Status note (updated 25 August 2026). This article describes the May 2026 versions: acmap v0.5 and noise-barrier-calc v0.8. The shared engine has changed substantially since — ISO 9613-2:2024 screening became the default, §7.4.3 lateral diffraction was added, and the composition of ground effect and screening, the K_met distances, power conservation in source discretisation and spectrum normalisation were all corrected. For the current state see the acmap and noise-barrier-calc wikis and the live tools.
Live in the Open Lab section are the two browser-based acoustic tools ST-LINE maintains as a working lab for our own environmental-acoustics algorithms. Open source on GitHub, fully client-side (no backend, no data upload), sharing the same ISO 9613-2 §7.4 physics engine with selectable Maekawa diffraction.
acmap — acoustic map (v0.5)
Visualises a sound-level survey loaded as CSV (lat, lon, Leq) and simulates point-source propagation with Maekawa diffraction on real buildings fetched from OpenStreetMap via Overpass API. Two complementary modes: Measure shows the loaded survey as a coloured heatmap with arithmetic statistics over the loaded points; Predict computes the sound level over a receiver grid with diffraction on the OSM buildings (worst-screen-wins, one obstacle per source–receiver pair).
What’s new in v0.5: an explainer block above the statistics that makes clear what is being computed — the “Mean Leq” is the arithmetic mean of the loaded survey, not an energy-averaged mean, and the block says so explicitly to avoid the common misreading. Per-stat technical sublabels and a fixed offset on the OSM building tooltip round out the iteration.
noise-barrier-calc — noise barrier calculator (v0.8)
noise-barrier-calc — live tool · wiki
Verifies the effectiveness of a noise barrier for a point, line, or area source over a configurable receiver grid. Computes the insertion loss (IL = Leq_no_barrier − Leq_with_barrier) in octave bands with selectable ISO 9613-2 §7.4 screening (with K_met) or Maekawa 1968 diffraction, and draws a banded map plus 5 dB isolines.
v0.8 collects several months of work on robustness and UX:
- Adaptive receiver grid: auto-fits the source’s bounding box + a user-configurable buffer per side. v0.7 had a fixed 300×300 m grid that truncated extended line/area sources.
- Up to 5 manual receivers at specific points (e.g. sensitive façades) with a dedicated results table, numbered draggable markers, and a CSV-export
manualflag distinguishing grid points from pointwise ones. - Result-explainer notes, per-stat technical sublabels, and supplementary “diffracted only” statistics — more indicative of barrier effectiveness than the full-grid mean (which includes front-of-barrier zones with IL≈0).
- Auto-redraw on view-mode change (Insertion Loss / Leq with barrier / Leq without) — no recompute needed.
- Defensive fallback on an empty barrier-height field: previously produced
IL = ∞in the statistics, now falls back to “no barrier” with a clear status warning. - Compact panel layout that eliminates the wasted vertical space between sections of differing height.
Status, limits, and what they are NOT
Both tools are demonstrative: useful for exploration, teaching, and preliminary assessments. They do not replace a calculation model adequate to the required methodology — CadnaA, SoundPLAN and NoiseModelling are common examples, but “certified software” is not a category the regulation provides for — and they do not replace the judgement of a registered competent acoustic technician. Barrier/building diffraction is modelled as the single dominant obstacle per source–receiver pair (worst-screen-wins, not rigorous ISO 9613-2 multi-screen).
acmap roadmap: Overpass mirror fallback + embedded offline dataset. noise-barrier-calc roadmap: lateral diffraction, multi-barrier cascade, source spectra from real CSV.
Apache-2.0 with patent grant. Code, issues, contributions on GitHub.