# Occupational Noise Exposure (ISO 9612)

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

Canonical: https://docs.jmrp.io/phonometry/perception/hearing/occupational-exposure/
Language: en
Alternate: https://docs.jmrp.io/phonometry/es/perception/hearing/occupational-exposure/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: 2026-07-10
Updated: 2026-11-16T21:36:10Z
Standards: ISO 9612:2009; Directive 2003/10/EC
Verified: 3 conformance checks (ISO 9612:2009) https://docs.jmrp.io/phonometry/reference/conformance/#outdoor-propagation-occupational-exposure
License: MIT (https://github.com/jmrplens/phonometry/blob/main/LICENSE)
Summary: ISO 9612 daily noise exposure LEX,8h in Python: task-based, job-based and full-day strategies, the Annex C uncertainty and the one-sided 95 % upper limit.

---

A working day is rarely measured in one take: the daily exposure level a
regulation acts on has to be assembled from *samples* of a real shift, and
reported with an uncertainty a hygienist can defend. `lex_8h` (in
[Levels](https://docs.jmrp.io/phonometry/signals/levels/levels/)) turns *one* recording into a daily level. ISO 9612:2009,
the engineering method (accuracy grade 2), is the survey design *around* that
primitive: how to sample a real working day, how to combine the pieces, and how
to attach the normative uncertainty every occupational-hygiene report needs. The
`occupational_exposure` module adds the three **measurement strategies** and the
**Annex C** uncertainty budget on top of the energy-average machinery.

![Left: a worker wearing a personal sound exposure meter (IEC 61252), its microphone mounted about 0.04 m above the shoulder and at least 0.1 m from the entrance of the most-exposed ear canal, per ISO 9612 Clause 12.3. Right: the three measurement strategies drawn as timelines over an eight-hour working day, task-based (the day split into labelled tasks, at least three samples plus a duration each), job-based (five or more random samples spread over the homogeneous exposure group) and full-day (the whole shift measured at least three times), all feeding the LEX,8h and its Annex C uncertainty, chosen by work pattern from Table B.1](https://docs.jmrp.io/phonometry/media/diagram_dosimeter_iso9612.svg)

## How do I compute ISO 9612 daily noise exposure LEX,8h in Python?

Describe each activity as a `hearing.Task` with its samples and duration, then
call `hearing.task_based_exposure(tasks)`. For the Annex D welder's day below
it returns `lex_8h` = 84.3 dB with an `expanded_uncertainty` $U = 3.2$ dB, so
the one-sided 95 % `upper_limit` is 87.5 dB, because the default includes the
Annex C duration term $u_{1\mathrm{b}}$ built from each task's `duration_range_hours`. The §1
snippet passes `include_duration_uncertainty=False` to reproduce Annex D case a,
where the durations are treated as exact, and that is what drops $U$ to 2.7 dB
and the upper limit to 87.0 dB. Unstructured days use
`job_based_exposure()` or `full_day_exposure()` on random samples instead.

## 1. The three measurement strategies (Clauses 9-11)

The *task-based* strategy (Clause 9) splits the nominal day into tasks, takes
$I \ge 3$ samples per task, and energy-sums the task contributions

$$
L_{\mathrm{EX,8h},m} = L_{p,\mathrm{A,eq}T,m} + 10 \log_{10}(T_m/T_0), \qquad T_0 = 8\ \text{h},
$$

so a loud but short task contributes little. The *job-based* (Clause 10) and
*full-day* (Clause 11) strategies instead take random samples over a homogeneous
exposure group ($N \ge 5$ samples for the job-based strategy, at least three
whole shifts for the full-day one), energy-average them, and normalise that
average to the nominal eight-hour day with the effective working-day duration
$T_\mathrm{e}$ in the same $10\log_{10}(T_\mathrm{e}/T_0)$ term. The daily level is the same
either way; the strategies differ in how the **uncertainty** is built.

**Choosing a strategy.** ISO 9612 Table B.1 picks the strategy from the *work
pattern*, not from convenience, and the trade-off is coverage against effort.
The **task-based** strategy is the recommended default when the day
decomposes into a small number of well-defined tasks: because it measures each
task separately, it explains *where* the dose comes from (the per-task
breakdown above) and lets a short, loud task be sampled properly without
dragging out the whole survey, but it needs a reliable work analysis, and its
uncertainty grows if the task durations are themselves uncertain. The
**job-based** strategy suits a mobile worker with an unpredictable pattern or
a homogeneous group doing "the same job": random samples across the group
average over the variability instead of resolving it, which is robust when the
day cannot be cleanly cut into tasks but blind to which activity dominates.
The **full-day** strategy (a worn dosimeter capturing the entire shift,
repeated on several days) needs the least analysis and captures everything
including the unexpected, at the cost of the most wearer-days and the weakest
diagnostic value (one number per day, no breakdown, and false contributions
like knocks on the microphone are hard to spot). As a rule: resolve the dose
with task-based when you can, fall back to job-based for irregular work, and
use full-day when the pattern defies description or an independent whole-shift
check is wanted.

### 1.1 Taking the samples

Everything below this line happens before a single number reaches a `Task`, and
the three commonest ways of invalidating a survey all live here.

**How long each sample runs (Clause 9.3).** A task measurement shall last at
least 5 min; if the task itself is shorter than 5 min, the measurement lasts as
long as the task. It may be shortened only when the level is found to be
constant or repeatable, or when the task is a minor contributor to the total
exposure. Cyclic noise is sampled over **at least three well-defined cycles**,
still at least 5 min, and always over a whole number of cycles. A partial
cycle biases the task level by whatever part of the cycle it over-samples.
Randomly fluctuating noise is sampled long enough to be representative, which
is a judgement the report has to record.

**How many, and taken when (Clause 9.3).** At least three measurements per
task, and the standard recommends taking them at different times during the
task or on different workers in the group, so that the scatter feeding
$u_{1\mathrm{a}}$ is real variability rather than one lucky minute. The Annex D day in
the snippet below is exactly this rule in action: welding keeps its three
samples, while cutting/grinding has six, because the first three spanned more
than 3 dB. Note what that buys: Eq. C.6 divides the scatter by $I(I-1)$, so it
is the sample *count*, not the sample length, that shrinks the sampling term.

**Where the microphone goes.** Two configurations, and the guide's opening
diagram draws only the first:

- **Worn instrument (Clause 12.3).** On the top of the shoulder, at least
  0.1 m from the entrance of the external ear canal on the most exposed side,
  approximately 0.04 m above the shoulder, with microphone and cable fastened
  so that clothing cannot cover them or rub against them. The body screening
  and reflections this introduces are exactly what the $u_3 = 1.0$ dB term of
  the Annex C budget pays for; where the ear is close to a small source, NOTE 1
  suggests measuring on both sides of the head to find which ear really is the
  most exposed.
- **Sound level meter (Clause 12.4).** With the worker absent, the microphone
  goes at the location of the worker's head, preferably in the centre plane of
  the head, on a line with the eyes and with its axis parallel to the worker's
  line of sight; where the head position is not well defined, 1.55 m ± 0.075 m
  above the floor for a standing worker and 0.80 m ± 0.05 m above the middle of
  the seat plane for a seated one. The workstation average may instead be swept
  at constant speed along a path shaped like an infinity sign. With the worker
  present, the microphone is held between 0.1 m and 0.4 m from the ear-canal
  entrance on the most exposed side, and if the work makes 0.4 m impossible the
  standard recommends the worn instrument instead.

![Two panels of the ISO 9612 Clause 12.4 sound level meter geometry. Left, worker absent: a tripod-mounted microphone stands at the location of the worker's head in front of a machine, its capsule on the eye line with the axis parallel to the line of sight, dimensioned 1.55 m plus or minus 0.075 m above the floor for a standing worker, with an inset giving 0.80 m plus or minus 0.05 m above the middle of the seat plane for a seated one, and a plan view of the constant-speed sweep along an infinity-shaped path. Right, worker present: the same workstation with a hand-held meter held between 0.1 m and 0.4 m from the entrance of the most exposed external ear canal, carrying a windscreen of at least 60 mm diameter, with a note that a fixed position under-reads a hand-held tool close to the ear](https://docs.jmrp.io/phonometry/media/diagram_slm_workstation_iso9612.svg)

**The two checks that decide whether the data survive.** A field calibration
with adjustment before each series and at the start of each daily series, and a
field calibration *without* adjustment at the end of each; if the two readings
differ by more than **0.5 dB** at any frequency, the results of that series
shall be discarded (Clause 12.2). This is what the example fiche's "field
checks within 0.3 dB" is claiming. Then, before trusting the day, compare the
level-versus-time log against the observation notes: a dosimeter records a
knock or a rub on the microphone as a genuine high-level event (Clause 13.2),
and one such artefact can move the day's $L_\mathrm{EX,8h}$; any unexplained peak
shall be investigated and the measurement repeated if the result is
significantly affected, and with a non-logging instrument, repeated outright.

**And the rest of the rejection rules.** Avoid high airflow, and fit hand-held
meters with a windscreen of at least 60 mm diameter (Clause 13.3; a 60 mm
windscreen keeps flow up to about 10 m/s from significantly affecting an 80 dB
level). Observe the worker, or reconstruct the activities from interviews and
work records and report that you did (Clause 9.3). Radio, speech and alarm
noise count as relevant if the work analysis says they are part of normal
working conditions; excluding them is allowed but is a *reportable decision*,
not a silent edit, and non-typical behaviour during the measurement has to be
evaluated and, if it matters, re-measured (Clause 13.4).

```python
from phonometry import hearing

# ISO 9612 Annex D — a welder's day split into three tasks. Each task level is
# the energy average of its Lp,A,eqT samples; durations carry a measured range.
tasks = [
    hearing.Task(samples=(70.0,), duration_hours=1.5, label="planning/breaks"),
    hearing.Task(samples=(80.1, 82.2, 79.6), duration_hours=5.0,
         duration_range_hours=(4.0, 6.0), label="welding"),
    hearing.Task(samples=(86.5, 92.4, 89.3, 93.2, 87.8, 86.2), duration_hours=1.5,
         duration_range_hours=(1.0, 2.0), label="cutting/grinding"),
]
res = hearing.task_based_exposure(tasks, include_duration_uncertainty=False, warn=False)
print(f"LEX,8h = {res.lex_8h:.1f} dB   U = {res.expanded_uncertainty:.1f} dB")
# LEX,8h = 84.3 dB   U = 2.7 dB
print(f"one-sided 95 % upper limit LEX,8h + U = {res.upper_limit:.1f} dB")   # 87.0 dB
for t in res.tasks:
    print(f"  {t.label:<16} Lp,A,eqT = {t.lp_aeqt:5.1f}   contributes {t.lex_8h_contribution:5.1f} dB")
#   planning/breaks  Lp,A,eqT =  70.0   contributes  62.7 dB
#   welding          Lp,A,eqT =  80.8   contributes  78.7 dB
#   cutting/grinding Lp,A,eqT =  90.1   contributes  82.8 dB

# The same shift measured job-based (Annex E) and full-day (Annex F): both use
# the Eq C.9 / Table C.4 sampling budget with k = 1.65 (one-sided 95 %).
job = hearing.job_based_exposure([88.1, 86.1, 89.7, 86.5, 91.1, 86.7], effective_duration_hours=7.5)
full = hearing.full_day_exposure([88.0, 91.9, 87.6, 90.4, 89.0, 88.4], effective_duration_hours=9.25)
print(f"job      LEX,8h = {job.lex_8h:.1f} dB   U = {job.expanded_uncertainty:.1f} dB")
# job      LEX,8h = 88.2 dB   U = 3.8 dB
print(f"full-day LEX,8h = {full.lex_8h:.1f} dB   U = {full.expanded_uncertainty:.1f} dB")
# full-day LEX,8h = 90.1 dB   U = 3.4 dB

res.plot()   # the figure below: task contributions with LEX,8h and LEX,8h + U
```

**Reading the job and full-day numbers.** The task-based case gets its per-task
breakdown; the other two deserve the same walk-through, because in both of them
the answer is moved by a term that is not a measured level at all. $T_\mathrm{e}$ is the
**effective duration of the working day** established by the work analysis: the
time the worker is actually exposed, breaks away from the noise excluded,
breaks spent in it included. The six job samples energy-average to 88.4 dB, and
with $T_\mathrm{e} = 7.5$ h the normalisation $10\log_{10}(7.5/8)$ subtracts 0.3 dB, so
$L_\mathrm{EX,8h}$ = 88.2 dB, which is *below* the sampled average. The six full-day samples
energy-average to 89.5 dB, but that shift lasts 9.25 h, so the same term adds
$10\log_{10}(9.25/8)$ = 0.6 dB and $L_\mathrm{EX,8h}$ = 90.1 dB, which is higher than
four of the six individual samples. That is not an error: normalising a long
shift to a nominal eight-hour day is *supposed* to penalise it. Get $T_\mathrm{e}$ wrong
by a factor and the whole result moves by $10\log_{10}$ of it, which is why the
task-based strategy carries a separate duration-uncertainty term $u_{1\mathrm{b}}$ for
exactly this quantity.

![ISO 9612 Annex D task-based exposure: the three task LEX,8h contributions as bars, the energy-summed daily LEX,8h line and the one-sided 95 % upper limit LEX,8h + U band above it](https://docs.jmrp.io/phonometry/media/exposure_uncertainty.svg)

**Show the code for this figure**

```python
import matplotlib.pyplot as plt
from phonometry import hearing

# The ISO 9612 Annex D welder's day of the previous snippet.
tasks = [
    hearing.Task(samples=(70.0,), duration_hours=1.5, label="planning/breaks"),
    hearing.Task(samples=(80.1, 82.2, 79.6), duration_hours=5.0,
         duration_range_hours=(4.0, 6.0), label="welding"),
    hearing.Task(samples=(86.5, 92.4, 89.3, 93.2, 87.8, 86.2), duration_hours=1.5,
         duration_range_hours=(1.0, 2.0), label="cutting/grinding"),
]
res = hearing.task_based_exposure(tasks, include_duration_uncertainty=False, warn=False)

# One line: task contribution bars plus the LEX,8h and LEX,8h + U lines.
res.plot()
plt.show()
```

### 1.2 Planning the survey

ISO 9612 is a planning standard as much as a computation standard, and the two
questions a hygienist asks *before* going on site are each one call away. How
many hours must the survey cover for a group of this size? Table 1 fixes the
minimum cumulative measurement duration for a homogeneous exposure group as a
function of its size, and `min_cumulative_duration_hours` reads it:

```python
for n in (4, 10, 25, 50):
    print(n, hearing.min_cumulative_duration_hours(n))
# 4 5.0
# 10 7.5
# 25 12.5
# 50 17.0
```

Three regimes and a ceiling: 5 h up to five workers, then half an hour per
extra worker to fifteen, then a quarter of an hour per worker to forty, then a
flat 17 h. Beyond forty, the standard's advice is to split the group rather
than measure longer. Pass the planned figures as `n_workers=` and
`sample_duration_hours=` to `job_based_exposure` and the result raises the
advisory when the coverage falls short.

The second question is how many samples the sampling term needs.
`table_c4_contribution(n_samples, u1)` interpolates Table C.4, so the answer can
be chosen against a target instead of discovered afterwards:

```python
print([round(hearing.table_c4_contribution(n, 3.0), 1) for n in (3, 5, 10, 20, 30)])
# [11.5, 3.3, 1.5, 0.9, 0.7]
```

For a 3 dB sample scatter, the contribution $c_1 u_1$ collapses from 11.5 dB at
three samples to 3.3 dB at five and 1.5 dB at ten, and then flattens: going
from five to ten samples buys 1.8 dB, going from twenty to thirty buys 0.2 dB.
That is where the Clause 10.4 advisory threshold of 3.5 dB sits. A job-based
survey of a group with 3 dB of scatter crosses it below five samples. This is
the quantitative form of the "more samples, not a better meter" rule of §2.
Note that the $N = 3$ and $N = 4$ rows of Table C.4 apply to full-day
measurements only (Table C.4, NOTE 1).

## 2. The Annex C uncertainty budget

Two subtleties are worth spelling out. First, the coverage factor is
$k = 1.65$ for a **one-sided** 95 % interval (Clause 14), because a hygienist
cares only about the *upper* bound: `res.upper_limit` = $L_\mathrm{EX,8h} + U$ is the
value 95 % of measurements fall below, the number compared against an action
limit. Second, the task and job methods weight the *same* spread of samples
differently. The task sampling uncertainty $u_{1\mathrm{a}}$ (Eq. C.6) divides the summed
squared deviations by $I(I-1)$ (the standard error of the mean, smaller by a
factor $\sqrt{I}$), whereas the job/full-day sampling uncertainty $u_1$ (Eq. C.12)
is the plain sample standard deviation with denominator $N-1$, whose contribution
$c_1 u_1$ is then read from **Table C.4** as a function of $(N, u_1)$. The same
raw scatter therefore inflates the job estimate more, which is the standard's
built-in penalty for coarser, fewer samples. (The printed job $L_\mathrm{EX,8h}$ is
$88.2$ dB where Annex E reports $88.1$: the standard rounds the effective-day
level to $88.4$ before the duration normalisation; the library keeps it
unrounded.)

**What dominates the budget.** Annex C combines four sources in quadrature
(Table C.1): the **sampling** uncertainty ($u_{1\mathrm{a}}$/$u_1$), the **duration**
uncertainty ($u_{1\mathrm{b}}$, task-based only), the **instrument** ($u_2$, Table C.5)
and the **microphone position** ($u_3$, Clause C.6). The last two are small
and roughly fixed ($u_2 = 0.7$ dB for a class 1 sound level meter, 1.5 dB for
a class 2 meter or a personal exposimeter, and $u_3 = 1.0$ dB by default),
which makes them a **floor** no survey design gets under: with a personal
exposimeter, $U = 1.65\sqrt{1.5^2 + 1.0^2} = 3.0$ dB even if every sample came
out identical. Everything above that floor is the sampling and duration terms,
which are also the only two the surveyor controls.

![Two panels. Left: the ISO 9612 Annex D day as stacked variance bars, one per task plus a whole-day bar, split into the sampling, duration, instrument and position terms; the planning task contributes nothing visible, welding a small stack, and cutting/grinding 3.5 of the 3.85 dB squared total, annotated as 1.5 hours of the day and 91 percent of the variance, with the whole-day bar labelled u = 1.96 dB giving U = 3.2 dB. Right: the expanded uncertainty against the number of samples for a fixed 3 dB sample scatter, the task-based curve falling from 4.1 dB at three samples toward the floor, the job-based Table C.4 curve starting far higher at 19 dB and converging by about twenty samples, with dotted horizontal floors at 3.0 dB for a personal exposimeter or class 2 meter and 2.0 dB for a class 1 meter](https://docs.jmrp.io/phonometry/media/exposure_budget.svg)

*Two things the figure says that the formulae do not.* First, every term is
weighted by $c_{1\mathrm{a}}$, the task's share of the daily energy, so the budget
follows the dose rather than the clock: cutting and grinding is 1.5 h of the
eight and carries **91 %** of the variance, welding's five hours carry 9 %, and
the quiet planning task carries nothing measurable. Second, "the sampling term
dominates" is a rule of thumb, not an identity. With the default personal
exposimeter the task sampling term overtakes the instrument term only once the
sample standard deviation exceeds $1.5\sqrt{I}$ (about 3.7 dB at six samples),
and in this Annex D day it does not: cutting and grinding scatters by 3.0 dB
over six samples, so $u_{1\mathrm{a}} = 1.2$ dB against $u_2 = 1.5$ dB. The
**duration** of that short, loud task weighs more than its sampling, because
$c_{1\mathrm{b}} = 4.34\,c_{1\mathrm{a}}/\bar{T}_m$ grows as the task shortens: a
$\pm 0.5$ h uncertainty on a 1.5 h task costs more than the sampling term and
almost as much as the instrument term, the largest segment of the day.

**Show the code for this figure**

```python
import matplotlib.pyplot as plt
import numpy as np

# `hearing` and `res` (the Annex D day, with duration uncertainty) come from
# the snippets above; here it is rebuilt with the duration term included.
res_d = hearing.task_based_exposure(tasks, warn=False)
for t in res_d.tasks:
    print(f"task {t.label:<16} sampling {(t.c1a * t.u1a) ** 2:5.3f}  "
          f"duration {(t.c1b * t.u1b) ** 2:5.3f}  "
          f"instrument {(t.c1a * t.u2) ** 2:5.3f}  "
          f"position {(t.c1a * t.u3) ** 2:5.3f}")
    # task planning/breaks  sampling 0.000  duration 0.000  instrument 0.000  position 0.000
    # task welding          sampling 0.049  duration 0.059  instrument 0.175  position 0.078
    # task cutting/grinding sampling 0.757  duration 1.070  instrument 1.148  position 0.510

counts = np.arange(3, 31)
u2, u3, scatter = 1.5, 1.0, 3.0
expanded = lambda term: 1.65 * np.sqrt(np.asarray(term) ** 2 + u2**2 + u3**2)
fig, ax = plt.subplots()
ax.plot(counts, expanded(scatter / np.sqrt(counts)), label="task-based")
ax.plot(counts, expanded([hearing.table_c4_contribution(int(n), scatter)
                          for n in counts]), "--", label="job-based")
ax.axhline(expanded(0.0))            # the instrument + position floor
ax.legend()
plt.show()
```

The practical consequence still holds, and the right panel gives it a shape: a
quadrature budget is set by its largest term, so tightening the instrument
grade buys nothing while the sampling term sits above the floor, and the
productive move is more samples, with returns that fall off fast. At a 3 dB
scatter the task-based $U$ goes 4.1 dB at three samples, 3.4 dB at ten, 3.1 dB
at thirty, while the job-based strategy pays a far steeper penalty for a small
$N$ (19.2 dB at three, 6.2 dB at five, 3.9 dB at ten) because Table C.4 does
not simply divide by $\sqrt{N}$. That gap is the standard's built-in penalty
for coarser, fewer samples, and it is what the Clause 9.3 / 10.4 advisories
push against. Because peak $L_{p,\mathrm{Cpeak}}$ carries no Annex C sampling model
(Table C.5, Note 1), it is reported without an uncertainty, not with a zero one.

When a task's samples span **3 dB or more** (Clause 9.3), or the job contribution
$c_1 u_1$ exceeds 3.5 dB (Clause 10.4), or too few workers are covered
(Table 1 cumulative-duration), the result sets `sampling_advisory=True` and, with
`warn=True`, emits an `OccupationalExposureWarning` recommending more
measurements. The
three Annex D/E/F worked examples above are reproduced to the standard's printed
precision (Annex E's final rounding is disclosed above), and the theory is
derived on the [Theory](https://docs.jmrp.io/phonometry/reference/theory/environment-transport/) page.

### `task_based_exposure()` / `job_based_exposure()` / `full_day_exposure()` parameters

| Parameter | Applies to | Type | Units | Range / default | Notes |
| :--- | :--- | :--- | :--- | :--- | :--- |
| `tasks` | task | list of `Task` | — | $\ge 1$ | Each `Task` has `samples`, `duration_hours`, optional `duration_range_hours`/`duration_observations_hours`, `label`, `instrument` |
| `samples` | job / full-day | sequence | dB | $\ge 2$ ($\ge 5$ / $\ge 3$ advised) | Random $L_{p,\mathrm{A,eq}T}$ samples |
| `effective_duration_hours` | job / full-day | float | h | $> 0$ | Effective working-day duration $T_\mathrm{e}$ |
| `instrument` | all | str | — | `'class1'`, `'class2'`, `'personal_exposimeter'` (default) | Selects $u_2$ (Table C.5) |
| `u3` | all | float | dB | default `1.0` | Microphone-position uncertainty (Clause C.6) |
| `include_duration_uncertainty` | task | bool | — | default `True` | `False` omits the $(c_{1\mathrm{b}}u_{1\mathrm{b}})^2$ term (Annex D case a) |
| `n_workers` / `sample_duration_hours` | job | int / float | — / h | default `None` | Table 1 cumulative-duration check |
| `warn` | all | bool | — | default `True` | Emit `OccupationalExposureWarning` for the sampling advisories |

All three return an `ExposureResult` with `lex_8h`, `combined_standard_uncertainty`
$u$, `expanded_uncertainty` $U = 1.65\ u$, `upper_limit` = $L_\mathrm{EX,8h} + U$,
`sampling_advisory`, and (task-based) the per-task `tasks` breakdown; the
result's `.plot()` draws the per-task contribution bars with the $L_\mathrm{EX,8h}$
and upper-limit lines (task-based results only, since the other strategies
carry no per-task breakdown).

## 3. The measurement report (Clause 15)

An exposure determination ends as a *document*: ISO 9612 Clause 15 lists what
the measurement report shall state, from the strategy that was applied and the
work analysis down to the requirement that the noise exposure level and the
measurement uncertainty be reported *as separate values*, each rounded to one
decimal place. `ExposureResult.report()` writes that report as a one-page PDF
fiche laid out like a prevention-service measurement sheet: the standard-basis
line naming the applied strategy, a header grid (company, worker(s)/job,
workplace, and the Clause 15 c instrumentation and calibration traceability),
the work analysis (the per-task table of durations, $L_{p,\mathrm{A,eq}T,m}$ levels
and $L_{\mathrm{EX,8h},m}$ contributions for a task-based result with its contribution
chart, or the sampling summary with the Formula C.9 budget for a job-based or
full-day result) and the boxed $L_\mathrm{EX,8h}$ with $U$, $k$ and the one-sided
95 % upper limit.

Because the number exists to be compared with the law, the fiche then assesses
the result against **Directive 2003/10/EC** (Article 3): the lower and upper
exposure action values (80 and 85 dB(A)) and the exposure limit value
(87 dB(A)), each marked exceeded / not exceeded on the value exactly as
displayed, with a PASS/FAIL verdict against the limit value. A printed note
records that the limit value applies to the *effective* exposure, with the
attenuation of the worn hearing protectors taken into account, which the
measured $L_\mathrm{EX,8h}$ does not include. `verbose=True` adds the per-task
Annex C uncertainty columns ($u_{1\mathrm{a}}$, $u_{1\mathrm{b}}$, $u_2$), and `language="es"`
renders the Spanish fiche (the vocabulary of the Spanish transposition,
RD 286/2006: *nivel de exposición diario equivalente*, *valor límite de
exposición*, with comma decimals). Rendering needs reportlab and, for the contribution chart the fiche of a
task-based result embeds, matplotlib (`pip install "phonometry[report,plot]"`).

```python
from phonometry import ReportMetadata, hearing

# The Annex D welders' day from section 1, with duration uncertainty.
res = hearing.task_based_exposure(tasks, warn=False)

res.report(
    "lex8h.pdf",
    metadata=ReportMetadata(
        client="Example fabrication works",
        specimen="Welders (homogeneous exposure group, 4 workers)",
        test_room="Steel assembly hall, line 2",
        instrumentation="Personal sound exposure meter (IEC 61252), s/n 0042",
        calibration="Calibrator IEC 60942 class 1; field checks within 0.3 dB",
        test_date="2026-07-20",
        laboratory="Phonometry reference example",
        report_id="EXAMPLE-9612",
    ),
)   # LEX,8h = 84.3 dB, U = 3.2 dB -> lower action value exceeded, limit PASS
```

The example fiche is regenerated with `make reports` and kept rendered in the
repository; click the preview to open the PDF.

**Report:** [ISO 9612 occupational noise exposure example report (PDF)](https://docs.jmrp.io/phonometry/media/iso9612_exposure_example.pdf): One-page ISO 9612 occupational noise-exposure fiche: a header with the company, the welders exposure group, the workplace and the instrumentation and calibration traceability, the task-based work-analysis table (planning/breaks, welding, cutting/grinding with durations, sample counts, task levels and LEX,8h contributions closed by the nominal-day totals row), the per-task contribution chart, the boxed LEX,8h = 84.3 dB with U = 3.2 dB, k = 1.65 and the 87.5 dB upper limit, and the Directive 2003/10/EC assessment table where the 80 dB(A) lower action value is exceeded, the 85 dB(A) upper action value and the 87 dB(A) limit value are not, ending in a PASS verdict.

Occupational noise-exposure fiche (ExposureResult.report), the ISO 9612 Annex D task-based day with the Clause 15 work analysis and the Directive 2003/10/EC assessment.

## What this guide covers

- **Covered.** The three measurement strategies of ISO 9612:2009 Clauses 9-11,
  `task_based_exposure`, `job_based_exposure` and `full_day_exposure`, with the
  Clause 9.3 sampling rules and the Clause 12.3 / 12.4 microphone geometry
  stated here as the procedure that produces their inputs; the Annex C
  uncertainty budget (Formulae C.6, C.9 and C.12 with Tables C.4 and C.5, and
  $u_3$ from Clause C.6); the one-sided coverage factor $k = 1.65$ of Clause 14;
  the Clause 9.3 / 10.4 / Table 1 advisories, reachable before the survey
  through `min_cumulative_duration_hours` and `table_c4_contribution`; and
  the Clause 15 measurement report through `ExposureResult.report()`, with the
  Directive 2003/10/EC assessment on top of it. The three worked examples of
  Annexes D, E and F are reproduced.

- **Not covered.** The work analysis of Clause 7 and the microphone placement of Clause 12.3 /
  12.4 are procedures, not calculations: they stay with the surveyor, and this
  guide documents them so the numbers fed to a `Task` mean something. Table B.1
  is reproduced as guidance and the caller picks the strategy; nothing here
  applies it. Hearing-protector attenuation (the ISO 4869 series) is not
  applied, so the $L_\mathrm{EX,8h}$ computed here (and the limit-value comparison the
  fiche prints on it) is the **unprotected** exposure, which is also what the
  ISO 1999 hearing-loss model wants. And Annex C offers no uncertainty model for
  $L_{p,\mathrm{Cpeak}}$ (Table C.5, Note 1), which is why the peak is reported without
  one rather than with a zero one.

## See also

- [Levels](https://docs.jmrp.io/phonometry/signals/levels/levels/): the `lex_8h` / `sound_exposure` dose primitives
  (IEC 61252) and the $L_{p,\mathrm{Cpeak}}$ these strategies report alongside.
- [Measurement uncertainty](https://docs.jmrp.io/phonometry/signals/metrology/gum-uncertainty/): the GUM machinery behind
  combined and expanded uncertainties.
- [Theory](https://docs.jmrp.io/phonometry/reference/theory/environment-transport/): the derivation of the strategy formulas and the
  Annex C budget.
- API reference: [`hearing.occupational_exposure`](https://docs.jmrp.io/phonometry/reference/api/hearing/occupational-exposure/).

## References

- **European Parliament and Council. (2003).** [*Directive 2003/10/EC on the minimum health and safety requirements regarding the exposure of workers to the risks arising from physical agents (noise)*](https://eur-lex.europa.eu/eli/dir/2003/10/oj/eng) (Directive 2003/10/EC). Official Journal of the European Union.
  The EU exposure action values (80/85 dB) and limit value (87 dB) that the LEX,8h and its upper limit are assessed against.
- **International Organization for Standardization. (2009).** [*Acoustics — Determination of occupational noise exposure — Engineering method*](https://www.iso.org/standard/41718.html) (ISO 9612:2009).
  The implemented method: the task-based (Clause 9), job-based (Clause 10) and full-day (Clause 11) strategies, the microphone placement of Clause 12.3, the Annex C uncertainty budget (Formulae C.6, C.9 and C.12, Tables C.4/C.5) and the one-sided coverage factor k = 1.65 (Clause 14), validated against the worked examples of Annexes D, E and F; the Clause 15 measurement report is written by ExposureResult.report().
- **National Institute for Occupational Safety and Health. (1998).** [*Criteria for a recommended standard: Occupational noise exposure — Revised criteria 1998*](https://www.cdc.gov/niosh/docs/98-126/pdfs/98-126.pdf) (DHHS/NIOSH Publication No. 98-126). https://doi.org/10.26616/NIOSHPUB98126
  The freely available criteria document (the linked PDF is the free download) behind the 85 dB(A) recommended exposure limit and the hearing-conservation rationale for measuring the daily dose.
