# Errata in ISO/TS 12913

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

Canonical: https://docs.jmrp.io/phonometry/reference/errata/iso-12913/
Language: en
Alternate: https://docs.jmrp.io/phonometry/es/reference/errata/iso-12913/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 ISO/TS 12913: 7 defects in the printed text, with the evidence and what phonometry does about each.

---

These are the 7 entries of the [errata registry](https://docs.jmrp.io/phonometry/reference/errata/) on ISO/TS 12913-3:2019 and ISO/TS 12913-2:2018, 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.

## ISO/TS 12913-3:2019, Annex A.3 (the perceived affective quality called part 3)

- **Location:** Annex A (informative), A.3, the paragraph that introduces
  Formulas (A.1) and (A.2).
- **The print:** "The results from part 3 (see A.1) are further processed to
  derive the values on two dimensions (pleasantness and eventfulness) for each
  site."
- **The problem:** the eight attributes the two formulas read (annoying, calm,
  chaotic, eventful, monotonous, pleasant, uneventful, vibrant) are the
  perceived affective quality, which is part 2 of the Method A questionnaire
  everywhere else the documents name it: in Table A.1 on the page before
  ("2 (perceived affective quality)"), in the A.2 paragraph that assigns its
  scale values 5 to 1 ("questionnaire part 2 (see Figure C.4 ...)"), in the
  title of A.3 itself ("based on perceived affective quality responses"), and
  in ISO/TS 12913-2:2018, C.3.1.3 and Figure C.4, "Questionnaire part 2:
  Perceived affective quality". Part 3 is the single overall rating of
  Figure C.5, "Overall, how would you describe the present surrounding sound
  environment?", which has no attributes and cannot feed either formula. The
  cross-reference "(see A.1)" does not help: A.1 is the general clause and
  names no part. The sentence should read "part 2 (see A.2 and Table A.1)".
- **Evidence:** the sentence on PDF page 11 (printed p. 5), Table A.1 and the
  A.2 paragraphs on PDF page 10 (printed p. 4), both of ISO/TS 12913-3:2019
  (first edition, 2019-12); C.3.1.3 on PDF page 21 (printed p. 15) and
  Figures C.4 and C.5 on PDF page 22 (printed p. 16) of ISO/TS 12913-2:2018
  (first edition).
- **Library behaviour:** `pleasantness_eventfulness` applies Formulas (A.1)
  and (A.2) to the eight attributes of part 2, the only reading under which
  they can be evaluated
  ([`soundscape.py`](https://github.com/jmrplens/phonometry/blob/main/src/phonometry/environment/assessment/soundscape.py)).
  No change was required. The 2025 edition of ISO/TS 12913-3 revises Annex A
  and has not been checked for this entry.
- **Status:** unreported.

## ISO/TS 12913-3:2019, Formula (A.3) (a stray factor 1)

- **Location:** Annex A (informative), A.4, Formula (A.3), Spearman's rank
  correlation coefficient for untied ranks.
- **The print:**
  $r_\mathrm{spearman} = 1 - 1\,\dfrac{6\cdot\sum_{i=1}^{n} d_i^2}{n\cdot(n^2 - 1)}$,
  with a "1" standing between the minus sign and the fraction.
- **The problem:** the coefficient for untied ranks is
  $1 - 6\sum d_i^2 / \left[n(n^2 - 1)\right]$, which is what the page gives if
  the stray "1" is read as a factor of one. Read the way a mixed number is
  written, $1\,\tfrac{a}{b} = 1 + \tfrac{a}{b}$, it would give
  $r = -6\sum d_i^2 / \left[n(n^2 - 1)\right]$, which is zero for identical
  rankings instead of one. Formula (A.4) on the same page reduces to the
  usual coefficient when there are no ties, so the intended form is not in
  doubt; the "1" is a typesetting remnant.
- **Evidence:** Formula (A.3) on PDF page 12 (printed p. 6) of
  ISO/TS 12913-3:2019 (first edition, 2019-12).
- **Library behaviour:** `spearman_rank_correlation` evaluates
  $1 - 6\sum d_i^2 / \left[n(n^2 - 1)\right]$ without ties and Formula (A.4)
  with them; the conformance rows hold the first to Pearson's coefficient of
  the ranks and the second to `scipy.stats.spearmanr`
  ([`soundscape.py`](https://github.com/jmrplens/phonometry/blob/main/src/phonometry/environment/assessment/soundscape.py)).
  No change was required.
- **Status:** unreported.

## ISO/TS 12913-3:2019, Formula (A.4) (the where-list of the tie counts)

- **Location:** Annex A (informative), A.4, the where-list under Formula
  (A.4), Spearman's rank correlation coefficient for tied ranks.
- **The print:** "$t_j$ is the number of in $t_j$ tied ranks of the variable
  $x$; $u_j$ is the number of in $u_j$ tied ranks of the variable $y$;
  $k(x)$ and $k(y)$ are the numbers of tied ranks of the variables $x$ and
  $y$", under
  $T = \sum_{j=1}^{k(x)} (t_j^3 - t_j)/12$ and
  $U = \sum_{j=1}^{k(y)} (u_j^3 - u_j)/12$.
- **The problem:** the first two definitions are not sentences ("the number
  of in $t_j$ tied ranks") and define $t_j$ by itself. The sums need $t_j$
  to be the number of values sharing the $j$-th tied rank of $x$ ($u_j$ the
  same for $y$), and $k(x)$, $k(y)$ to be the number of such groups of ties
  in each variable, which "the numbers of tied ranks" does not say. The
  intended reading is the usual tie correction of Spearman's coefficient,
  which is what makes (A.4) Pearson's coefficient of the average ranks.
- **Evidence:** the where-list on PDF page 13 (printed p. 7), under Formula
  (A.4) on PDF page 12 (printed p. 6), of ISO/TS 12913-3:2019 (first
  edition, 2019-12).
- **Library behaviour:** `spearman_rank_correlation` sums
  $(t_j^3 - t_j)/12$ over the groups of equal values of each variable, and a
  conformance row holds Formula (A.4) so read to `scipy.stats.spearmanr` on
  93 real answers with heavy ties
  ([`soundscape.py`](https://github.com/jmrplens/phonometry/blob/main/src/phonometry/environment/assessment/soundscape.py)).
  No change was required.
- **Status:** unreported.

## ISO/TS 12913-3:2019, Formula (B.2) ($x_I$ for $x_i$)

- **Location:** Annex B (informative), B.3, the where-list under Formula
  (B.2), the covariance of Pearson's correlation coefficient.
- **The print:** "$\bar{x}$ is the arithmetic mean value of the array $x_I$;"
  with a capital $I$, followed by "$\bar{y}$ is the arithmetic mean value of
  the array $y_i$;".
- **The problem:** the index is the lower-case $i$ of the sum in (B.2),
  $\sum_{i=1}^{n} (x_i - \bar{x})(y_i - \bar{y})/n$, and of the line for
  $\bar{y}$ just under it; $x_I$ names no array of the annex.
- **Evidence:** the where-list on PDF page 15 (printed p. 9) of
  ISO/TS 12913-3:2019 (first edition, 2019-12).
- **Library behaviour:** `pearson_correlation` takes $\bar{x}$ as the mean of
  the $x_i$ ([`soundscape.py`](https://github.com/jmrplens/phonometry/blob/main/src/phonometry/environment/assessment/soundscape.py)).
  No change was required.
- **Status:** unreported.

## ISO/TS 12913-2:2018, A.3 f), NOTE (exponent 3 for a cube root)

- **Location:** Annex A (normative), A.3 f), the NOTE on the root mean cubed
  loudness $N_\mathrm{rmc}$.
- **The print:** "The root mean cubed loudness (cubic mean), Nrmc, is
  computed by determining the mean of all loudness values raised to the power
  of 3 with a subsequent application of the exponent 3 as shown in the
  following formula:
  $N_\mathrm{rmc} = \sqrt[3]{\frac{1}{n}\sum_{i=1}^{n} N_i^3}$".
- **The problem:** the text and the formula under it disagree. The formula
  takes the cube root of the mean of the cubes, a subsequent exponent of
  $1/3$, which is what a cubic mean is and what returns a loudness in sone;
  the text says the subsequent exponent is 3, which would give the mean cube
  raised to the third power, $\left(\frac{1}{n}\sum N_i^3\right)^3$, in
  sone to the ninth. The sentence should read "a subsequent application of
  the exponent 1/3". ISO 532-1:2017, 6.4, NOTE, describes the energy mean of
  the loudness level in the same shape and gets it right: a power of about
  3.322 and then a power law "with the exponent lg(2)", its inverse.
- **Evidence:** the NOTE and its formula on PDF page 14 (printed p. 8) of
  ISO/TS 12913-2:2018 (first edition, 2018-08); ISO 532-1:2017, 6.4, on PDF
  page 22 (printed p. 16).
- **Library behaviour:** `binaural_indicators` computes $N_\mathrm{rmc}$ of
  each ear by the formula, the cube root of the mean of the cubes of the
  loudness over time, and a conformance row holds it to that formula
  ([`soundscape_binaural.py`](https://github.com/jmrplens/phonometry/blob/main/src/phonometry/environment/assessment/soundscape_binaural.py)).
  No change was required.
- **Status:** unreported.

## ISO/TS 12913-2:2018, C.3.2.3 against Figure C.7 (three scales in the text, four in the figure)

- **Location:** Annex C (informative), C.3.2.3, "Soundwalk data collection
  part 1: Assessment of the sound environment", and Figure C.7 below it.
- **The print:** the text reads "The participants should assess a site on
  three different five-point unipolar continuous-category scales with
  additional verbal labelling ranging from "not at all" to "extremely"." The
  figure prints four scales: "How loud is it here?", "How unpleasant is it
  here?" and "How appropriate is the sound to the surrounding?", labelled from
  "not at all" to "extremely", and "How often would you like to visit this
  place again?", labelled "never", "rarely", "sometimes", "often",
  "very often".
- **The problem:** the text and the figure it introduces disagree on the
  number of scales and on their labels. Either the fourth scale belongs to
  Method B, and the text should say four and name its second set of labels,
  or it does not, and the figure should not print it. ISO/TS 12913-3:2019 B.2
  and Table B.1 speak of "the five-point unipolar continuous-category scales"
  without a number and do not settle it.
- **Evidence:** C.3.2.3 and Figure C.7 on PDF page 24 (printed p. 18) of
  ISO/TS 12913-2:2018 (first edition); B.2 and Table B.1 on PDF page 14
  (printed p. 8) of ISO/TS 12913-3:2019.
- **Library behaviour:** `METHOD_B_SCALES` holds the four scales of the
  figure, as printed, and `method_b_summary` takes a table of three or four
  of them, so a study that used either reading is summarised under its own
  questions ([`soundscape.py`](https://github.com/jmrplens/phonometry/blob/main/src/phonometry/environment/assessment/soundscape.py)).
- **Status:** unreported.

## ISO/TS 12913-2:2018, Figures C.2 to C.4 ("extend" for "extent", "reponse" for "response")

- **Location:** Annex C (informative), the questionnaire of Method A: the
  questions of Figures C.2, C.3 and C.4 and the instruction line under each.
- **The print:** "To what extend do you presently hear the following four
  types of sounds?" (Figure C.2), "To what extend do you presently hear the
  following three types of sounds?" (Figure C.3), "For each of the 8 scales
  below, to what extend do you agree or disagree that the present surrounding
  sound environment is..." (Figure C.4); and "Please tick off one reponse
  alternative per type of sound" (Figures C.2 and C.3), "Please tick off one
  reponse alternative per scale" (Figure C.4).
- **The problem:** "extend" is a verb; the question asks "to what extent",
  which Figure C.6 of the same annex spells correctly ("Overall, to what
  extent is the present surrounding sound environment appropriate to the
  present place?"). "reponse" is a misspelling of "response". The figures are
  a questionnaire meant to be put in front of participants as printed, so the
  slips reach the field unless the study corrects them.
- **Evidence:** Figures C.2 and C.3 on PDF page 21 (printed p. 15), Figures
  C.4 and C.6 on PDF page 22 (printed p. 16), all of ISO/TS 12913-2:2018
  (first edition).
- **Library behaviour:** `METHOD_A_SCALES` and `METHOD_A_ALTERNATIVE_PART_1`
  transcribe the questions and instructions as printed, misspellings
  included, so the table can be compared with the page; the docstring of
  `QuestionnaireScale` says a study printing its own questionnaire from it
  should correct them
  ([`soundscape.py`](https://github.com/jmrplens/phonometry/blob/main/src/phonometry/environment/assessment/soundscape.py)).
- **Status:** unreported.
