# Errata in DIN 45672

> One page of the phonometry documentation, as Markdown. Index: https://docs.jmrp.io/phonometry/llms.txt

Canonical: https://docs.jmrp.io/phonometry/reference/errata/din-45672/
Language: en
Alternate: https://docs.jmrp.io/phonometry/es/reference/errata/din-45672/index.md
Version: phonometry 4.0.0
Install: pip install "phonometry[full]"
Author: José Manuel Requena Plens (https://orcid.org/0000-0003-1250-6212)
Published: 2026-10-28
Updated: 2026-11-16T21:36:14Z
License: MIT (https://github.com/jmrplens/phonometry/blob/main/LICENSE)
Summary: Errata in DIN 45672: 9 defects in the printed text, with the evidence and what phonometry does about each.

---

These are the 9 entries of the [errata registry](https://docs.jmrp.io/phonometry/reference/errata/) on DIN 45672-1:2009-12, DIN 45672-2:1995-07 and E DIN 45672-3:2023-02, in the order the registry lists them. The registry's introduction says how each claim is checked against the printed page and what the status of an entry means.

## DIN 45672-1:2009-12, Clause 4.5.1, Formulae (1) and (5) (the compression-wave speed of a thin rod, and a radicand short of a factor 2)

- **Location:** Clause 4.5.1, Formulae (1) to (4) on printed page 7 and
  Formula (5) on printed page 8 (PDF pages 7 and 8 of the copy read here, which
  prints its folios without an offset).
- **The print:** Formula (1) gives the compression-wave speed as
  $\mathrm{v}_\mathrm{p} = \sqrt{E/\rho} = \sqrt{G(1-\nu)/(\rho(1-2\nu))}$, Formula (3) gives
  Poisson's ratio as $\nu = (\mathrm{v}_\mathrm{p}^2 - 2 \mathrm{v}_\mathrm{s}^2)/(2(\mathrm{v}_\mathrm{p}^2 - \mathrm{v}_\mathrm{s}^2))$, and
  Formula (5) gives the two moduli as $G = \mathrm{v}_\mathrm{s}^2 \rho$ and $E = \mathrm{v}_\mathrm{p}^2 \rho$.
- **The problem:** the three cannot all hold. In the unbounded continuum the
  clause says it is describing, the compression wave travels at
  $\mathrm{v}_\mathrm{p} = \sqrt{2G(1-\nu)/(\rho(1-2\nu))}$, and Formula (3) is exactly the
  inversion of that together with $\mathrm{v}_\mathrm{s} = \sqrt{G/\rho}$ of Formula (2). The
  second radical of Formula (1) is short of the factor 2, which makes it
  $\sqrt{2}$ too slow at every Poisson's ratio. The first, $\sqrt{E/\rho}$, is
  the speed of a longitudinal wave in a thin rod: with $E = 2G(1+\nu)$ it gives
  $\mathrm{v}_\mathrm{p}^2/\mathrm{v}_\mathrm{s}^2 = 2(1+\nu)$, against $2(1-\nu)/(1-2\nu)$ in the continuum. The
  two expressions Formula (1) sets equal agree with each other only at
  $\nu = (\sqrt{17}-1)/8 \approx 0.39$. Formula (5) is the rod speed solved
  for $E$, so from a measured $\mathrm{v}_\mathrm{p}$ it returns the P-wave modulus
  $M = 2G(1-\nu)/(1-2\nu)$ rather than $E$, which overstates $E$ by 35 % at
  $\nu = 0.3$ and by a factor of 3.8 at $\nu = 0.45$, the range of a
  saturated soil.
- **Evidence:** the three formulas against each other on the same two pages,
  and against the P-wave speed of an isotropic elastic continuum. Verified on
  PDF page 7 (printed p. 7) for Formulae (1) to (4) and on PDF page 8 (printed
  p. 8) for Formula (5) of DIN 45672-1:2009-12: the radicals, the factor
  $(1-\nu)$ over $(1-2\nu)$ and the absence of the factor 2 are all legible on
  the page.
- **Consequence for the standard's own tables:** none; the clause prints no
  worked values. What it changes is a modulus read from two measured speeds.
- **Library behaviour:**
  [`compression_wave_speed`](https://github.com/jmrplens/phonometry/blob/main/src/phonometry/vibration/immission/ground.py)
  and
  [`youngs_modulus_from_wave_speeds`](https://github.com/jmrplens/phonometry/blob/main/src/phonometry/vibration/immission/ground.py)
  implement the continuum relations Formula (3) is the inverse of, and the
  tests hold both printed forms to the factors above, so an edit back to the
  print fails.
- **Status:** unreported.

## DIN 45672-2:1995-07, Clause 4 (the start-up of the running r.m.s., quoted in mean square)

- **Location:** Clause 4, last paragraph of printed page 3, and Figure 3 on
  printed page 4 (PDF pages 3 and 4 of the copy read here, which prints its
  folios without an offset).
- **The print:** the running r.m.s. "erst nach einer Dauer von 2τ mit einer
  Unsicherheit von 14 % und nach einer Dauer von 4τ mit einer Unsicherheit von
  2 % zur Verfügung steht, wobei der Mittelwert des gleitenden Effektivwertes
  für ein harmonisches Signal zugrunde gelegt wurde (siehe Bild 3)": it is only
  available after $2\tau$ to within 14 % and after $4\tau$ to within 2 %, taking
  the mean of the running r.m.s. of a harmonic signal.
- **The problem:** 14 % and 2 % are the shortfalls of the running mean square,
  $\mathrm{e}^{-2} = 13.5$ % and $\mathrm{e}^{-4} = 1.8$ %, and not of the running r.m.s. the
  sentence names. Formula (1) started from rest gives a mean square that grows
  as $(1 - \mathrm{e}^{-t/\tau})$ of its final value once the ripple is averaged out, so
  the r.m.s. grows as the square root of that, and it is short by
  $1 - \sqrt{1 - \mathrm{e}^{-2}} = 7.0$ % after $2\tau$ and by
  $1 - \sqrt{1 - \mathrm{e}^{-4}} = 0.9$ % after $4\tau$: about half the printed
  figures.
- **Evidence:** Formula (1) on the same page and Figure 3 on the next, which
  draws $\tilde v_\mathrm{F}/\hat v$ of an 8 Hz and a 20 Hz sine against
  time in units of $\tau$ with the mean marked at 0.707. At $2\tau$ both
  curves oscillate around 0.66, which is 93 % of 0.707, and at $4\tau$
  around 0.70. Verified on
  PDF page 3 (printed p. 3) and PDF page 4 (printed p. 4) of
  DIN 45672-2:1995-07.
- **Consequence for the standard's own tables:** none. The advice the sentence
  gives, to start the averaging before the train arrives, stands either way;
  what is overstated is the size of the error a late start costs.
- **Library behaviour:**
  [`running_velocity_rms`](https://github.com/jmrplens/phonometry/blob/main/src/phonometry/vibration/immission/railway.py)
  says which quantity each figure belongs to, and the conformance report
  reproduces the printed 14 % and 2 % from the mean square of Formula (1)
  started from rest, which is the reading that matches them.
- **Status:** unreported.

## E DIN 45672-3:2023-02, Annex C, C.3 (two assessment severities that weight by the factor once where Formula (11) squares it)

- **Location:** C.3 on printed page 34 (PDF page 34 of the copy read here,
  which prints its folios without an offset) and printed page 35, against
  Formula (11) on printed page 23 and Annex E on printed page 37.
- **The print:** with 200 passages by day and 20 by night, the factor
  $\alpha$ = 0,7 of a surface tram and $KB_\mathrm{FTm,Zug}$ = 0,4, Formula (11)
  gives "$KB_\mathrm{FTr,Zug,Tag}$ = 0,11" and "$KB_\mathrm{FTr,Zug,Nacht}$ = 0,05", and C.4
  finds the day exceeded, "0,11 > $A_\mathrm{r,Tag}$ = 0,1".
- **The problem:** Formula (11) with those inputs and $N_\mathrm{r}$ = 1920 by day and
  960 by night is $0.7 \cdot 0.4 \cdot \sqrt{200/1920} = 0.090$ and
  $0.7 \cdot 0.4 \cdot \sqrt{20/960} = 0.040$. The printed values are
  what the same inputs give with $\alpha$ under the root once instead of
  squared, $0.4 \sqrt{0.7 \cdot 200/1920} = 0.108$ and
  $0.4 \sqrt{0.7 \cdot 20/960} = 0.048$: the example weights the energy
  of the category by the factor where Formula (11) weights its amplitude.
  With the 0.090 of the formula the verdict of C.4 on the day turns round:
  0.09 is below the $A_\mathrm{r}$ of 0.1 and the requirement is met.
- **Evidence:** the inputs on printed page 34, the two results at the top
  of printed page 35 and the assessment below them, the formula and its
  $N_\mathrm{r}$ on printed page 23 and the factors on printed page 37. Verified on PDF page 34 (printed p. 34), PDF
  page 35 (printed p. 35), PDF page 23 (printed p. 23) and PDF page 37
  (printed p. 37) of E DIN 45672-3:2023-02.
- **Consequence for the standard's own tables:** the example's daytime
  conclusion, that mitigation is to be planned, does not follow from its
  numbers.
- **Library behaviour:**
  [`train_assessment_severity`](https://github.com/jmrplens/phonometry/blob/main/src/phonometry/vibration/immission/train_categories.py)
  squares the factor as Formula (11) prints it, and the conformance rows of
  C.3 compare with the 0.090 and 0.040 it gives for the printed inputs,
  saying that the print reads 0.11 and 0.05.
- **Status:** not reported; the document is a draft under comment.

## E DIN 45672-3:2023-02, Annex C, Table C.1 (a sum level formed without the weighting Clause 7.1 prescribes)

- **Location:** Table C.1 and the results of C.3 on printed page 34 (PDF
  page 34), against Formulae (8) and (9) on printed pages 21 and 22.
- **The print:** the table closes with the row
  "Schwinggeschwindigkeitssummenpegel der betrachteten Zugkategorie
  ($L_{v,\mathrm{Zug}}$):" and 78,1 dB in its last column, and C.3 feeds it to
  Formula (9),
  $KB_\mathrm{FTm,Zug} = c_{\mathrm{T}1} v_0 10^{L/20} = 0{,}4$, then 0,6 by Formula (10) and
  "$v_{\max}$ = 1,81 mm/s" by Formula (12).
- **The problem:** Clause 7.1 a) first adds the KB weighting of Table 2 to
  each band (Formula (8)) and sums the bands from 4 Hz to 80 Hz. The energy
  sum of the 19 printed $L_v$ without any weighting is 78.08 dB, which is
  the 78.1 printed; the weighted sum over 4 Hz to 80 Hz is 77.7 dB, and the
  chain from it is 0.385, 0.577 and 1.73 mm/s. The printed 1.81 mm/s is
  $3 \cdot 1.5 \cdot 5 \cdot 10^{-5} \cdot 10^{78.1/20} = 1.808$, so the
  example carried the unweighted sum through.
- **Evidence:** the sum and the three results on printed page 34 and the
  formulas on printed pages 21 and 22. Verified on PDF page 34 (printed
  p. 34), PDF page 21 (printed p. 21) and PDF page 22 (printed p. 22) of
  E DIN 45672-3:2023-02.
- **Consequence for the standard's own tables:** the results of C.3 are
  4 % high against the standard's own procedure; the verdicts of C.4 are
  the same either way, 0.6 rounds both.
- **Library behaviour:**
  [`predict_train_category`](https://github.com/jmrplens/phonometry/blob/main/src/phonometry/vibration/immission/railway_prediction.py)
  weights the bands as 7.1 a) says before it sums them; the conformance rows
  hold the printed chain from 78.1 dB and the test holds the weighted one.
- **Status:** not reported; the document is a draft under comment.

## E DIN 45672-3:2023-02, Annex C, C.2 and Table C.1 (a floor transfer that comes from no table of Annex A, and a figure cited by the wrong number)

- **Location:** C.2 on printed page 33 (PDF page 33) and the column
  $\Delta L_{v,\mathrm{DF}}$ of Table C.1 on printed page 34, against Table A.5 on
  printed page 27 and Figure 4 on printed page 15.
- **The print:** C.2 says the foundation-to-floor transfer of the example
  is "die Übertragungen vom Fundament zur Geschossdecke mit einer
  Deckeneigenfrequenz von 20 Hz aus Bild 3", and Table C.1 prints the
  column 1,9; 2,3; 3,1; 3,5; 5,0; 6,9; 11,5; 17,3; 10,0; 5,4; 1,9; 1,5;
  −0,8; −2,3; −3,8; −5,4; −6,5; −8,1; −9,6 dB from 4 Hz to 250 Hz.
- **The problem:** Figure 3 is the transfer from the ground into a
  foundation at ground level; the foundation-to-floor transfer of a concrete
  floor is Figure 4 and Table A.5. Read at the ratios of the bands to 20 Hz,
  Table A.5's mean gives 1.60; 2.06; 2.52; 3.26; 4.19; 6.35; 9.94; 17.26;
  9.85; 4.41; 3.27; 3.25; 1.42; 3.89; 2.83 dB up to 100 Hz and nothing
  above a ratio of 5. Only the peak agrees; neither the mean nor either
  deviation of Table A.5, nor the 20 Hz column of Table A.1, gives the
  printed column, and the table has no values for the last four bands the
  column fills.
- **Evidence:** the sentence on printed page 33, the column on printed
  page 34 and Table A.5 on printed pages 27 and 28. Verified on PDF page 33
  (printed p. 33), PDF page 34 (printed p. 34), PDF page 27 (printed p. 27)
  and PDF page 28 (printed p. 28) of E DIN 45672-3:2023-02.
- **Consequence for the standard's own tables:** the example cannot be
  reproduced from the standard's own tables; its transfer column is an
  input.
- **Library behaviour:** the conformance rows of Table C.1 take the printed
  column as an input of Formula (1) and hold the sum;
  [`foundation_to_floor_level_difference`](https://github.com/jmrplens/phonometry/blob/main/src/phonometry/vibration/immission/railway_prediction.py)
  reads Table A.5 as printed.
- **Status:** not reported; the document is a draft under comment.

## E DIN 45672-3:2023-02, Annex A, Table A.6 (a lower deviation printed above the mean it deviates from)

- **Location:** Table A.6 on printed page 28 (PDF page 28), and Figure 5 on
  printed page 16.
- **The print:** at the ratios 0,20, 3,10, 4,00 and 5,00 the column
  "Standardabweichung nach unten (E–u)" reads 2,37; 5,26; 5,42 and 7,63 dB
  against a "Mittelwert (E–m)" of 2,19; 3,55; 3,24 and 3,01 dB, and at 5,00
  the "Standardabweichung nach oben (E–o)" is 5,40 dB, below the lower
  one.
- **The problem:** a deviation downward from a mean cannot lie above it, and
  the upper deviation cannot lie below the lower. Figure 5 draws the same
  crossing, its curve 1 ending above curves 2 and 3, so the figure was made
  from the same data; whether two columns were swapped at the tail or the
  statistics are wrong cannot be told from the page. Table A.5, the concrete
  floor, keeps its order in every row.
- **Evidence:** the four rows on printed page 28 and the tail of the curves
  on printed page 16. Verified on PDF page 28 (printed p. 28) and PDF page
  16 (printed p. 16) of E DIN 45672-3:2023-02.
- **Consequence for the standard's own tables:** a reader who takes the
  lower deviation as the safe side of a timber floor's transfer is above
  the mean at those ratios.
- **Library behaviour:**
  [`FOUNDATION_TO_FLOOR_LEVEL_DIFFERENCES`](https://github.com/jmrplens/phonometry/blob/main/src/phonometry/vibration/immission/railway_prediction.py)
  carries the table as printed, its docstring says where the order fails,
  and no order between the three statistics is enforced.
- **Status:** not reported; the document is a draft under comment.

## E DIN 45672-3:2023-02, Figures 5, 6 and 7 (legends that name the wrong table and the wrong quantity)

- **Location:** the legend of Figure 5 on printed page 16 (PDF page 16) and
  the axis legends of Figures 6 and 7 on printed pages 17 and 18.
- **The print:** Figure 5, the foundation-to-floor transfer of timber
  floors, labels its curves 1, 2 and 3 "Übertragung Fundament → Erdgeschoss
  und Obergeschosse bei Holzbalkendecken" with "(D–u)", "(D–m)" and
  "(D–o)"; Figures 6
  and 7, the ground-to-floor transfer, label their vertical axis
  "Pegeldifferenz $\Delta L_{v,\mathrm{DF}}(f_\mathrm{Tn})$ in dB".
- **The problem:** the table Figure 5 draws, Table A.6, names its columns
  E–u, E–m and E–o; D–u, D–m and D–o are the columns of Table A.5, the
  concrete floor of Figure 4. And Figures 6 and 7 draw $\Delta L_{v,\mathrm{DB}}$,
  the ground-to-floor difference of Tables A.1 and A.2, as their captions
  say; $\Delta L_{v,\mathrm{DF}}$ is the foundation-to-floor difference of Figures 4
  and 5.
- **Evidence:** the legends on printed pages 16, 17 and 18 and the column
  headings on printed pages 28, 24 and 25. Verified on PDF page 16 (printed
  p. 16), PDF page 17 (printed p. 17), PDF page 18 (printed p. 18), PDF page
  24 (printed p. 24), PDF page 25 (printed p. 25) and PDF page 28 (printed
  p. 28) of E DIN 45672-3:2023-02.
- **Consequence for the standard's own tables:** none; the captions and the
  tables are right.
- **Library behaviour:** none to take; the tables are what is implemented.
- **Status:** not reported; the document is a draft under comment.

## E DIN 45672-3:2023-02, 5.4.4 (an annex called normative where it is printed informative)

- **Location:** 5.4.4 on printed page 16 (PDF page 16 of the copy read here,
  which prints its folios without an offset), against the heading of
  Annex A on printed page 24.
- **The print:** "Im normativen Anhang A sind die Werte in Tabellenform für
  alle relevanten Deckeneigenfrequenzen zusammengefasst."; the annex is
  headed "Anhang A (informativ)".
- **The problem:** the six tables the prediction is made from are either
  part of the requirements or an information, and the two pages say one
  each.
- **Evidence:** the sentence on printed page 16 and the heading on printed
  page 24. Verified on PDF page 16 (printed p. 16) and PDF page 24 (printed
  p. 24) of E DIN 45672-3:2023-02.
- **Consequence for the standard's own tables:** none to their values.
- **Library behaviour:** the tables are implemented as printed, whatever
  their status.
- **Status:** not reported; the document is a draft under comment.

## E DIN 45672-3:2023-02, Formula (11) (the assessment sum printed without the rule that zeroes a quiet category)

- **Location:** Formula (11) and its symbols on printed page 23 (PDF page
  23), against Formula (6) of E DIN 4150-2:2023-08 on its printed pages 19
  and 20.
- **The print:** "$KB_\mathrm{FTr} = \sqrt{\sum_{\mathrm{Zug}=1}^{N_\mathrm{Z}}
  \tfrac{n_\mathrm{Zug}}{N_\mathrm{r}} (\alpha_\mathrm{Zug} \cdot
  KB_\mathrm{FTm,Zug})^2}$", introduced by "Berechnung der
  Beurteilungs-Schwingstärke ($KB_\mathrm{FTr}$) für den jeweiligen
  Beurteilungszeitraum entsprechend DIN 4150-2", with $N_\mathrm{r}$,
  $N_\mathrm{Z}$, $n_\mathrm{Zug}$, $KB_\mathrm{FTm,Zug}$ and
  $\alpha_\mathrm{Zug}$ "nach informativem Anhang E" listed under it and
  nothing else.
- **The problem:** the sum, its symbols and the factors of Annex E are
  Formula (6) and Table 2 of the draft of DIN 4150-2, which prints under
  its formula that a category whose $KB_\mathrm{FTm,Zug}$ is at or below 0.1 enters
  as zero. The sentence is not reproduced, so a predicted category at or
  below 0.1 counts here and not in the assessment the formula says it
  performs.
- **Evidence:** the formula and its symbol list on printed page 23, and
  the sentence under Formula (6) on printed page 20 of the other draft.
  Verified on PDF page 23 (printed p. 23) of E DIN 45672-3:2023-02 and PDF
  page 20 (printed p. 20) of E DIN 4150-2:2023-08.
- **Consequence for the standard's own tables:** none; the category of
  Annex C is at 0.4.
- **Library behaviour:** the chain of
  [`predict_train_category`](https://github.com/jmrplens/phonometry/blob/main/src/phonometry/vibration/immission/railway_prediction.py)
  ends in
  [`train_assessment_severity`](https://github.com/jmrplens/phonometry/blob/main/src/phonometry/vibration/immission/train_categories.py),
  which applies the rule of the assessment it stands for.
- **Status:** not reported; the document is a draft under comment.
