# ISO 9613-2:2024/DAmd 1: the draft amendment takes barrier diffraction back to the 1996 formulas
> The draft amendment rewrites Dz and Kmet in Clause 7.4 and restores their 1996 form. How much the numbers move, the two errata in Annexes A and B, and why the third edition already in drafting argues against hard-coding anything.
Published: 2026-09-03
Category: regolatorio
Tag: iso-9613-2, standards, environmental-acoustics, diffraction, barriers, deep-dive
Page:
---
In January 2024 the second edition of ISO 9613-2 rewrote Clause 7.4, the barrier diffraction term: a new form for the barrier attenuation Dz, a new meteorological correction Kmet, and a threshold zmin the previous edition never named. We covered that rewrite in [what changes against the 1996 edition](/en/blog/iso-9613-2-2024-diffrazione-barriere/).
Since then an amendment has appeared in the ISO catalogue: **ISO 9613-2:2024/DAmd 1**, three pages, ISO/TC 43/SC 1, status *Under development*, stage *DIS ballot initiated: 12 weeks* [40.20]. It rewrites those same two formulas again — and restores the form they had in 1996.
Before anything else: **this is a draft**. An expert report cites the published edition, not a DIS out for ballot, and a draft's content can still change at any stage. This page is for something else: knowing what is about to move, and by how much, before it moves.
## The two formulas that change
The formulas quoted here are the ones the amendment modifies, reproduced for technical comparison only. The text, tables and figures of the standard are not reproduced: ISO 9613-2:2024 and the DAmd 1 draft are sold through the ISO catalogue and national standards bodies.
```
Dz — barrier attenuation, formula (18)
1996: Dz = 10·log₁₀(3 + (C₂/λ)·C₃·z·Kmet)
2024: Dz = 10·log₁₀[1 + (2 + (C₂/λ)·C₃·z)·Kmet] for z > z_min
Dz = 0 for z ≤ z_min
DAmd 1: Dz = 10·log₁₀[3 + (C₂/λ)·C₃·z·Kmet] for z > z_min
Dz = 0 for z ≤ z_min
```
```
Kmet — meteorological correction, formula (21)
1996: Kmet = exp{-(1/2000)·√[d_SS·d_SR·d / (2z)]}
Kmet = 1 for z ≤ 0
2024: Kmet = exp{-(1/2000)·√[(max(d_SS,d_SR)+e)·min(d_SS,d_SR)·d / (2(z-z_min))]}
DAmd 1: Kmet = exp{-(1/2000)·√[d_SS·d_SR·d / (2z)]}
Kmet = 1 for z ≤ 0
```
The DAmd 1 column and the 1996 column are the same formula. The amendment moves the factor 3 back outside the bracket multiplied by Kmet, restores the symmetric product dSS·dSR·d under the square root, puts 2z back in the denominator in place of 2(z − zmin), and reinstates the clause Kmet = 1 for z ≤ 0.
Three things from the 2024 edition stay: the zmin threshold, lateral diffraction around the vertical edges (7.4.3) with its combination rule (7.4.4), and the upper limits — 20 dB for a single edge, 25 dB for multiple diffraction. The amendment leaves all three alone.
One line of the earlier post on this blog also needed fixing: it stated that the direct distance `d` appeared only in the 2024 edition. It does not — the canonical 1996 Kmet already carries it, and that is exactly what the draft brings back. That post has been updated.
## zmin stays, and changes not one decibel
The threshold survives, which makes it easy to read the amendment as a partial retreat. The arithmetic says otherwise.
zmin = −2λ / (C₂·C₃). Substitute it into the argument of the 1996 logarithm:
```
3 + (C₂/λ)·C₃·z_min = 3 + (C₂/λ)·C₃·(−2λ/(C₂·C₃)) = 3 − 2 = 1
```
and 10·log₁₀(1) = 0. The same substitution in the 2024 form gives 1 + (2 − 2)·Kmet = 1, which is again 0 dB. **zmin is the point where both formulas already vanish on their own**: it is not a physical threshold introduced by the 2024 edition, it is the boundary of the domain where the logarithm has an argument ≥ 1, written out explicitly. It is needed — below that value the 2024 form would take the logarithm of a negative number — but it produces no numbers different from the ones the formula produces anyway.
The consequence: with the amendment, **Dz becomes numerically identical to the 1996 one in every geometry**, with the threshold left in place. For diffraction over the top edge the retreat is not partial: it is the 1996 formulation reinstated, plus a domain clause.
## How much the numbers move
Source at 0.5 m, receiver at 4 m, barrier 20 m from the source (30 m in the tall-barrier case), C₂ = 20, C₃ = 1, c = 343 m/s. The right-hand column is the change in the level at the receiver for an urban road spectrum, A-weighted: negative means that under the amendment the barrier screens more and the level drops.
| Geometry | z (m) | Dz 2024 at 500 Hz | Dz DAmd 1 at 500 Hz | Level at receiver |
|---|---|---|---|---|
| 4 m barrier, receiver at 50 m | +0.18 | 8.7 dB | 8.8 dB | −0.06 dB(A) |
| 8 m barrier, receiver at 500 m | +0.93 | 10.9 dB | 11.2 dB | −0.23 dB(A) |
| 2.5 m barrier, receiver at 300 m | +0.08 | 3.7 dB | 5.4 dB | −1.44 dB(A) |
| 2.5 m barrier, receiver at 1000 m | +0.09 | 0.4 dB | 4.8 dB | −4.37 dB(A) |
| obstacle grazing the line of sight, 300 m | −0.0005 | 1.3 dB | 4.8 dB | −4.15 dB(A) |
The short urban scenario — an ordinary barrier height with the receiver a few tens of metres away — moves by hundredths of a decibel: nobody working there has anything to redo. The two middle rows and the last one are the cases the 2024 edition declared it was correcting, and they are where the amendment gives back up to 4.4 dB per band.
## The case that grazes the line of sight
The last row of the table is the interesting one. An obstacle that brushes the line of sight without intercepting it has a slightly negative z. There:
- **1996 and DAmd 1** set Kmet = 1 for z ≤ 0, so Dz = 10·log₁₀(3 + something small and negative) ≈ **4.8 dB, at any distance**;
- **2024** applies Kmet all the way down to zmin, and with a distant receiver Kmet collapses (0.003 at 500 Hz over 300 m in our case): Dz tends to zero.
Those 4.8 dB credited to an obstacle that does not screen are the shortcoming the second edition declared it was eliminating, and the same one Probst's 2020 investigation isolated when it proposed the change that went into the 2024 edition. Under the amendment they come back.
The 2024 *Foreword* listed among the introduced changes the modified specification of Dz and Kmet "to eliminate well known shortcomings with low barriers and large source-to-receiver distances". The draft deletes that item from the list. It states no reason why the change should be withdrawn, and no technical contribution on that point appears to have been published yet: the DIS ballot opened recently, and anyone with an argument makes it through their national standards body — the ISO record itself says so.
## Two errata that weigh more than the diffraction change
The amendment is not only Clause 7.4. It fixes two annex formulas that, as published, do not produce the result they describe.
**Formula (A.1), basal area of a wood.** The published form feeds a diameter in centimetres into an expression that returns square metres per hectare; the draft divides the diameter by 100. For a wood of 25 cm stems at 400 stems per hectare:
| | G as computed |
|---|---|
| formula (A.1) as published | 1.96 × 10⁵ m²/ha |
| formula (A.1) with the correction | 19.63 m²/ha |
A factor of 10⁴. The corrected value sits in the order of magnitude forestry literature associates with mature woodland; the other one is larger than the hectare itself.
**Formula (B.1), radiation angle.** The published form contains arctan(dp/(zS − zR)): the denominator goes to zero when source and receiver sit at the same height, and the expression switches branch when the receiver is higher than the source. The draft writes 90° + arctan((zS − zR)/dp) − arcsin(d/2r), which is defined in every case:
| zS | zR | dp | (B.1) as published | (B.1) corrected |
|---|---|---|---|---|
| 100 m | 5 m | 300 m | 105.8° | 105.8° |
| 100 m | 5 m | 1000 m | 89.7° | 89.7° |
| 30 m | 30 m | 500 m | undefined | 87.1° |
| 20 m | 40 m | 400 m | 264.8° | 84.8° |
With the source lower than the receiver, the published form returns an angle outside the physical range. Anyone who implemented Annex B by reading the formula literally has a broken case in production, and it is not a corner case: it is the geometry of a stack seen from a receiver on higher ground.
The rest of the amendment is editorial: two Table B.1 values with a missing decimal separator (101 and −77 become 10.1 and −7.7), one source height in Table B.2, a mismatched symbol in (B.7), and Figure 15 redrawn around the angle bisector, with the paragraph describing it rewritten.
## The larger signal is in the ISO register
While the amendment is out for ballot, the catalogue also holds **ISO/AWI 9613-2, Edition 3**: new project approved, status *Under development*, stage 10.99 — "Drafting has started" — and the explicit note **"Will replace ISO 9613-2:2024"**.
So the second edition, published in January 2024, is being both amended and replaced within the same period. For anyone writing reports the consequence is the usual one, only sharper: the edition belongs in the document, because 1996, 2024, amended 2024 and the third edition do not give the same numbers. For anyone writing software the consequence is that **hard-coding one edition's formulas into the engine is the choice that ages worst**: the edition has to stay explicit and selectable, and every result has to say which formulation produced it.
## What the literature says
On the withdrawal itself there is, so far, nothing published. On the 2024 rewrite there is, and it converges on one point:
- **Davis and Callahan** (NOISE-CON 2025) compared the two editions on ground-level sources typical of energy projects and conclude that the 2024 changes "result in significant reductions to calculated effectiveness of sound barriers", to the point that in some cases the difference is between a buildable project and one that cannot be mitigated at reasonable cost. That is the effect the table above measures from the other side: if the amendment passes, those reductions are undone.
- **Odeh, Khayyat and Kocsis** (Forum Acusticum / EuroNoise 2025) review what the 2024 edition introduced — meteorological corrections, ground effect, source characterisation, reflections from cylindrical surfaces — assessing it with commercial noise mapping software.
- **Schaal and Zollitsch** (Forum Acusticum / EuroNoise 2025) criticise instead what ISO 9613-2 still does **not** model: transmission through semi-transparent obstacles, and more generally the limits of quality assurance when the scenario is more complex than the engineering method describes. That criticism is independent of the amendment, and neither the 2024 edition nor the draft addresses it.
- The investigation the 2024 change came from — **Probst, 2020**, an analysis of alternatives for a revised ISO 9613-2 — proposed integrating the long-range effect through Kmet in exactly the way the second edition adopted.
Environmental noise prediction software, for its part, responded to the second edition by parameterising: the formulation is exposed as an option — with or without the limit, current or previous edition — rather than swapped inside a version update. With an amendment out for ballot and a third edition in drafting, that choice explains itself.
## What changes in our tools: nothing, for now
The [barrier calculator](/en/tools/barrier-calculator/) and the [Maekawa explorer](/en/tools/maekawa/) stay as they are: ISO 9613-2:2024 as the default, a comparison mode on the previous formulation, Maekawa as a third option, and the method printed next to the result. An amendment out for ballot does not go into a calculation engine: it goes in once published, and when it does there will be one more entry in the selector, not a silent change of default.
One thing the draft does make more visible. Our comparison mode on the previous edition omits the direct distance `d` in Kmet, which is why it is documented as *legacy* rather than as "ISO 9613-2:1996": it exists to reproduce the tool's own historical results, not to represent the standard. If DAmd 1 is published, that canonical Kmet becomes the formulation in force again, and at that point it will be implemented as such — declared, alongside the others.
Method behind the numbers on this page: the three formulations were implemented separately and compared on the same geometries, with C₂ = 20, C₃ = 1 (single edge), c = 343 m/s and an A-weighted urban road spectrum. The differences measure the gap between formulas, not the agreement of any one of them with a field measurement. ISO records consulted on 3 September 2026.
The scope of the models, the assumptions and the declared limits are in the [noise-barrier-calc](/en/wiki/noise-barrier-calc/) and [Maekawa](/en/wiki/maekawa/) wiki pages; the edition in force for every standard cited is in the [regulatory references](/en/wiki/riferimenti-normativi-acustica/).
## References
- **[ISO 9613-2:2024/DAmd 1](https://www.iso.org/standard/93178.html)** — *Amendment 1*, ISO/TC 43/SC 1, ICS 17.140.01, 3 pages. Status *Under development*, stage *DIS ballot initiated: 12 weeks* [40.20]. The record states how to take part in the draft: through your national standards body.
- **[ISO/AWI 9613-2](https://www.iso.org/standard/92873.html)** — third edition, new project, stage 10.99: "Drafting has started", "Will replace ISO 9613-2:2024".
- **[ISO 9613-2:2024](https://www.iso.org/standard/74047.html)** — the edition in force: Clause 7.4 diffraction, Annex A propagation through foliage and industrial sites, Annex B source geometry.
- **Davis, S., Callahan, R.** (2025) — *A comparison of sound barrier effectiveness when using ISO 9613-2:1996 vs. ISO 9613-2:2024*, NOISE-CON 2025, Stowe (VT), pp. 787-793, [DOI 10.3397/NC_2025_0135](https://doi.org/10.3397/NC_2025_0135).
- **Odeh, O., Khayyat, A., Kocsis, D.** (2025) — *Key Updates in ISO 9613-2:2024 — What's New in Outdoor Noise Prediction Standards?*, Forum Acusticum / EuroNoise 2025, Málaga, pp. 1259-1264, [DOI 10.61782/fa.2025.0511](https://doi.org/10.61782/fa.2025.0511).
- **Schaal, J., Zollitsch, D.** (2025) — *Quality Assurance and Modeling Limitations in Noise Emission Control Using ISO 9613-2 and ISO 17534*, Forum Acusticum / EuroNoise 2025, Málaga, pp. 1293-1299, [DOI 10.61782/fa.2025.0718](https://doi.org/10.61782/fa.2025.0718).
- **Probst, W.** (2020) — *Investigation of ISO 9613-2: alternatives with improvements for a revised ISO 9613-2*. The analysis behind the Dz and Kmet change the second edition adopted.
- **Maekawa, Z.** (1968) — *Noise reduction by screens*, Applied Acoustics 1(3), 157-173, [DOI 10.1016/0003-682X(68)90020-0](https://doi.org/10.1016/0003-682X(68)90020-0).
- **Odeh, O., Kocsis, D.** (2026) — *From models to reality: how CNOSSOS-EU and ISO 9613-2 perform against measured road traffic noise*, Noise Mapping 13, [DOI 10.1515/noise-2025-0023](https://doi.org/10.1515/noise-2025-0023): the two methods compared against road traffic noise measurements.