# Airflow Resistance

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

Canonical: https://docs.jmrp.io/phonometry/materials/absorbers/airflow-resistance/
Language: en
Alternate: https://docs.jmrp.io/phonometry/es/materials/absorbers/airflow-resistance/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: 2026-07-28
Updated: 2026-11-08T02:07:59Z
Standards: ISO 9053-1:2018; ISO 9053-2:2020
Verified: 2 conformance checks (ISO 9053-2:2020) https://docs.jmrp.io/phonometry/reference/conformance/#materials-absorption-airflow-impedance
License: MIT (https://github.com/jmrplens/phonometry/blob/main/LICENSE)
Summary: Airflow resistance, specific airflow resistance and resistivity of porous materials by the static (ISO 9053-1) and alternating (ISO 9053-2) methods.

---

Push air slowly through a porous absorber and it pushes back. That viscous
drag, exerted by the pore walls on the air threading through them, is the
single most informative number a porous material has: it sets how much sound
the material dissipates at low frequency, it is the first input of every
equivalent-fluid model, and it is measured with nothing more exotic than a
pump and a manometer. ISO 9053 standardises the measurement twice over: the
**static method** of Part 1 drives a steady laminar flow through the specimen
and reads the pressure drop, and the **alternating method** of Part 2 replaces
the steady flow with a 2 Hz piston so that specimens too leaky or too delicate
for a stable static reading can be measured acoustically. This guide covers
the three quantities and their units, both methods, the accredited-style test
fiche, and what the measured resistivity feeds afterwards.

## 1. The three quantities (ISO 9053)

The airflow resistance quantifies how strongly a porous material opposes a steady
or slowly-oscillating flow. Both parts share the same three quantities and units
(ISO 9053-1:2018, Clause 3):

$$
R = \frac{\Delta p}{q_v}\ \left[\text{Pa·s/m}^3\right], \qquad
R_\mathrm{s} = R\,A\ \left[\text{Pa·s/m}\right], \qquad
\sigma = \frac{R_\mathrm{s}}{d}\ \left[\text{Pa·s/m}^2\right],
$$

with $\Delta p$ the pressure difference across the specimen, $q_v$ the volumetric
flow, $A$ the cross-section and $d$ the thickness. Note the specific airflow
resistance $R_\mathrm{s}$ is in **Pa·s/m** (not Pa·s/m²); the airflow resistivity $\sigma$
is $R_\mathrm{s}$ per metre of thickness. Two helpers walk that chain from what an
instrument actually shows: `linear_airflow_velocity(q_v, area)` is the
Clause 3.4 definition $u = q_v/A$ (a rig reads a volumetric flow off a
rotameter, never a velocity), and
`specific_airflow_resistance(resistance, area)` is Clause 3.2's
$R_\mathrm{s} = R\,A$; it also accepts the other route,
`specific_airflow_resistance(pressure_drop_pa=…, velocity=…)`, for a bench
that logs $\Delta p$ and $u$ and never forms $R$ at all.

**Which way the flow must go.** $A$ is defined as the area *perpendicular to
the direction of flow*, and that direction is a choice, not a formality: a
fibrous board whose fibres lie in the plane resists through-thickness flow
quite differently from in-plane flow, so $\sigma$ is a directional property of
any material with an oriented microstructure. ISO 9053-2 Clause 7.1 makes it
explicit. A non-homogeneous specimen is measured in one or more chosen
orientations, and the orientation goes in the report (ISO 9053-1 Clause 9 (g)
asks for the same thing). Drive the specimen in the direction the sound will
travel through it in service, which for a wall-mounted layer means through the
thickness. A resistivity measured edge-on will fit neither the layer nor the
models of section 4, and an unnoticed orientation mismatch is one of the
standard reasons a measured impedance-tube spectrum refuses to agree with a
model fitted from a catalogue value.

![Airflow resistance measurement rigs. Left, the ISO 9053-1 static method: a specimen sealed at its rim inside a measurement cell, resting on a perforated support grid, with at least one bore of free space above it, a thickness gauge in contact with the specimen in position, a flow source feeding a flowmeter below and a differential manometer across the specimen, both instruments annotated with their five percent tolerance. Right, the ISO 9053-2 alternating method: an oscillating piston driving a cavity terminated either by the measurement cell carrying the specimen or by an airtight plug, each with its own stroke, and a microphone reading the cavity level](https://docs.jmrp.io/phonometry/media/diagram_airflow_resistance.svg)

## 2. Static method (ISO 9053-1)

In the static method (ISO 9053-1:2018) a steady laminar flow is stepped up and the
pressure difference plotted against the linear velocity $u = q_v/A$. A regression
of at least second order **constrained through the origin**, $\Delta p = a\,u +
b\,u^2$, is fitted, and $\Delta p$ and $R_\mathrm{s}$ are read at the reference velocity
$u = 0.5\ \text{mm/s}$ (Clause 7.5); the highest velocity must not exceed
15 mm/s. Because $R_\mathrm{s} = \Delta p/u = a + b\,u$, the linear term $a$ is the
zero-velocity specific airflow resistance.

Neither velocity is arbitrary, and Clause 7.5 says why. A plane wave of
particle velocity 0.5 mm/s carries a sound pressure of 0.2 Pa (80 dB), so the
reference velocity puts the steady-flow rig at the amplitude the material will
actually meet in use, which is the whole reason an aerodynamic measurement
predicts an acoustic quantity. The 15 mm/s ceiling is the same arithmetic at
110 dB, and the standard warns that some materials are already non-linear
there. That non-linearity is also why the fit is constrained through the
origin and read at the bottom: $\Delta p$ must vanish with $u$, the linear
coefficient $a$ is the viscous (Darcy) resistance acoustics wants, and the
quadratic coefficient $b$ is the inertial correction that grows with velocity.
One trap follows from the same clause: a textile "air permeability" to
ISO 9237 is measured at velocities ten to a hundred times higher and therefore
yields a *different*, larger resistivity, so a datasheet permeability figure
can never be substituted for a measured $\sigma$ in the models of section 4.

![ISO 9053-1 static-method airflow resistance: the measured pressure drop against linear airflow velocity, fitted with a through-origin quadratic, with the specific airflow resistance evaluated at the 0.5 mm/s reference velocity](https://docs.jmrp.io/phonometry/media/airflow_resistance.svg)

*The slightly super-linear pressure drop is fitted through the origin; the
specific airflow resistance is the fit read at 0.5 mm/s.*

**Show the code for this figure**

```python
import matplotlib.pyplot as plt
import numpy as np
from phonometry import materials

area = np.pi * 0.05**2                      # 100 mm diameter cell [m^2]
u = np.array([0.5, 1, 2, 4, 8, 12]) * 1e-3  # linear velocity [m/s]
dp = 1.6e4 * u + 4.0e5 * u**2               # measured pressure drop [Pa]
r = materials.static_airflow_resistance(u, dp, area=area, thickness_m=0.05)

u_fit = np.linspace(0.0, 13e-3, 200)
dp_fit = r.linear_coefficient * u_fit + r.quadratic_coefficient * u_fit**2
fig, ax = plt.subplots()
ax.plot(u_fit * 1e3, dp_fit, label="Through-origin fit  dp = a u + b u^2")
ax.plot(u * 1e3, dp, "o", label="Measured pressure drop")
ax.plot(r.evaluation_velocity * 1e3, r.pressure_drop_pa, "D",
        label="Evaluation at 0.5 mm/s")
ax.set_xlabel("Linear airflow velocity u [mm/s]")
ax.set_ylabel("Pressure drop dp [Pa]")
ax.set_title(rf"$R_\mathrm{{s}}$ = {r.specific_resistance:.0f} Pa·s/m")
ax.legend()
plt.show()
```

```python
import numpy as np
from phonometry import materials

area = np.pi * 0.05**2            # 100 mm diameter cell [m^2]

# What the rig actually reads: a rotameter between the source and the specimen
# (in ml/min here) and a differential manometer. Clause 3.4 turns the first
# into the linear velocity the standard evaluates at.
q_v = np.array([236.0, 471.0, 942.0, 1885.0, 3770.0, 5655.0]) * 1e-6 / 60.0
u = np.array([materials.linear_airflow_velocity(q, area) for q in q_v])
print(np.round(u * 1e3, 2))       # [ 0.5  1.   2.   4.   8.  12. ] mm/s
dp = 1.6e4 * u + 4.0e5 * u**2                    # measured pressure drop [Pa]

r = materials.static_airflow_resistance(u, dp, area=area, thickness_m=0.05)
print(round(r.specific_resistance))   # 16200   R_s [Pa*s/m]
print(round(r.resistivity))           # 324000  sigma [Pa*s/m^2]
print(round(r.linear_coefficient))    # 16000   a = R_s at u -> 0
r.plot()   # the figure above: the fitted dp(u) with the evaluation point

# The same R_s straight from the pair the manometer and the rotameter give,
# without forming R first (Clause 3.2, second call form).
print(round(materials.specific_airflow_resistance(pressure_drop_pa=8.1, velocity=0.5e-3)))
```

Those six flows are what a 100 mm cell needs to cover the clause's range: the
0.5 mm/s reference velocity is 236 ml/min through that cross-section and the
15 mm/s ceiling is about 7 l/min, so the rotameter has to resolve a factor of
thirty and land accurately at the bottom of it, where the evaluation happens.
`linear_airflow_velocity` is scalar-only, hence the comprehension.

### 2.1 Running the static rig

The arithmetic above is the easy half. Everything that decides whether the
number is right happens before the first flow step, and ISO 9053-1 spells it
out in Clauses 5 to 7.

**The specimen and the cell.** The specimen must be wide enough to be
representative: at least 10 pores across for a foam, 10 fibres for a fibrous
material, 10 grains for a granular one, and where the microstructure is
unknown, at least 95 mm in diameter or 90 mm on the smallest edge
(Clause 6.2.1). The measurement cell has the *same* lateral dimensions as the
specimen (its own floor is 29 mm per Clause 5.2) and needs a free space of at
least one diameter in front of the specimen so the flow arrives laminar and
unidirectional, with more than a diameter behind it recommended for thin,
low-porosity specimens. The specimen rests on a perforated support or a
thin-wire grid with at least 50 % open area evenly distributed, holes no
smaller than 3 mm, and an airflow resistance under 1 % of the specimen's;
Clause 5.2 even gives the closed form $R_\mathrm{s} = 8\eta h/(\phi r^2)$ for a
perforated plate so the support can be checked rather than assumed.

**The two ways to get it wrong, and their signs.** Both dominant error
mechanisms are one-sided, which makes them diagnosable. A leak around the
specimen rim is a resistance *in parallel* with the specimen, so the measured
$R$ always comes out **too low**; Clause 7.2 seals the rim with a thin layer
of petroleum jelly, thread seal tape or rings, and warns that the jelly must
not penetrate the specimen, because inside the pores it does the opposite and
reads high. Compression is the mirror image: $\sigma$ rises steeply with bulk
density, so a specimen forced into the cell reads **too high**, repeatably and
convincingly, which is why Clause 6.2.2 requires the thickness not to be
modified when the specimen is placed in the cell. Thin specimens may be
stacked only if the microstructure survives it, and for fibrous materials or
non-wovens only with matching fibre orientation; where that is impossible,
Annex A gives an impedance-tube route instead. The diagnostic worth
remembering: a resistance that drifts *downward* as the flow rate rises is
usually a leak opening under pressure, not material non-linearity, which
pushes the other way.

**In-position thickness.** The thickness that defines $d$, and with it the
tested bulk density, is measured with the specimen in the cell, by bringing
the gauge into contact with the upper surface and compressing it lightly
(Clauses 7.3 and 7.4), not with a rule on the bench beforehand.

**Instrument tolerances.** The volumetric flow and the differential pressure
must each be good to ±5 % of the indicated value (Clauses 5.4 and 5.5), the
flow tap sits between the source and the specimen, and the pressure instrument
must resolve down to 0.1 Pa; the worked example above needs this, since the
whole reading at the reference velocity is 8.1 Pa. The flow source should
reach 0.5 mm/s and hold it stable (Clause 5.3).

**Is the rig itself usable?** Clause 5.6 is the only check that answers that,
and nothing in this library performs it: at least one *calibrated* test
specimen shall be measured before a series of measurements, at least once a
day, after any hardware or software change, and after any significant change
of conditions (0.5 kPa of pressure, 5 °C of temperature or 5 % of humidity).
The artefact is normally made of straight cylindrical pores precisely because
its resistance follows in closed form from the same $8\eta h/(\phi r^2)$, so
the check is against theory rather than against another laboratory, and the
measured value must land within ±10 % of it. Record the check beside the
measurement (the `notes` field of the fiche is the place to put it), because
the result object carries only the fit.

**How much agreement to expect.** Usually three specimens are measured
(Clause 6.3), and Clause 8 puts the inter-laboratory reproducibility at around
15 % for open-cell foams, "strongly" worse for granular materials or materials
with semi-closed pores, dominated by exactly the leakage and compression above.
A 20 % disagreement with a catalogue value is therefore not evidence of a
faulty instrument; a 20 % disagreement between your own three specimens is.

### ISO 9053-1 report (`.report()`)

`StaticAirflowResult.report(path)` renders a one-page PDF fiche laid out like an
accredited airflow-resistance test report (ISO 9053-1:2018, static method): a
standard-basis line, an optional metadata header block, a two-panel body with a
metrics table (the evaluation velocity, the fitted pressure difference, the
airflow resistance $R$, the specific airflow resistance $R_\mathrm{s}$, the airflow
resistivity $\sigma$ when a thickness is available, and the through-origin fit
coefficients $a$ and $b$) beside the fitted $\Delta p(u)$ curve (the result's
own `.plot()`), the boxed specific airflow resistance $R_\mathrm{s}$ with $R$ and
$\sigma$ alongside, and a footer with the fixed disclaimer. ISO 9053-1 is a
material characterisation, so the fiche carries no pass/fail verdict.

It uses the same `ReportMetadata` container and rendering engine as the other
fiches. The descriptive fields that apply here are `client`, `manufacturer`,
`specimen`, `thickness_m` (the specimen thickness $d$, in metres, shown in
millimetres), `test_room`, `test_date`, `temperature_c`,
`relative_humidity_percent`, `measurement_standard`, `laboratory`, `operator`,
`report_id` and `notes`. The `requirement` field is ignored (ISO 9053-1 has no
verdict). The fiche embeds the fitted curve, so rendering needs both reportlab
and matplotlib (`pip install "phonometry[report,plot]"`); only
`engine="reportlab"` is supported. The fiche renders in English by default;
pass `language="es"` for a Spanish fiche (translated fixed strings and a comma
decimal separator).

```python
from phonometry import materials, ReportMetadata

r = materials.static_airflow_resistance(u, dp, area=area, thickness_m=0.05)
r.report(
    "airflow_fiche.pdf",
    metadata=ReportMetadata(
        specimen="50 mm porous absorber (open-cell)",
        thickness_m=0.050,
        measurement_standard="ISO 9053-1",
        test_room="Static airflow rig, 100 mm cell",
        laboratory="Phonometry Reference Laboratory",
    ),
)                                  # R_s, R and sigma at u = 0.5 mm/s
```

**Report:** [ISO 9053-1 static airflow-resistance example report (PDF)](https://docs.jmrp.io/phonometry/media/iso9053_airflow_resistance_example.pdf): One-page static airflow-resistance fiche: a metadata header (client, manufacturer, specimen, specimen thickness, test facility, temperature and humidity), a metrics table with the evaluation velocity, the fitted pressure difference, the airflow resistance R, the specific airflow resistance R_s, the airflow resistivity sigma and the through-origin fit coefficients a and b beside the fitted pressure-drop curve, and the boxed specific airflow resistance R_s with R and sigma alongside, read at the 0.5 mm/s reference velocity.

Static airflow-resistance fiche (StaticAirflowResult.report), R_s, R and sigma at 0.5 mm/s.

## 3. Alternating method (ISO 9053-2)

In the alternating method (ISO 9053-2:2020) a piston oscillating at 1–4 Hz drives
an alternating flow into a cavity terminated either by the specimen or by an
airtight plug; the resistance follows from the sound-pressure-level difference
between the two terminations (Formula (2)):

$$
R = \frac{\kappa'\,P_\mathrm{S}}{2\pi f\,V}\cdot\frac{h_\mathrm{t}}{h_\mathrm{s}}\cdot
    10^{(L_{p,\mathrm{s}} - L_{p,\mathrm{t}})/20},
$$

with $P_\mathrm{S}$ the static (atmospheric) pressure, $f$ the piston frequency, $V$ the
volume of the air cavity (including every connecting pipe, with the piston at
mid-stroke), $L_{p,\mathrm{s}}$ and $L_{p,\mathrm{t}}$ the cavity sound pressure levels measured
with the measurement cell carrying the specimen and with the airtight
termination, $h_\mathrm{s}$ and $h_\mathrm{t}$ the corresponding piston stroke amplitudes, and
$\kappa'$ the **effective** ratio of specific heats.

**What the formula is saying.** The airtight termination is the reference: with
it the cavity is a sealed compliance, the piston's whole swept volume goes into
compressing the air, and the level is high. Put the specimen in its place and
most of that flow bleeds through the pores instead, so the cavity level falls.
The drop $L_{p,\mathrm{t}} - L_{p,\mathrm{s}}$ *is* the measurement. The stroke ratio
$h_\mathrm{t}/h_\mathrm{s}$ only normalises the two runs to the same volume velocity, which is
why the two strokes are deliberately unequal: ISO 9053-2 Clause 6.2 NOTE 1
gives 1.4 mm with the airtight termination against 14 mm with the specimen for
a 10 mm piston, a 100 mm cell and a cavity near $10^{-3}$ m³ (exactly the
example below), so that both levels land inside the analyser's useful range.
Two consequences are worth having. Only a level *difference* at one frequency
enters, so the sound measuring device needs no absolute calibration and its
frequency response is not critical; what matters is level linearity
(Clause 8.7). And the run order is fixed by Clause 8.5/8.6: measure $P_\mathrm{S}$, set
the stroke to $h_\mathrm{s}$, read the piston frequency, the background level $L_{p,\mathrm{b}}$
(source running, neither cell nor termination fitted) and then $L_{p,\mathrm{s}}$; only
then swap in the airtight termination, set $h_\mathrm{t}$, confirm the frequency has not
moved, and read $L_{p,\mathrm{t}}$.

Heat conduction
between the oscillating air and the cavity walls makes the compression not fully
adiabatic; the normative Annex A corrects $\kappa$ down to

$$
\kappa' = \frac{\kappa}{\sqrt{1 + (\kappa-1)\tfrac{S}{V}b
          + \tfrac{1}{2}\big((\kappa-1)\tfrac{S}{V}b\big)^2}},
\qquad b = \sqrt{\frac{2 c_0 l_\mathrm{h}}{\omega}},\quad
l_\mathrm{h} = \frac{k_\mathrm{a}}{\rho_0 c_0 C_\mathrm{P}},
$$

with $S$ and $V$ the cavity surface and volume, $b$ the thermal boundary-layer
thickness (`thermal_boundary_layer_thickness`), $l_\mathrm{h}$ the characteristic
thermal diffusion length built from the thermal conductivity $k_\mathrm{a}$, the density
$\rho_0$, the speed of sound $c_0$ and the specific heat at constant pressure
$C_\mathrm{P}$, and $\omega = 2\pi f$. For the Annex A.3 example cavity this gives
$\kappa' = 1.370$, about 2 % below the adiabatic 1.4008.

```python
from phonometry import materials

# Annex A.3 cavity: closed cylinder 100 mm x 100 mm, piston at 2 Hz
kp = materials.effective_kappa(cavity_surface=0.0471, cavity_volume=7.854e-4, frequency=2.0)
print(round(kp, 3))               # 1.37   effective ratio of specific heats
print(round(materials.thermal_boundary_layer_thickness(2.0) * 1e3, 2))   # 1.83 mm

R = materials.alternating_airflow_resistance(
    level_specimen=74.0, level_termination=90.0,
    piston_stroke_specimen=14e-3, piston_stroke_termination=1.4e-3,
    frequency=2.0, cavity_volume=7.854e-4, kappa_prime=kp,
)
print(round(R))                   # 222949  airflow resistance R [Pa*s/m^3]
```

$b$ is 1.83 mm at 2 Hz, which is millimetre-scale, against a cavity 100 mm
across. That is why $\kappa'$ enters through the *surface-to-volume ratio*
$S/V$ and not through the cavity's absolute size: only the air within a
boundary layer of the
wall exchanges heat with it, so a rig scaled down at constant shape keeps the
same $b$ while $S/V$ grows, and the correction grows with it.

Pass the `effective_kappa` result to `alternating_airflow_resistance` for an
Annex-A-conforming figure; its `kappa_prime` argument otherwise defaults to the
uncorrected adiabatic 1.4.

![Effective ratio of specific heats against piston frequency over the normative 1 to 4 Hz band for three cavities of surface-to-volume ratio 30, 60 and 120 per metre, all lying below the dashed adiabatic value 1.4008, with the ISO 9053-2 Annex A.3 point marked at 2 hertz and a right-hand axis reading the percentage error in R that using the adiabatic value would cause](https://docs.jmrp.io/phonometry/media/effective_kappa.svg)

*The Annex A correction over the whole 1–4 Hz piston band. It falls as the
frequency falls (a thicker boundary layer reaches more of the air) and rises
with the cavity's surface-to-volume ratio, which is the only geometric
quantity it depends on. For the Annex A.3 cavity ($S/V = 60$ 1/m), the
correction is about 2 %, and since $R$ is proportional to $\kappa'$, skipping
it biases every reported resistance by that much. A cavity twice as compact
doubles the penalty, so a scaled-down rig cannot treat the correction as
optional.*

**Show the code for this figure**

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

f = np.linspace(1.0, 4.0, 120)
for s_over_v in (30.0, 60.0, 120.0):
    volume = 7.854e-4                      # the Annex A.3 cavity volume
    kappa = [materials.effective_kappa(cavity_surface=s_over_v * volume,
                                       cavity_volume=volume, frequency=fi)
             for fi in f]
    plt.plot(f, kappa, label=f"S/V = {s_over_v:g} 1/m")
plt.axhline(1.4008, ls="--", label="adiabatic 1.4008")
plt.plot([2.0], [materials.effective_kappa(cavity_surface=0.0471,
                                           cavity_volume=7.854e-4,
                                           frequency=2.0)], "D")
plt.xlabel("Piston frequency f [Hz]")
plt.ylabel("Effective ratio of specific heats")
plt.legend()
plt.show()
```

**Acceptance criteria (Clause 8.7).** Formula (2) is only valid when two
conditions hold. Formula (3),
$(h_\mathrm{t}/h_\mathrm{s})\,10^{(L_{p,\mathrm{s}}-L_{p,\mathrm{t}})/20} < 0.3$, keeps the linearised Formula (2)
close enough to the exact relation of Annex B; Formula (4),
$L_{p,\mathrm{s}} - L_{p,\mathrm{b}} > 10$ dB, keeps the background's influence on $L_{p,\mathrm{s}}$
below 0.4 dB for uncorrelated noise (a correlated background needs more).
`alternating_airflow_resistance` accepts `background_level` precisely so it can
check the second, and warns through `AirflowResistanceWarning` when either
fails or when the piston frequency leaves 1–4 Hz. When one does fail, the
standard offers four knobs: change the specimen length, the specimen diameter,
the cavity volume, the piston frequency or the piston stroke. It is also worth
checking that the flow the piston drives through the specimen lands in the
recommended window: Clause 6.2 gives $q_v = 2\pi f h A_\mathrm{P}$ and Formula (1)
$v_\mathrm{s} = q_v/A$, and recommends rms velocities between 0.5 mm/s and 4 mm/s.
This is the same physical regime as the static method's reference velocity.

```python
# The same run, with the background level so the Formula (4) check is live,
# and the flow the piston actually drives through a 100 mm cell.
R = materials.alternating_airflow_resistance(
    level_specimen=74.0, level_termination=90.0,
    piston_stroke_specimen=14e-3, piston_stroke_termination=1.4e-3,
    frequency=2.0, cavity_volume=7.854e-4, kappa_prime=kp,
    background_level=52.0,          # Formula (4): 74.0 - 52.0 = 22 dB > 10 dB
)
piston_area = np.pi * 0.005**2                          # 10 mm piston
q_s = materials.piston_volume_flow_rate(2.0, 14e-3, piston_area)
print(round(q_s / area * 1e3, 2))   # 1.76 mm/s, inside the 0.5-4 mm/s window

# Formula (3): the ratio the standard caps at 0.3.
print(round((1.4e-3 / 14e-3) * 10 ** ((74.0 - 90.0) / 20), 3))   # 0.016

# Finish the chain the static method gets for free: R -> R_s -> sigma.
r_s = materials.specific_airflow_resistance(R, area)
print(round(r_s, 1), round(r_s / 0.05))    # 1751.0 Pa*s/m, 35021 Pa*s/m^2
```

**What Part 2 returns, and what it does not.** Unlike the static method,
`alternating_airflow_resistance` returns a bare resistance in Pa·s/m³, not a
result object: there is no `StaticAirflowResult` equivalent, so no `.plot()`
and no ISO 9053-2 fiche. The two steps to $R_\mathrm{s}$ and $\sigma$ are the
caller's responsibility; that is what the last two lines above do. What Part 2
buys instead is the specimens Part 1 cannot hold: very open or very thin
materials whose static pressure drop is too small to read against the rig's own
leakage, and delicate or granular samples a steady flow would deform or
displace. The two
methods should agree inside the reproducibility quoted above when the specimen
is properly sealed, with one documented exception. ISO 9053-2 Clause 5 NOTE
warns that for materials whose visco-inertial transition frequency lies below
100 Hz (coarse metal or plant fibres, low-porosity foams with big pores,
coarse road pavements) the static method can legitimately give a different
answer. Part 1 remains the reference method whenever a stable static reading
is possible.

## 4. What the resistivity feeds

The flow resistivity is not an end in itself: almost every consumer of the
number is a **porous-material model**. The empirical Delany-Bazley and Miki
regressions predict a porous material's complex characteristic impedance and
wavenumber from $\sigma$ *alone*, through the dimensionless ratio
$X = \rho_0 f/\sigma$; the physics-based Johnson-Champoux-Allard model keeps
$\sigma$ as the first of its five parameters. Fitted to a tube or
reverberation-room measurement once, those models then predict the layer at
any thickness, backing or incidence angle. The whole model family, and the
multilayer solver that stacks the layers, lives in
[Porous and Multilayer Absorbers](https://docs.jmrp.io/phonometry/materials/absorbers/porous-absorbers/); the
regression validity window of Delany-Bazley ($0.01 < X < 1$) is stated there
in the same $X$ that $\sigma$ defines.

Because $\sigma$ enters the models as a ratio, its useful range is bounded on
both sides. A layer with too little resistivity barely couples to the sound
field: air moves through it freely and little energy is dissipated. Too much
resistivity and the layer behaves like a wall: the wave reflects off the
front face before the pores can absorb it. Between the two lies the classic
design window for the total flow resistance of a layer of thickness $d$,

$$
\rho_0 c_0 \;\lesssim\; R_\mathrm{s} = \sigma\,d \;\lesssim\; 4\,\rho_0 c_0,
$$

that is, a specific flow resistance of one to four times the characteristic
impedance of air ($\rho_0 c_0 \approx 410$ Pa·s/m). A 50 mm blanket therefore
wants a resistivity of roughly 8 kPa·s/m² to 33 kPa·s/m², which is exactly
where commercial absorber products cluster.

![Left panel: normal-incidence absorption against frequency for a 50 mm hard-backed porous layer at flow resistivities of 2, 8, 20, 33 and 100 kPa s per square metre, showing a transparent layer at the lowest value and a reflecting one at the highest. Right panel: absorption at 500 Hz and 1 kHz and the diffuse-field value against the dimensionless ratio sigma d over rho0 c0 on a logarithmic axis, with the region between 1 and 4 shaded](https://docs.jmrp.io/phonometry/media/flow_resistivity_window.svg)

*The window, drawn. Left: the same 50 mm hard-backed layer at five
resistivities. The 2 kPa·s/m² layer is too transparent everywhere. The wave
passes through it, reflects off the wall and comes back out. The
100 kPa·s/m² one leads below about 400 Hz, where more resistance still helps,
and then saturates near 0.89 while the three window curves reach 1: above
1 kHz the wave is reflecting off the face before the pores can act. Right: the
same family read against the dimensionless $\sigma d/(\rho_0 c_0)$, with the
$1$ to $4$ window shaded. The optimum is flat-topped, which is the practical
point. A catalogue spread of tens of percent inside the window costs almost
nothing, and the same spread just outside it costs a great deal. Note also
that the diffuse-field curve peaks lower on the axis than the
normal-incidence one, so a layer optimised in the tube is slightly
over-resistive for a room. The window is engineering guidance and forms no
part of either ISO 9053 part.*

**Show the code for this figure**

```python
f_win = np.geomspace(100.0, 5000.0, 200)
for sigma in (2e3, 8e3, 20e3, 33e3, 100e3):
    layer = materials.layered_absorber(
        f_win, [materials.PorousLayer(thickness_m=0.05, medium=materials.miki(f_win, sigma))]
    )
    plt.semilogx(f_win, layer.absorption, label=f"{sigma/1e3:g} kPa s/m^2")
plt.xlabel("Frequency [Hz]")
plt.ylabel("Normal-incidence absorption")
plt.legend()
plt.show()
```

The worked specimen of section 2 is a useful counter-example: its
324 kPa·s/m² is an order of magnitude above the top of the window, so at 50 mm
it is not a bulk absorber at all but a resistive facing, of the kind that
belongs in front of a cavity rather than alone against a wall.

Typical orders of magnitude, for orientation rather than design: glass wools
run from a few kPa·s/m² in light thermal grades to some tens of kPa·s/m² in
dense acoustic boards; rock wools sit somewhat higher at equal density;
open-cell foams (melamine, polyurethane) span roughly 5 kPa·s/m² to
30 kPa·s/m²; and fibrous felts and compressed boards can exceed
100 kPa·s/m², at which point they act more as resistive facings than as bulk
absorbers. Within one product family the resistivity rises steeply with bulk
density and falls with fibre diameter, which is why nominally identical
products from two production runs can differ by tens of percent, and why a
measured $\sigma$, not a catalogue value, should anchor any model fit that
will be trusted quantitatively.

The number also travels beyond the porous models. The perforated-panel and
microperforated-panel impedances contain the same viscous physics evaluated
in a single hole; the [slit absorber](https://docs.jmrp.io/phonometry/materials/absorbers/metamaterial-absorbers/)
inherits it per slit; and when an
[impedance tube](https://docs.jmrp.io/phonometry/materials/absorbers/impedance-tube/) measurement disagrees
with a model prediction, the first parameter to re-measure, before touching
the model, is $\sigma$.

## What this guide covers

- **Covered.** Both parts of ISO 9053: the static method of Clause 7.5
  (`static_airflow_resistance`, the through-origin regression and the 0.5 mm/s
  evaluation velocity) and the alternating method of ISO 9053-2 Formula (2) with
  the Annex A effective ratio of specific heats (`effective_kappa`,
  `thermal_boundary_layer_thickness`, `alternating_airflow_resistance`),
  including the 1-4 Hz and Formula (3)/(4) validity warnings, and the
  accredited-style ISO 9053-1 fiche. The definitional conversions that connect
  them to what a rig reads are `linear_airflow_velocity` (Clause 3.4),
  `specific_airflow_resistance` (Clause 3.2, in both call forms) and
  `piston_volume_flow_rate` (ISO 9053-2 Clause 6.2).

- **Not covered.** The equivalent-fluid models that consume the resistivity (Delany-Bazley, Miki,
  Johnson-Champoux-Allard) live in [Porous and Multilayer
  Absorbers](https://docs.jmrp.io/phonometry/materials/absorbers/porous-absorbers/), and the
  $\sigma d$ design window of section 4 is engineering guidance, not part of
  either standard. ISO 9053-1 defines no pass/fail verdict or single-number
  rating, so the fiche carries neither.

## See also

- [Porous and Multilayer Absorbers](https://docs.jmrp.io/phonometry/materials/absorbers/porous-absorbers/): the
  Delany-Bazley, Miki and Johnson-Champoux-Allard models that consume the
  resistivity, and the multilayer transfer-matrix solver.
- [Impedance Tube](https://docs.jmrp.io/phonometry/materials/absorbers/impedance-tube/): the normal-incidence
  measurement that a $\sigma$-anchored model is usually fitted against.
- [Sound Absorption Measurement and Rating](https://docs.jmrp.io/phonometry/materials/absorbers/absorption-measurement/):
  the reverberation-room measurement and the ISO 11654 rating of the finished
  absorber.
- API reference: [`materials`: airflow resistance](https://docs.jmrp.io/phonometry/reference/api/materials/airflow-resistance/).
- Theory: [Acoustic material characterisation](https://docs.jmrp.io/phonometry/reference/theory/materials-surfaces/#acoustic-material-characterisation-iso-11654-iso-9053-12-iso-10534-12-astm-e2611): the ISO 9053 definitions of resistance, resistivity and the flow regime they assume.
- [Published catalogues](https://docs.jmrp.io/phonometry/reference/catalogues/porous-materials/): the
  flow resistivity, porosity and tortuosity the books print, for named
  specimens and for whole classes of material.

## References

- **Allard, J. F., & Atalla, N. (2009).** *Propagation of sound in porous media: Modelling sound absorbing materials* (2nd ed.). Wiley. https://doi.org/10.1002/9780470747339
  The porous-material theory that consumes the measured resistivity: viscous and thermal dissipation in the pores, and the equivalent-fluid models of section 4. ISBN 978-0-470-74661-5.
- **International Organization for Standardization. (2018).** [*Acoustics — Determination of airflow resistance — Part 1: Static airflow method*](https://www.iso.org/standard/69869.html) (ISO 9053-1:2018).
  The static method of section 2, with the through-origin regression and the 0.5 mm/s reference velocity.
- **International Organization for Standardization. (2020).** [*Acoustics — Determination of airflow resistance — Part 2: Alternating airflow method*](https://www.iso.org/standard/76744.html) (ISO 9053-2:2020).
  The alternating method of section 3, with the Annex A effective ratio of specific heats, validated against the standard's own Annex A.3 worked example.
