How to choose noise prediction software: the questions that decide it
Noise Mapping · SoundPLAN Asia · · 8 min read
Noise prediction packages are compared badly. Buyers line up feature lists, find them largely identical, and then decide on price or on whichever demonstration was most recent. The feature lists look identical because at this level they mostly are: every serious package models sources, propagates sound and draws contours. What separates them is narrower and much more consequential, and it is worth knowing which questions actually determine the outcome before you sit through any demonstration.
1. Which calculation standard must your report cite?
This is the first question and it disqualifies faster than any other. Not "which standards does the software support" - which standard does the authority reviewing your report require, by name and version.
An industrial permit in most of Asia and the Gulf will expect ISO 9613-2. A strategic map against a European-style framework will expect CNOSSOS-EU, and which amendment matters. Occupational work will expect ISO 9612. National road and rail methods differ from country to country and are often mandatory rather than optional.
Write the list down before you shortlist. Then ask each vendor to confirm, in writing, how each named standard and version is implemented. A package that covers four of your five required methods means subcontracting the fifth, forever.
2. What scale will your largest model be?
A single industrial site with 200 sources and a city-wide map with several hundred thousand receiver points are different computational problems. The second is governed by calculation time, memory behaviour, multi-core use and whether the workflow supports splitting a job.
The honest test is not a benchmark figure in a brochure. It is to take your own largest realistic dataset into a trial and time it. Vendors who are confident will help you do this; the ones who deflect are telling you something.
3. Which source types, and in what combination?
Most projects need more than one. A corridor scheme needs road and rail. A plant near an airport needs industrial and aircraft. Combined assessment matters more than any individual source model, because the reviewer wants the total at the receptor. Check that the package combines sources under the method your authority requires, not just that it models each one separately.
4. What does it have to produce?
Contours are the easy part. Ask about the outputs that end up in the report:
- Facade levels per floor, not just a ground-level grid
- Population exposure statistics, if your framework requires exposure counts rather than contours
- Comparison scenarios - with and without mitigation - as a difference map, not two separate pictures
- Tabulated receptor results that can be pasted into a report without retyping
- Export formats your GIS and your client's GIS both read
5. What can it import, and how painful is that?
Model building, not calculation, consumes the project budget. The questions that matter are how terrain, building footprints with heights, road networks with traffic attributes and GIS layers come in, and how much manual repair is needed after import. Bring a real dataset - your own DTM and building layer, not a sample - to the trial.
6. Can the result be defended two years later?
Noise reports get challenged, sometimes long after the author has moved on. The package should let you record source data provenance, keep scenarios side by side, and reproduce a result from an archived model. If the answer to "where did this number come from" lives only in somebody's memory, the software has failed at the one thing that matters in a dispute.
7. Licensing, and what happens at the edges
- Single-seat or network licence, and what happens when a second person needs it during a deadline week
- Whether modules can be added later or whether the tier is fixed at purchase
- What a version upgrade costs and whether it affects reproducibility of an archived model
- Whether a temporary licence is available for a peak workload
8. Who answers when it goes wrong on a Friday
Under-weighted in most comparisons and decisive in practice. Support in your time zone and your language, from somebody who understands the standard you are working to rather than only the software, is worth more than any feature. Ask who answers, where they are, and what their response time actually is.
A note on comparing against another package
If you are evaluating us against another product, run the comparison on your own project rather than on either vendor's material - ours included. Take a model you already have, with a receptor whose answer you know, build it in both, and see which one gets there with less repair work and a clearer audit trail. Any claim a vendor makes about a competitor's product, including a claim we might make, is worth less than one afternoon of your own testing.
SoundPLAN supports 50+ international calculation standards including ISO 9613-2, CNOSSOS-EU and ISO 9612, and is supplied across Asia and the Middle East by SoundPLAN Asia. Send us your list from question 1 and we will confirm how each one is handled before you commit to anything - that conversation is free and it is quicker than a demonstration.
Compare the SoundPLAN products · The full standards list · Where to buy in your country