones, and LGA overrides where applicable. Day, evening, and night limits are then computed relative to the IF. Character adjustments are applied as described above.
  • NSW — Rating Background Level (RBL) plus intrusiveness and amenity assessments per the Noise Policy for Industry (2017). Project noise trigger levels are the lower of the intrusiveness criterion (RBL + 5 dB) and the amenity criterion (area-type dependent). Sleep disturbance screening is applied at night (LAmax vs RBL + 15 dB). Modifying factor corrections are applied via the Characteristic penalties / Modifying factors panel.
Sound power spectra and period normalisation
Each spectrum-based source holds an 8-band octave sound power spectrum (63 Hz to 8 kHz) plus a broadband Lw per period (day / evening / night / Lmax). Before propagation, a constant per-period offset specAdj = Lw(period) − 10·log10(Σ 10Lwi/10) is computed so that the 8-band energy sum matches the entered broadband Lw for that period. The shape of the spectrum is preserved; only its overall level is anchored to the period Lw. This lets the user enter a representative spectrum shape once and vary only the broadband Lw across day, evening, and night. For flat (uniform) spectra specAdj ≈ 0 and the ISO prediction converges to the simple-method result to within the ground-effect and barrier-effect contributions.
Building sources
Building sources model noise radiating from industrial or commercial building facades and roofs where the internal noise level (Lp) is known. The user draws a building polygon on the map, enters the building height, internal Lp per period, and wall and roof construction (Rw). The radiated sound power per facade or roof panel is calculated as Lw = Lp − Rw + 10·log10(S) + 6 dB, where S is the facade or roof area. Facade elements (roller doors, louvres, glazing) with their own Rw can be added per wall. Propagation from the resulting sub-sources uses the same method as point sources.
Line sources
Line sources model extended linear noise emitters (conveyors, pipelines, unbroken plant lines, rail corridors modelled outside CoRTN) where a length-normalised sound power is known. The user draws the line on the map and enters Lw′ per period (dB re 1 pW per metre) along with the source height. Each line is discretised into short segments with adaptive spacing based on total length; each segment is treated as a point sub-source with Lw,seg = Lw′ + 10·log10(lseg) and propagated to each receiver using the selected propagation method. Per-receiver contributions are energy-summed across sub-sources. Barrier, terrain, ground, reflection, and octave-band spectrum treatments are identical to point sources. Noise-map rendering uses the same sub-source expansion per grid cell.
Area sources
Area sources model spatially distributed emitters (open yards, stockpiles, open-air process equipment, car parks) where either a total sound power or a per-area sound power is known. The user draws the polygon on the map and chooses Total Lw or Lw per m2 as the input mode, entering values per period along with source height above ground. When the input is total Lw, the code distributes it across the polygon area via Lw,m² = Lw − 10·log10(A). The polygon is then sampled by a grid of sub-sources whose spacing is chosen to keep each sub-source’s contribution small relative to the total at the receiver; each sub-source is propagated independently via the selected method and energy-summed at the receiver. Barrier, terrain, ground, and reflection treatments are applied per sub-source. Large areas relative to the receiver distance produce a softer falloff than a point source (closer to 3 dB per doubling in the near-field, transitioning to 6 dB per doubling far from the area).
Road traffic — CoRTN with Australian adjustments
Road traffic noise is modelled using the UK Calculation of Road Traffic Noise (CoRTN) method with an Australian +2.5 dB adjustment. The user draws a road polyline on the map and enters, per period: hourly flow Q, mean speed V, percentage heavy vehicles %HV, gradient (%), road surface, and (optionally) a noise-reducing barrier. The code queries the SA DIT speed-zone service to pre-fill V where available. The correction chain follows CoRTN Charts 2–12 in order:
  • Basic noise level from flow (L10(18h) Chart 2 base curve)
  • Speed correction (Chart 4) and %HV correction (Chart 5)
  • Gradient correction (Chart 6)
  • Distance correction (Chart 7) using perpendicular distance to the line
  • Ground absorption (Chart 8) using the user’s G value along the propagation path
  • Angle of view correction (Chart 10) for finite road segments
  • Reflection correction from opposing facades (Figure 5)
  • Road surface correction
  • Low volume correction (Chart 12, LA10 only) for flows below 200 veh/h
  • Australian +2.5 dB adjustment applied as the final step
Barrier screening for CoRTN uses Charts 9 and 9a applied per path-length difference δ, with dual-carriageway geometry handled as two parallel lines and (in NSW) a 3-source-height model (0.5 m, 1.5 m, 3.75 m) energy-summed. LA10 is the primary CoRTN output; LAeq is derived via a periodic offset (approximately LA10 − 3 dB for typical continuous flows). Both metrics are available. The CoRTN engine is implemented as a shared-calc namespace (SharedCortn) so the same math runs on both the main thread (receiver-point calculations) and in a dedicated Web Worker (cortn-worker.js) that sweeps the full map grid for the CoRTN road noise map. The CoRTN noise map is independent of the ISO 9613-2:1996 map — both can be active simultaneously as separate Leaflet overlays — and exposes its own Day/Night, LA10/LAeq, height, and grid controls in the Modelling panel. Grid computation includes all correction chain steps except reflections (grid-mode reflections are forced to zero) and terrain profile screening; per-road barrier screening is applied with the user’s configured barrier geometry.
Barrier attenuation (ISO 9613-2:1996 §7.4)
Barrier screening is calculated from user-drawn barriers and OSM building footprints. For each source–receiver pair the dominant intervening barrier is identified and its Maekawa single- and double-diffraction attenuation Dz is computed per octave band from path-length difference. Three diffraction paths are energy-summed: over the barrier top and around each lateral end. The over-top path is capped at 20 dB (single diffraction) or 25 dB (double diffraction); lateral paths are uncapped. A line-of-sight check ensures no attenuation is applied when the direct ray clears the barrier top.
ISO 9613-2:1996 defines the barrier term Abar as an insertion loss, not a separate additive attenuation. The combined ground-plus-barrier contribution is therefore computed as AgrBar = max(Dz, Agr,bar) per band, where Agr,bar is the ground effect recomputed along two sub-paths (source → barrier top, barrier top → receiver) with the barrier top acting as a pseudo-receiver then a pseudo-source for Table 3 region factors. This §7.4 sub-path treatment is threaded through every ISO propagation call site — point, line, area, building, and road sources, for both receiver predictions and the noise map. It is most consequential for grazing geometries where per-band ground attenuation can exceed the barrier’s diffraction loss at some frequencies; for tall barriers at short distances the barrier IL dominates and the sub-path calculation is a no-op.
CONCAWE propagation model (Report 4/81)
The CONCAWE model predicts noise propagation from industrial plants for a range of meteorological conditions. It was developed from field measurements at three European petrochemical complexes and uses octave band calculations from 63 Hz to 4 kHz (this tool extends to 8 kHz using the 4 kHz attenuation values as a conservative approximation). The predicted sound pressure level at a receiver is:
Lp(f) = Lw(f) + D − K1 − K2 − K3 − K4 − K5 − K6
where:
  • K1 — Geometric spreading: 10·log10(4πd²) dB. Spherical propagation from a point source (identical to ISO 9613-2:1996 Adiv).
  • K2 — Atmospheric absorption: per octave band, calculated from temperature and relative humidity per ISO 9613-1 (identical to ISO 9613-2:1996 Aatm).
  • K3 — Ground attenuation: for acoustically hard surfaces (concrete, water), K3 = −3 dB. For soft ground (grass, soil), K3 is determined from empirical curves as a function of frequency and distance (CONCAWE Figure 1, Appendix II polynomials). Distance clamped to 100–2000 m.
  • K4 — Meteorological correction: accounts for refraction by wind and temperature gradients. Six meteorological categories are defined from the Pasquill atmospheric stability class and the vector wind speed (component of wind from source to receiver). Category 1 represents strong upwind propagation with unstable atmosphere (greatest attenuation). Category 4 is neutral (zero correction). Category 6 represents moderate to strong downwind propagation (greatest enhancement). This implementation uses the Simplification 2 distance-independent K4 values from §6.2, which the authors showed to have only 0.5 dB(A) accuracy loss compared to the full distance-dependent polynomial model.
  • K5 — Source height correction: reduces the combined ground and meteorological effect (K3 + K4) for elevated sources (height > 2 m) based on the grazing angle between the source, receiver, and ground plane (CONCAWE Figure 9).
  • K6 — Barrier attenuation: calculated using the Maekawa diffraction method, consistent with SoundPLAN’s implementation of CONCAWE which substitutes the General Prediction Method (GPM) barrier algorithm.
The model is validated for source–receiver distances of 100–2000 m and wind speeds up to 7 m/s.
Key differences from ISO 9613-2:1996:
  • Ground effect: CONCAWE uses single empirical curves based on neutral-condition measurements (Parkin & Scholes). ISO 9613-2:1996 uses a three-region detailed method (source, middle, receiver) based on favourable (downwind) conditions. CONCAWE ground attenuation (K3) is applied independently of barrier screening — there is no §7.4 insertion-loss interaction.
  • Meteorological conditions: CONCAWE explicitly models six weather categories. ISO 9613-2:1996 assumes favourable (downwind) conditions; an optional Cmet correction (§8) is available in the Propagation accordion to convert worst-case downwind to a long-term average (see “Meteorological correction” below). CONCAWE is preferred when assessments require prediction under specific meteorological conditions, particularly in Victoria where neutral conditions are often specified.
  • Frequency range: CONCAWE covers 63 Hz to 4 kHz. This tool extends to 8 kHz using the 4 kHz attenuation values.
Reference: Manning, C.J. (1981). The propagation of noise from petroleum and petrochemical complexes to neighbouring communities. CONCAWE Report 4/81.
LAmax barrier and terrain screening
LAmax barrier and building attenuation is calculated using the Maekawa diffraction formula applied per octave band to the source Lmax spectrum. Ground effect (Agr) is not applied to LAmax predictions. Terrain screening is applied where enabled. Barrier and terrain screening is applied regardless of the selected propagation method.
Terrain screening — Deygout 3-edge method
Where ground elevation data is available, the terrain profile between each source and receiver is sampled from DEM data. The Geoscience Australia 5 m LiDAR WCS service is queried as the primary source; Open-Elevation (SRTM ∼30 m resolution) is used as a fallback where WCS data is unavailable. The terrain profile is processed with the Deygout construction: the profile point with the greatest path-length difference is selected as the primary diffracting edge, and the source-side and receiver-side sub-profiles are then searched recursively for secondary edges. Up to three edges (one primary, two secondary) are considered per profile and their Maekawa Dz contributions combined per the Deygout energy-sum rule. Terrain diffraction is computed per octave band. Where terrain and building barriers coincide, the screening term is Ascreen = max(Abar, Aterr) per band, and this combined insertion loss enters the ground-barrier interaction (§7.4) as a single Dz-equivalent value.
Ground absorption zones
In addition to the global ground factor G set in the Propagation method panel, the user can draw local ground absorption polygons on the map with individual G values (0 = hard/reflecting, 0.5 = mixed, 1 = soft/porous). For each source–receiver ray the three ISO 9613-2:1996 Table 3 regions (source side, middle, receiver side) are each assigned a length-weighted average G from the zones intersecting that region (_gzWeightedG). The resulting Gs / Gm / Gr triplet drives the per-band ground attenuation and, when a barrier is present, also the two sub-path Agr,bar calculations described under Barrier attenuation (§7.4) above. When no zones are drawn, all three regions use the single global G value.
Facade reflections
Facade reflections are calculated using the image source method per ISO 9613-2:1996 §7.5. For each source–receiver pair, building and barrier edges within 50 m of the receiver are tested as reflecting surfaces. The reflected contribution from the dominant surface is calculated as a separate propagation path (using the same propagation method as the direct path) and energy-added to the direct path. A single dominant reflection is considered per source–receiver pair. Grazing incidence is excluded (incidence angle >80° relative to surface normal). The reflection coefficient ρ is configurable per custom building via the Surface type dropdown in the building edit panel (ISO 9613-2:1996 Table 4): flat hard walls (ρ = 1.0, default), walls with windows/openings (ρ = 0.8), factory walls 50% openings (ρ = 0.4), open installations (ρ = 0.0), or a custom value. The reflected contribution is adjusted by 10·log10(ρ) dB before energy summation. Barrier surfaces are assumed hard (ρ = 1.0). Reflections are enabled via the Reflections toolbar button and apply to both receiver predictions and the noise map.
Meteorological correction (ISO 9613-2:1996 §8)
The ISO 9613-2:1996 §8 meteorological correction Cmet converts the worst-case downwind prediction to a long-term average. It is applied as a broadband subtraction after per-band energy summation, using equations (21) and (22):
  • Near-field (dp ≤ 10·(hs + hr)): Cmet = 0 — eq. (21)
  • Far-field (dp > 10·(hs + hr)): Cmet = C0 · [1 − 10·(hs + hr) / dp] — eq. (22)
C0 is a site-specific long-term meteorological factor (0–5 dB; default 2 dB per §8). Cmet is applied to all Leq paths (single-point, noise map grid, area and line sub-sources, building sub-sources). It is explicitly excluded from LAmax predictions, which remain worst-case downwind. The toggle defaults to OFF, preserving conservative worst-case downwind predictions as the default. The UI is in the Propagation accordion, visible only when ISO 9613-2:1996 is the active method.
Operating % of period
The calculation provides an option to enter the percentage of the assessment period during which each source is operating. A 10·log10(pct/100) weighting is applied to the source level: 100 % → 0 dB, 50 % → −3 dB, 25 % → −6 dB.
Character penalties and modifying factors
A noise that contains a dominant character (tonal, impulsive, low-frequency, intermittent, or modulating) is generally more intrusive than steady broadband noise at the same sound pressure level. The applicable state policy may therefore require a per-receiver, per-period decibel penalty to be added to the predicted Leq before comparison with the criterion. Penalties are set in the Characteristic penalties / Modifying factors panel and flow through to all downstream outputs (compliance strip, per-case comparison, PDF and Word reports, and the compliance noise-map overlay). Jurisdictional summary:
  • SA (Noise Policy 2023, clause 13(3)) — tonal, impulsive, low-frequency, intermittent, and modulating characters each attract a decibel penalty. The policy requires the character to be fundamental to the noise impact, not incidental.
  • NSW (Noise Policy for Industry, Fact Sheet C) — tonal (+2 or +5 dB), low-frequency (+2 or +5 dB), intermittent (+5 dB night-time only). Maximum combined correction per period: 10 dB. Per Fact Sheet C note 2, if a tone is ≤ 160 Hz and is also classified as low-frequency, only one 5 dB correction is applied.
  • VIC (EPA Publication 1826.5) — tonal (2–5 dB), impulsive (2–5 dB), intermittent (3–5 dB). Maximum total adjustment: 10 dB.
Multiple sources — energy summation
Where more than one noise source is present, contributions at each receiver are combined by energy summation: Ltotal = 10·log10(Σ 10Li/10). The combined level will always be higher than the loudest individual source, but by no more than 3 dB when two equal sources are present.
Assessment cases
An assessment case is a named, independent configuration of the tool’s inputs (source operating states, movements, barriers, receiver penalties) representing a single scenario — for example “existing operations”, “proposed — worst case”, or “with mitigation”. Multiple cases can be enabled simultaneously and are compared side-by-side in the Assessment cases panel. For each enabled case the tool runs the full propagation chain and produces a per-receiver predicted level, a per-case compliance outcome, and (optionally) a per-case noise map overlay. The governing case is the one that produces the highest level at each receiver for each period; it drives the overall compliance outcome reported in the summary. Saved assessments serialise the full list of cases (including enabled state) so the comparison is reproducible.
Noise map
The noise map generates a predicted sound level grid across the map extent using the selected propagation method. Grid resolution is adaptive (5–50 m based on zoom level). Contour lines are interpolated from the grid and represent equal noise level lines in dB(A). The noise map uses the same source inputs, propagation method, and receiver height as the receiver predictions. Contour accuracy degrades at the edges of the grid and at large distances from sources.
Terrain screening and heatmap smoothing. Per-grid-cell terrain insertion loss is computed using the Deygout three-edge method (ISO 9613-2:1996 / ISO/TR 17534-3) at each grid node. To suppress radial spike artefacts arising from DEM grid discretisation, the per-band terrain IL field is then spatially filtered with a separable 5 × 5 Gaussian kernel (σ = 0.5 cells, radius 2 cells). This smoothing can reduce the apparent terrain screening level by up to 3–6 dB at receivers located inside narrow terrain shadow zones (shadow width < 4 ×  grid spacing). The point-receiver compliance values shown in the receiver panel are computed along the exact source–receiver ray using the same Deygout method but without grid smoothing, and are therefore the authoritative values for compliance assessment. The noise map should be used as a spatial screening tool only, not as the basis for compliance decisions.
3D Scene Viewer
The 3D Scene Viewer is a read-only visualisation of the current assessment, intended for visual verification of spatial relationships — source heights relative to barriers, screening geometry, building obstruction, and terrain profile — in a single orbital view. It is not a calculation tool: no propagation, criteria, or noise-map result depends on it, and editing the assessment is done in the 2D map and side panels as before. The viewer is opened from the toolbar (3D View button) or with the V shortcut, and is available regardless of whether terrain is enabled (a flat fallback plane is rendered when no DEM data is cached for the current viewport).
Rendering uses Three.js (r128) loaded lazily on first open. The scene uses an equirectangular local projection with a cos(centre latitude) compensation for 1° lng ≠ 1° lat in metres; X is east, Z is −north, and Y is elevation. Elevations are normalised to a local datum (the minimum sampled DEM elevation in the viewport at modal-open time is subtracted from every Y value), so a site at 30 m ASL with 22 m of relief renders anchored at Y=0 with a visible 22 m variation rather than as a floating mesa above the building plane. Objects placed on top of a building footprint render at roof + object_height in 3D rather than ground + object_height — a rooftop fan with height_m = 2 on a 10 m custom building lands at Y=12, matching what the propagation engine sees.
Objects rendered in the scene: terrain mesh (vertex-coloured green–brown–tan ramp by normalised elevation, with cells skipped where DEM coverage is missing); OSM buildings (single merged grey mesh, opacity 0.7); custom buildings (blue, opacity 0.8); building sources (orange, opacity 0.8); barriers (semi-transparent green walls with slope-following bases and a darker crest accent line); ground absorption zones (semi-transparent G-factor colour ramp from grey at G=0 through olive at G=0.5 to green at G=1); point sources (red spheres); line sources (red tubes following per-vertex elevation); area sources (semi-transparent red polygons); receivers (R1–R4-coloured spheres matching the 2D marker palette); and sprite labels for sources and receivers (drawn through geometry with anti-stacking offsets when multiple labels are placed within 5 m of each other).
Code quality and validation
ISO/TR 17534-3 validation
The core propagation calculations are validated against the reference test cases defined in ISO/TR 17534-3:2015 — Acoustics: Software for the calculation of sound outdoors. This technical report defines input geometries and reference output values for testing implementations of ISO 9613-2:1996. The in-app “Run validation” button (in the Propagation method panel) executes the three ground-attenuation test cases and reports the computed vs reference total for each:
  • T01: Reflecting ground (G = 0) — reference 44.29 dB — pass (±0.05 dB)
  • T02: Mixed ground (G = 0.5) — reference 41.53 dB — pass (±0.05 dB)
  • T03: Porous ground (G = 1.0) — reference 39.14 dB — pass (±0.05 dB)
In addition to the in-app validation, a headless vitest regression suite of 233 tests across 5 files (calc, geometry, multi-source, iso17534, area-source) is run on every source change via npm test. The vitest ISO/TR 17534-3 suite additionally exercises T04 (spatially-varying ground — pass) and the combined barrier-plus-ground cases T08, T09, and T11. Current conformance:
  • T08: long barrier, varying G — reference 32.48 dB — within ±0.6 dB
  • T09: short barrier, varying G — reference 32.93 dB — within ±0.25 dB (tightened from ±1.0 dB after the §7.4 ground-barrier sub-path fix)
  • T11: cubic building barrier — reference 41.30 dB — within ±1.0 dB
The residual error in T08 and T11 is not from the §7.4 sub-path treatment — in those cases the combined barrier/terrain screen Ascreen exceeds Agr,bar in every band, so the sub-path calculation is a numerical no-op. It comes from approximations elsewhere: the lateral-diffraction energy summation, the Fresnel z construction, and the spatially-varying ground-region extent averaging. These cases are noted as partial conformance. Users can re-run the in-app ground-attenuation validation at any time via the “Run validation” button.
Terrain screening validation
No terrain-specific test cases exist in ISO/TR 17534-3. The terrain diffraction implementation is validated by regression test: for a flat terrain profile (no elevation variation), the tool is required to return 0 dB terrain insertion loss for all source–receiver pairs. This is confirmed automatically as part of the validation suite.
Conformity statement
This tool implements ISO 9613-2:1996 (Attenuation of sound during propagation outdoors — Part 2: General method of calculation) and is partially conformant with ISO/TR 17534-3:2015 based on the test cases listed above. The GoI-form (Geometry of Implementation) identifies the following implementation status:
  • Ground effect (§7.3): fully implemented, per-band
  • Atmospheric absorption (§7.2): fully implemented
  • Geometric divergence (§7.1): fully implemented
  • Barrier / building screening (§7.4): implemented, single dominant path, top and lateral diffraction
  • Terrain screening: partial — single dominant ridge, per-band Maekawa diffraction, DEM from Geoscience Australia WCS LiDAR (5 m, primary) with Open-Elevation SRTM fallback
  • Meteorological correction (§8): not implemented in the ISO 9613-2:1996 path (Cmet correction) — available via the CONCAWE propagation model, which provides explicit six-category meteorological corrections based on Pasquill stability class and vector wind speed
Security
User-entered data (source names, receiver names) is escaped before insertion into the interface using DOM-native escapeHTML() to prevent cross-site scripting (XSS). No user data is transmitted to any server — all calculations run entirely in the browser.
Recommended treatments (reverse calculations)
The Recommended treatments panel performs three reverse calculations for receivers that exceed their criteria. All outputs are indicative and subject to detailed acoustic design.
Maximum allowable Lw
For each point source and exceedance period, the panel computes the maximum source sound power level that would just hold the receiver at the criterion. The method energetically subtracts the subject source's contribution from the total predicted level to obtain the “other sources” linear sum. The allowable contribution for the subject source is then criterion_linear − other_linear, and the allowable Lw follows from the same attenuation path as the forward calculation. A source is flagged Others exceed (amber) when the combined contribution of all other sources alone exceeds the criterion, meaning no reduction of this source alone can achieve compliance.
Indicative barrier height
Barrier height is estimated using the simplified Maekawa formula at 500 Hz:
IL = min(20, 10 · log10(3 + 20 · 2δf/c))
where δ is the diffraction path-length difference (m), f = 500 Hz, and c = 340 m/s. The barrier is placed at the midpoint of the source–receiver path; placement closer to the source or receiver yields greater attenuation for the same height. The source height used is that of the dominant contributing source at the specific receiver and period. The search covers 1.5 m to 10.5 m in 0.5 m steps; a required insertion loss that cannot be achieved within this range is flagged as impractical (>10 m). A 2 dB robustness margin is included in the required insertion loss. The 500 Hz approximation is conservative for broadband sources but may overestimate barrier effectiveness for low-frequency-dominated sources (e.g. transformers, large reciprocating machinery) — detailed octave-band modelling is required in those cases.
Operating time restrictions
The tool applies a 10 · log10(%/100) weighting to source levels for the entered operating percentage. The maximum allowable operating percentage is the inverse of this weighting applied to the required contribution reduction: new% = current% × 10Δ/10, where Δ is the required reduction in contribution (negative). The output is presented as a percentage only — conversion to hours depends on the assessment period definition for the applicable jurisdiction and is the responsibility of the user.
Limitations
This tool is a screening tool intended to indicate whether more detailed assessment is warranted. It should not be used as a substitute for a full acoustic assessment by a qualified acoustic consultant. Predicted levels should be verified against site measurements before use in formal reports. Results are sensitive to source power level inputs — users are responsible for the accuracy of Lw values entered.