# Errata in EN 1793

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

Canonical: https://docs.jmrp.io/phonometry/reference/errata/en-1793/
Language: en
Alternate: https://docs.jmrp.io/phonometry/es/reference/errata/en-1793/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 EN 1793: 6 defects in the printed text, with the evidence and what phonometry does about each.

---

These are the 6 entries of the [errata registry](https://docs.jmrp.io/phonometry/reference/errata/) on EN 1793-5:2016, 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.

## EN 1793-5:2016, Formula (1) against Formula (4) (a gain factor that multiplies)

- **Location:** 5.2, Formula (1) and the list under it, printed folios 15 and
  16 (PDF pages 17 and 18); 5.5.1, Formula (4) and the paragraph above it,
  printed folio 25 (PDF page 27) of BS EN 1793-5:2016.
- **The print:** Formula (1) closes the ratio of the reflected to the incident
  band energy with the product
  $\cdot\, C_{geo,k} \cdot C_{dir,k}(\Delta f_j) \cdot C_{gain,k}(\Delta f_g)$,
  and the list under it calls $C_{gain,k}$ "the correction factor to account
  for a change in the amplification settings of the loudspeaker and in the
  sensitivity settings of the individual microphones when changing the
  measurement configuration from free field to in front of the sample under
  test or vice versa". 5.5.1 then defines it as "the ratio between the
  spectrum obtained from windowing the incident component from the impulse
  response in front of the sample under test and the spectrum of the incident
  component obtained from windowing the free field impulse response":
  $C_{gain,k}(\Delta f_g) = \int_{\Delta f_g} |F[h_{i,k,D}(t)\, w_{i,k}(t)]|^2\, df \,/\, \int_{\Delta f_g} |F[h_{i,k,FF}(t)\, w_{i,k}(t)]|^2\, df$.
- **The problem:** the two formulas put the factor on the same side. In
  Formula (1) the reflected component comes from the record in front of the
  device and the incident component from the free-field record, so a gain
  that differs by an amplitude factor $g$ between the two configurations
  already multiplies the ratio by $g^2$. Formula (4) measures exactly that
  $g^2$. Multiplying by it, as Formula (1) prints, leaves $g^4$ in the index;
  dividing by it removes the change, which is what the list under Formula (1)
  says the factor is for. At the 20 % that 5.5.1 still accepts
  ($C_{gain,k} = 1.2$), the product as printed reports the index 44 % high,
  where leaving the change uncorrected would have reported it 20 % high.
  The two other factors of Formula (1) are the right way round:
  $C_{geo,k} = (d_{r,k}/d_{i,k})^2$ puts back the spreading of the longer
  reflected path, and $C_{dir,k}$ divides the loudspeaker's output towards
  the microphone by its output towards the specular point.
- **Evidence:** both formulas as printed, read on the page: Formula (1) on
  PDF page 17 (printed p. 15), its list on PDF page 18 (printed p. 16), and
  Formula (4) with the paragraph that defines it on PDF page 27 (printed
  p. 25) of BS EN 1793-5:2016. The consequence is a recomputation: synthetic
  records of a reflection of amplitude 0.5, with the in-front records scaled
  by 1.08, return 0.25 in every band when the factor divides and
  $0.25 \times 1.08^4$ when it multiplies.
- **Library behaviour:**
  [`reflection_index`](https://github.com/jmrplens/phonometry/blob/main/src/phonometry/environment/propagation/barrier_reflection.py)
  divides by $C_{gain,k}$ and says so in its docstring. The conformance row
  "EN 1793-5:2016 Formulas (1) and (4) (a gain change between configurations)"
  and a test on the scaled records pin it.
- **Status:** unreported.

## EN 1793-5:2016, Table B.2 (seven expanded uncertainties a hundredth below 1.96 times the printed $s_R$)

- **Location:** Annex B.5 and Table B.2, printed folios 54 and 55 (PDF
  pages 56 and 57) of BS EN 1793-5:2016; Table A.1 on printed folio 47 (PDF
  page 49).
- **The print:** B.5 says the uncertainty is "estimated using the values for
  the standard deviation of reproducibility given in Table A.1", with "a
  coverage factor of 1,96" and "the maximal ones (last column of Table A.1)".
  Table B.2 prints, against $s_R$ (High) of 0,32, 0,18, 0,12, 0,14, 0,15,
  0,15 and 0,17 in the bands of 100 Hz, 125 Hz, 500 Hz, 630 Hz, 800 Hz,
  1 250 Hz and 3 150 Hz, an expanded uncertainty $U$ of 0,62, 0,34, 0,23,
  0,26, 0,28, 0,28 and 0,32.
- **The problem:** 1.96 times the printed $s_R$ gives 0.63, 0.35, 0.24,
  0.27, 0.29, 0.29 and 0.33 in those seven bands, a hundredth above each
  printed cell. The other eleven bands reproduce to the hundredth, and so does
  the row of $DL_{RI}$ (1.96 times 0.81 dB is 1.59 dB, and 7.68 dB plus or minus
  that is the [6.09; 9.27] dB of B.5). No single coverage factor reproduces
  the seven: the 100 Hz cell needs one between 1.922 and 1.953, the 125 Hz cell
  one between 1.861 and 1.917.
- **The likely mechanism:** every one of the seven is consistent with 1.96
  times an unrounded standard deviation that rounds to the printed one (0.316
  prints as 0.32, and 1.96 times 0.316 is 0.619). The print does not show
  which values the example multiplied; if it was those, they are values the
  reader of Table A.1 does not have.
- **Consequence:** a hundredth of the index, which is the resolution the
  report states it to (5.11). A laboratory following B.5 with Table A.1 as
  printed obtains the higher values.
- **Evidence:** the two tables as printed, read on the page: Table B.2 on PDF
  page 57 (printed p. 55) and Table A.1 on PDF page 49 (printed p. 47) of
  BS EN 1793-5:2016. The rest is a recomputation from the printed cells.
- **Library behaviour:**
  [`environment.reflection_index_expanded_uncertainty`](https://github.com/jmrplens/phonometry/blob/main/src/phonometry/environment/propagation/barrier_reflection.py)
  multiplies the values of Table A.1 as printed. The conformance row
  "EN 1793-5:2016 Table B.2 (expanded uncertainty per band)" checks the
  eleven cells that follow from them, and a test pins the other seven a
  hundredth above the print.
- **Status:** unreported.

## EN 1793-5:2016, 5.5.7 against 5.8 (a low frequency limit read against the bands two ways)

- **Location:** 5.5.7, fifth paragraph, printed folio 33 (PDF page 35), and
  5.8, the paragraph on other purposes, the third after the list under
  Formula (12), printed folio 43 (PDF page 45) of BS EN 1793-5:2016.
- **The print:** 5.5.7: the qualification sample of 4 m by 4 m gives "a low
  frequency limit for the reflection index ... of about 160 Hz at
  microphone 2, 170 Hz at microphone 5 and 220 Hz at microphone 8 ..., i.e.
  using the two window lengths specified in 5.5.5 measurements are valid down
  to the 200 Hz one-third octave band". 5.8: "for a 3,5 m high barrier the
  single number rating of sound reflection is $DL_{RI}$ (400 - 5 000 Hz),
  because 300 Hz is the lowest frequency applicable to all microphones and
  this invalidates the measurements in the 315 one-third octave band and
  below".
- **The problem:** the first example keeps a band whose centre, 200 Hz, lies
  below the limit of one of its microphones, 220 Hz; the second drops a band
  whose centre, 315 Hz, lies above the limit, 300 Hz. No single rule relating a
  limit to a band gives both: against the centre, the 4 m sample loses its
  200 Hz band; against the lower band edge (178 Hz and 282 Hz), it loses it as
  well; against the upper band edge (224 Hz and 355 Hz), the 3.5 m barrier
  keeps its 315 Hz band.
- **Consequence:** the lowest reliable band $m$ of Formula (12) cannot be
  derived from a low frequency limit by the rule the clause illustrates,
  because the two illustrations disagree.
- **Evidence:** both sentences as printed, read on the page: PDF page 35
  (printed p. 33) and PDF page 45 (printed p. 43) of BS EN 1793-5:2016. The
  band edges are the base-ten one-third octave edges,
  $f_m \cdot 10^{\pm 1/20}$.
- **Library behaviour:**
  [`sound_reflection_rating`](https://github.com/jmrplens/phonometry/blob/main/src/phonometry/environment/propagation/noise_reducing_devices.py)
  and
  [`reflection_index`](https://github.com/jmrplens/phonometry/blob/main/src/phonometry/environment/propagation/barrier_reflection.py)
  take the lowest reliable band from the caller, 200 Hz by default, which is
  what 5.5.7 states for the qualification sample and what Annex B uses;
  [`reflection_low_frequency_limit`](https://github.com/jmrplens/phonometry/blob/main/src/phonometry/environment/propagation/barrier_reflection.py)
  reports the limit of each microphone and leaves the choice of band to the
  report, which states both: the low frequency limit by 5.11 n), and the
  lowest reliable band used in the single number by 5.11 q).
- **Status:** unreported.

## EN 1793-5:2016, Table 1 against Formula (7) (what $T_{W,BH}$ is the length of)

- **Location:** Table 1, the row of $T_{W,BH}$, printed folio 14 (PDF
  page 16); 5.5.5, NOTE 1 and Formula (7), printed folio 31 (PDF page 33) of
  BS EN 1793-5:2016.
- **The print:** Table 1: "$T_{W,BH}$ Length of the Blackman-Harris trailing
  edge of the Adrienne temporal window". NOTE 1: "A four-term full
  Blackman-Harris window of length $T_{W,BH}$ is:" Formula (7), with
  $0 \le t \le T_{W,BH}$.
- **The problem:** Formula (7) rises from nearly zero to 1 over the first half
  of $T_{W,BH}$ and falls back over the second, so read with the length of the
  trailing edge in it, it draws a whole bell over those 2.22 ms rather than
  the fall that 5.5.5 describes. The edges of the Adrienne window are halves
  of Formula (7), each over a full window twice its own length, which is the
  reading NOTE 1 gives and Table 1 does not.
- **Evidence:** the row and the note as printed, read on the page: PDF
  page 16 (printed p. 14) and PDF page 33 (printed p. 31) of
  BS EN 1793-5:2016.
- **Library behaviour:**
  [`adrienne_reflection_window`](https://github.com/jmrplens/phonometry/blob/main/src/phonometry/environment/propagation/barrier_reflection.py)
  takes each edge as a half of Formula (7) over twice the edge's length, which
  gives the window Figure 13 draws; the conformance row
  "EN 1793-5:2016 5.5.5 and 5.5.1 (the three Adrienne windows)" checks the
  three lengths of each window. No number changes with the reading of the row.
- **Status:** unreported.

## EN 1793-5:2016, 5.5.5 ("microphones 1 to 3, microphones 4 to 6 and microphones 1 to 9")

- **Location:** 5.5.5, last paragraph, printed folio 32 (PDF page 34) of
  BS EN 1793-5:2016; the test report template of Annex B.1, item (m), printed
  folio 49 (PDF page 51).
- **The print:** 5.5.5: "different window lengths for microphones 1 to 3,
  microphones 4 to 6 and microphones 1 to 9 may be used". Annex B.1 (m):
  "Adrienne temporal window (State length for each microphone height)",
  followed by "Length for microphones 1–3", "Length for microphones 4–6" and
  "Length for microphones 7–9".
- **The problem:** the three groups of 5.5.5 overlap, since microphones 1 to
  9 include the other six, and they leave the bottom row without a group of
  its own. The template of the same standard names the rows of the grid, 1 to
  3, 4 to 6 and 7 to 9, one window length per microphone height; the bottom
  row is the one the ground reflection reaches first (5.5.7), so it is the one
  a reduced sample most needs a window of its own for. "1 to 9" reads as a
  misprint for "7 to 9".
- **Evidence:** the paragraph and the template as printed, read on the page:
  PDF page 34 (printed p. 32) and PDF page 51 (printed p. 49) of
  BS EN 1793-5:2016.
- **Library behaviour:**
  [`reflection_index`](https://github.com/jmrplens/phonometry/blob/main/src/phonometry/environment/propagation/barrier_reflection.py)
  takes one window length per microphone, so the three rows of the template,
  or any other split, can be given. No number changes with the reading.
- **Status:** unreported.

## EN 1793-5:2016, Figure 15 Key (a distance from the loudspeaker named $d_M$, and a garbled item B)

- **Location:** 5.6.2.3, the Key of Figure 15, printed folio 38 (PDF page 40);
  Table 1, printed folio 13 (PDF page 15) of BS EN 1793-5:2016.
- **The print:** Key item 1: "Distance from the loudspeaker front panel to the
  reference plane $d_M$ [m]". Key item B: "reference position midway and B of
  the sample under test". Table 1: "$d_M$ Horizontal distance from the source
  and microphone reference plane to the measurement grid; it is equal to
  $d_M$ = 0,25 m" and "$d_S$ Horizontal distance from the front panel of the
  loudspeaker to the source and microphone reference plane; it is equal to:
  $d_S$ = 1,50 m".
- **The problem:** item 1 puts the words of one distance to the symbol of
  another. The distance from the loudspeaker front panel to the reference
  plane is $d_S$, 1.50 m; $d_M$ is the 0.25 m from the plane to the grid, and
  that is what the dimension marked 1 spans in Figure 15 (a), from the
  microphones A, B and C to the dashed reference plane, with no loudspeaker
  drawn. The drawing and the symbol agree with Table 1; the words do not.
  Item B names B inside its own definition and has lost words: the caption of
  Figure 15 (a) calls B the position "midway", between the most protruding
  part (A) and the less protruding one (C).
- **Consequence:** none numerically, since Table 1 and 3.14 fix $d_M$. A
  reader who takes the key at its word would read $d_M$ as the 1.50 m of the
  loudspeaker rather than the 0.25 m of the grid.
- **Evidence:** the key, the drawing and Table 1 as printed, read on the page:
  PDF page 40 (printed p. 38) and PDF page 15 (printed p. 13) of
  BS EN 1793-5:2016.
- **Library behaviour:** none required; no value is read from Figure 15.
  `REFLECTION_SOURCE_DISTANCE_M` and `REFLECTION_MICROPHONE_DISTANCE_M` in
  [`barrier_reflection`](https://github.com/jmrplens/phonometry/blob/main/src/phonometry/environment/propagation/barrier_reflection.py)
  carry $d_S$ = 1.50 m and $d_M$ = 0.25 m as Table 1 defines them.
- **Status:** unreported.
