Insulation Ratings (ISO 717)

Every sound-insulation quantity in this documentation ends its journey the same way: a band spectrum walks in, a single number walks out. Whether the spectrum is a laboratory $R$, a field $R^{\prime }$ or $D_{nT}$, a façade $D_{2m,nT}$, an impact $L_n$ or $L_{nT}^{\prime }$, or a flanking $D_{n,f}$, what regulations quote is the weighted single number of ISO 717, and two reference-curve engines produce all of them: ISO 717-1 for airborne quantities, where an unfavourable deviation is a band falling below the reference, and ISO 717-2 for impact quantities, with the sign flipped because a higher impact level is worse. This guide covers the engines themselves: the shifting rule, the spectrum adaptation terms $C$, $C_{tr}$ and $C_I$, the enlarged-range and one-decimal variants, and the ISO 717 fiche. The measurements that produce the spectra live in the field, laboratory, intensity, survey and flanking guides; the prediction that consumes the ratings is EN 12354.

How do I compute Rw from third-octave sound reduction data in Python?

Pass the 16 one-third-octave values from 100 Hz to 3150 Hz to building.weighted_rating(R). It shifts the ISO 717-1 reference curve and returns the rating read at 500 Hz, plus the spectrum adaptation terms $C$ and $C_{tr}$. For the measured spectrum used below, w.rating, w.c, w.ctr prints 30 -2 -3: $R_w(C;C_{tr})=30(-2;-3)$. The same call rates $R^{\prime }$, $D_n$ and $D_{nT}$.

Airborne ratings (ISO 717-1)

The band spectrum is collapsed to one number by the reference-curve method of ISO 717-1: a fixed reference curve is shifted in 1 dB steps toward the measured curve until the sum of unfavourable deviations (where the measurement falls below the reference) is as large as possible but not more than 32.0 dB (16 one-third-octave bands) or 10.0 dB (5 octave bands). The two caps are one rule, not two constants: $32.0=2.0\times 16$ and $10.0=2.0\times 5$, so what the method actually permits is an average unfavourable deviation of 2 dB per band. The rating ($R_w$, $R_w^{\prime }$, $D_{nT,w}$ …) is the shifted reference read at 500 Hz.

The reference curve itself (ISO 717-1:2020, Table 3) is not arbitrary either. It rises 9 dB per octave from 100 Hz to 500 Hz, then only 4 dB over the next octave, and is flat from 1250 Hz up. A rating engine carrying that shape is deliberately lenient at low frequency and unforgiving above 1 kHz, which is why a construction with a low-frequency dip can still rate well — and why $C_{tr}$ exists to put the penalty back. The result object hands you the curve in its shifted position, so the unshifted one is a subtraction away:

import numpy as np
from phonometry import building

R = [20.4, 16.3, 17.7, 22.6, 22.4, 22.7, 24.8, 26.6,
     28.0, 30.5, 31.8, 32.5, 33.4, 33.0, 31.0, 25.5]
w = building.weighted_rating(R)

# The curve is shifted until it reads `rating` at 500 Hz; unshifted it reads
# 52 dB there, so the shift applied was 52 - rating.
reference = np.asarray(w.shifted_reference) + (52 - w.rating)
print(w.rating, 52 - w.rating)          # 30 22   (shifted down by 22 dB)
print(reference.astype(int).tolist())
# [33, 36, 39, 42, 45, 48, 51, 52, 53, 54, 55, 56, 56, 56, 56, 56]

The spectrum adaptation terms $C$ (pink noise) and $C_{tr}$ (urban traffic) add the low-frequency penalty of a real source. Both come from one definition (Clause 4.5, Formulae (1) and (2)): re-rate the measured spectrum against a standard source spectrum and subtract the curve-based rating,

$$C_j=X_{A,j}-X_w,X_{A,j}=-10{\log }_{10}\sum _{i}10^{(L_{i,j}-X_i)/10},$$

where $L_{i,j}$ is the level of source spectrum $j$ in band $i$ (both spectra are A-weighted and normalised to 0 dB overall) and $X_i$ the measured insulation in that band, to one decimal place. The two terms differ only in which spectrum is substituted: No. 1, A-weighted pink noise, representative of living activities (speech, music, radio, television), gives $C$; No. 2, A-weighted urban road traffic, whose energy sits at low frequency, gives $C_{tr}$. So the rating plus the term ($R_w+C$ for a laboratory index, $R_w^{\prime }+C$ or $D_{nT,w}+C$ for the quantities of the field guide, and likewise with $C_{tr}$) is the A-weighted level difference the construction achieves against that source — which is what makes the sum, and not the rating alone, the thing worth specifying. Reading them:

• $C$ stays small for most constructions (0 to −2 dB is typical): the pink spectrum is close to the weighting already implicit in the reference curve.
• $C_{tr}$ punishes weak low-frequency insulation. A lightweight double leaf with its mass-air-mass resonance near 100 Hz can carry a $C_{tr}$ of −5 to −10 dB, while a heavy monolithic wall with the same $R_w$ loses far less: two constructions with equal ratings can differ audibly against traffic.
• Design with the descriptor that matches the noise, carried by the field quantity the requirement rates: $R_w^{\prime }+C_{tr}$ for a façade on a busy road, $D_{nT,w}+C$ (or the plain rating, where the regulation says so) between dwellings, the two example requirements of ISO 717-1, 5.3.

Two constructions, one rating. Both spectra come from the library’s own panel models — Sharp’s method for the concrete leaf, the mass-air-mass double-wall model for the lightweight one — and both rate to $R_w=49$ dB, so a specification written on $R_w$ alone treats them as interchangeable. They are not: the double leaf enters the rated range still climbing out of its 82 Hz mass-air-mass resonance, and against traffic it gives up 6 dB more than the concrete wall. $R_w+C_{tr}$ is what a façade on a busy road is actually specified on.

Show the code for this figure

import matplotlib.pyplot as plt
import numpy as np
# `building` is the import of the reference-curve block above.

freqs = np.array([100, 125, 160, 200, 250, 315, 400, 500, 630, 800,
                  1000, 1250, 1600, 2000, 2500, 3150], dtype=float)

# 150 mm dense concrete: m' = 2300 x 0.15 = 345 kg/m2, fc = 125 Hz.
heavy = np.round(building.single_panel_transmission_loss(
    freqs, 345.0, critical_frequency=125.0).transmission_loss, 1)
# Metal-stud double leaf: 12 kg/m2 per leaf, 90 mm cavity.
light_result = building.double_wall_transmission_loss(freqs, mass1=12.0,
                                                      mass2=12.0, gap=0.09)
light = np.round(light_result.transmission_loss, 1)

w_heavy = building.weighted_rating(heavy)
w_light = building.weighted_rating(light)
print(w_heavy.rating, w_heavy.c, w_heavy.ctr)   # 49 -2 -7
print(w_light.rating, w_light.c, w_light.ctr)   # 49 -5 -13
print(round(light_result.resonance_frequency, 1))            # 81.5 Hz

fig, (ax_spec, ax_bar) = plt.subplots(1, 2, figsize=(13.0, 5.8))
x = np.arange(16)
ax_spec.plot(x, w_heavy.shifted_reference, "--", label="shifted reference")
ax_spec.plot(x, heavy, "-o", label="150 mm dense concrete")
ax_spec.plot(x, light, "-s", label="double leaf, 12 kg/m2 + 90 mm")
ax_spec.set_xticks(x, [f"{f:g}" for f in freqs], rotation=45, fontsize=8)
ax_spec.set(xlabel="Frequency [Hz]", ylabel="Sound reduction index R [dB]")
ax_spec.legend()

groups = ("Rw", "Rw + C", "Rw + Ctr")
xb = np.arange(3)
ax_bar.bar(xb - 0.19, [w_heavy.rating, w_heavy.rating + w_heavy.c,
                       w_heavy.rating + w_heavy.ctr], width=0.36,
           label="concrete")
ax_bar.bar(xb + 0.19, [w_light.rating, w_light.rating + w_light.c,
                       w_light.rating + w_light.ctr], width=0.36,
           label="double leaf")
ax_bar.set_xticks(xb, groups)
ax_bar.set_ylabel("Single number [dB]")
ax_bar.legend()
plt.show()

Every figure of a rating on this page shows the curve where it came to rest. The rule that put it there is iterative, and the two positions that decide the answer are the ones a finished figure has already discarded: the last shift whose unfavourable sum still exceeded 32.0 dB, and the next shift after the answer, which is legal but wastes the allowance. The clip walks both engines through those steps on the two worked spectra of this page.

The reference curve steps in 1 dB increments toward the measured spectrum while the unfavourable deviations shade in band by band and their sum falls toward the 32.0 dB cap. The airborne engine stops at 31.8 dB for Rw = 30 dB, one step before the position that would waste the allowance; the impact engine then runs the same rule with the deviation's sign reversed, the curve stepping upward and stopping at 28.0 dB for L'nT,w = 79 dB.

Download the animation (WebM)

The reference curve steps in 1 dB increments toward the measured spectrum while the unfavourable deviations shade in band by band and their sum falls toward the 32.0 dB cap. The airborne engine stops at 31.8 dB for Rw = 30 dB, one step before the position that would waste the allowance; the impact engine then runs the same rule with the deviation's sign reversed, the curve stepping upward and stopping at 28.0 dB for L'nT,w = 79 dB.

Download the animation (WebM)

Both walks stop where the standard says: at the largest unfavourable sum that does not exceed 32.0 dB. The step before it is rejected for overshooting the cap, and the step after it is legal but reads a decibel lower for no reason, which is what “as large as possible” in ISO 717-1:2020, 4.4 and ISO 717-2:2020, 4.3.1 excludes.

The shifted reference has come to rest where the shaded unfavourable area totals 31.8 dB: one step back would overshoot the cap at 44.1 dB, and one more step down would waste the allowance at 20.6 dB. The rating is then read off that curve at 500 Hz, not off the measurement: the measured spectrum touches 26.6 dB there, and $R_w$ is 30 dB.

# `building`, `R` and `w` come from the reference-curve block above.
print(w.rating, w.c, w.ctr)                          # 30 -2 -3  ->  Rw(C;Ctr) = 30(-2;-3)
print(w.unfavourable_sum)                            # 31.8  (cap 32.0 = 2.0 x 16)

w.plot()   # measured R vs shifted ISO 717-1 reference, deviations shaded (needs matplotlib)

Show the code for this figure

import matplotlib.pyplot as plt
from phonometry import building

# Single-number rating from a measured 16-band R spectrum (ISO 717-1 Annex C)
R = [20.4, 16.3, 17.7, 22.6, 22.4, 22.7, 24.8, 26.6,
     28.0, 30.5, 31.8, 32.5, 33.4, 33.0, 31.0, 25.5]
w = building.weighted_rating(R)

# One line — measured curve vs the shifted ISO 717-1 reference, deviations shaded:
w.plot()
plt.show()

# By hand, from the band curve the result now carries:
fig, ax = plt.subplots()
ax.semilogx(w.band_centers, w.measured, "o-", label="Measured R")
ax.semilogx(w.band_centers, w.shifted_reference, "s--", label="Shifted reference")
ax.fill_between(w.band_centers, w.measured, w.shifted_reference,
                where=w.measured < w.shifted_reference, interpolate=True,
                alpha=0.3, label="Unfavourable deviations")
ax.set_xlabel("Frequency [Hz]")
ax.set_ylabel("Sound reduction index [dB]")
ax.set_title(f"Rw = {w.rating} dB  (C={w.c:+d}; Ctr={w.ctr:+d})")
ax.legend()
plt.show()

weighted_rating() parameters

Parameter	Type	Units	Range / default	Notes
values_by_band	1D array	dB	16 (thirds) or 5 (octaves)	Measured $R$, $R^{\prime }$, $D_{nT}$ … per band
bands	str or None	—	'third-octave' / 'octave' / None	None infers from the count

weighted_rating() returns a WeightedRatingResult (rating, c, ctr, unfavourable_sum, all integers except the sum), plus the three fields the figures on this page are drawn from: band_centers, measured and shifted_reference, and quantity naming what was rated.

Impact ratings (ISO 717-2)

The single-number rating (ISO 717-2) shifts the same style of reference curve, but an unfavourable deviation now occurs where the measurement exceeds the reference (impact noise is worse when higher), the sign opposite to ISO 717-1. The rating ($L_{n,w}$, $L_{n,w}^{\prime }$, $L_{nT,w}^{\prime }$) is the shifted reference read at 500 Hz; for octave bands it is then reduced by 5 dB. The spectrum adaptation term $C_I=L_{n,\text{sum}}-15-L_{n,w}$ (ISO 717-2:2020 Formulae (A.1) to (A.3)) uses the energetic sum over 100–2500 Hz (the first 15 thirds, excluding 3150 Hz) or 125–2000 Hz (octaves).

The $-15$ dB is a normalising offset, not physics: it is chosen so that $C_I$ comes out about zero for a massive floor with an effective covering, slightly positive for a timber joist floor whose low-frequency peaks dominate, and down to $-15$ dB for a bare or poorly covered concrete floor (Annex A.1). Read that way, the $C_I=-11$ dB of the example below says the measured spectrum is far flatter and more high-frequency-weighted than the reference contour — a bare slab, not a floor with a covering — so the floor is better at the low frequencies footsteps excite than $L_{nT,w}^{\prime }=79$ dB alone suggests. The corollary is a design trap: a soft covering acts where the reference curve is most demanding, so it improves the single number more than it improves what the neighbour hears. The unweighted sum $C_I$ is built on exists precisely because it correlates better with A-weighted walking noise than the reference-curve rating does, which is why requirements are sometimes written on $L_{n,w}^{\prime }+C_I$ rather than on $L_{n,w}^{\prime }$.

For measurements extended down to 50 Hz, weighted_impact_rating_extended additionally returns the enlarged-range term $C_{I,50\text{–}2500}$ (A.2.1 NOTE), and with one_decimal=True the 0.1 dB-step rating used in uncertainty statements (it reproduces the printed $L_{n,r,0,w}=77.6$ dB and $C_{I,r,0}=-10.3$ dB of A.2.2).

The same fitting rule with the sign reversed: the shaded area is where the measurement rises above the reference, and the curve is shifted up until that area reaches 28.0 dB against the same 32.0 dB cap. A floor rated this way is worse the higher its number.

import numpy as np
from phonometry import building

# 16 one-third-octave impact levels (100 Hz - 3150 Hz), dB, from the
# ISO 717-2 Annex C worked example; measured with T = T0, so these are
# already the standardized L'nT values.
l_nt = np.array([62.1, 63.2, 63.5, 66.2, 68.5, 70.0, 71.7, 73.1,
                 73.8, 73.5, 73.8, 73.3, 73.1, 73.0, 72.4, 71.2])

# Weighted impact rating + spectrum adaptation term CI (ISO 717-2)
res_imp = building.weighted_impact_rating(l_nt)
print(res_imp.rating, res_imp.ci, res_imp.unfavourable_sum)   # 79 -11 28.0  ->  L'nT,w(CI)=79(-11)

# Octave-band data carry the extra -5 dB reduction (Clause 4.3.2)
octave = np.array([65.3, 64.5, 58.0, 55.8, 43.0])
print(building.weighted_impact_rating(octave).rating)  # 54

res_imp.plot()   # measured L'nT vs shifted ISO 717-2 reference, measured-above shaded (needs matplotlib)

Show the code for this figure

import matplotlib.pyplot as plt
import numpy as np
from phonometry import building

# 16 standardized one-third-octave impact levels L'nT, dB (ISO 717-2
# Annex C worked example, measured with T = T0).
l_nt = np.array([62.1, 63.2, 63.5, 66.2, 68.5, 70.0, 71.7, 73.1,
               73.8, 73.5, 73.8, 73.3, 73.1, 73.0, 72.4, 71.2])
# Weighted impact rating + spectrum adaptation term CI (ISO 717-2)
res_imp = building.weighted_impact_rating(l_nt)

# One line — measured L'nT vs the shifted ISO 717-2 reference (measured-above shaded):
res_imp.plot()
plt.show()

# By hand, from the band curve the result now carries (note the opposite sign:
# an unfavourable deviation is where the MEASURED level exceeds the reference).
# Here the input was l_n_t, so the rated quantity is the field level L'nT,w:
fig, ax = plt.subplots()
ax.semilogx(res_imp.band_centers, res_imp.measured, "o-", label="Measured L'nT")
ax.semilogx(res_imp.band_centers, res_imp.shifted_reference, "s--", label="Shifted reference")
ax.fill_between(res_imp.band_centers, res_imp.shifted_reference, res_imp.measured,
                where=res_imp.measured > res_imp.shifted_reference, interpolate=True,
                alpha=0.3, label="Unfavourable deviations")
ax.set_xlabel("Frequency [Hz]")
ax.set_ylabel("Impact sound pressure level [dB]")
ax.set_title(f"L'nT,w = {res_imp.rating} dB  (CI={res_imp.ci:+d})")
ax.legend()
plt.show()

Feeding the standardized spectrum into weighted_impact_rating reproduces the ISO 717-2 Annex C values (thirds $L_{nT,w}^{\prime }=79$, $C_I=-11$; octave 54, $C_I=0$).

weighted_impact_rating() parameters

Parameter	Type	Units	Range / default	Notes
values_by_band	1D array	dB	16 (thirds) or 5 (octaves)	Measured $L_n$, $L_n^{\prime }$ or $L_{nT}^{\prime }$ per band
bands	str or None	—	'third-octave' / 'octave' / None	None infers from the count

weighted_impact_rating() returns an ImpactRatingResult (rating, ci integers, unfavourable_sum in dB), with the same three plotting fields as the airborne result: band_centers, measured and shifted_reference.

Enlarged frequency ranges and one-decimal ratings

When the measurement covers more than the core 100–3150 Hz bands, ISO 717-1 Annex B defines additional adaptation terms with the range as a subscript ($C_{50\text{–}3150}$, $C_{50\text{–}5000}$, $C_{100\text{–}5000}$ and the $C_{tr}$ counterparts), computed with the Table B.1 spectra over the enlarged range. weighted_rating_extended takes the band values with their centre frequencies and returns the core rating plus every extended term the input covers (the impact counterpart weighted_impact_rating_extended adds $C_{I,50\text{–}2500}$). With one_decimal=True the reference curve shifts in 0.1 dB steps and all reductions keep one decimal: the variant ISO 717 prescribes “for the expression of uncertainty” and ISO 12999-1 Annex B requires when stating the uncertainty of a single-number value.

from phonometry import building

# Single-number rating from a measured 16-band R spectrum (ISO 717-1 Annex C)
R = [20.4, 16.3, 17.7, 22.6, 22.4, 22.7, 24.8, 26.6,
     28.0, 30.5, 31.8, 32.5, 33.4, 33.0, 31.0, 25.5]

freqs = [50, 63, 80, 100, 125, 160, 200, 250, 315, 400, 500,
         630, 800, 1000, 1250, 1600, 2000, 2500, 3150, 4000, 5000]
r_ext = [18.7, 19.2, 20.0, *R, 26.8, 29.2]     # ISO 717-1 Annex C, Table C.2
ext = building.weighted_rating_extended(r_ext, freqs)
print(ext.rating, ext.c, ext.ctr, ext.c_50_5000, ext.ctr_50_5000)
# 30 -2 -3 -2 -4   ->  Rw(C;Ctr;C50-5000;Ctr,50-5000) = 30(-2;-3;-2;-4)

one_dp = building.weighted_rating_extended(r_ext, freqs, one_decimal=True)
print(one_dp.rating)   # 30.0, the 0.1 dB-step rating for uncertainty statements

ext.plot()   # enlarged-range curve vs the shifted core reference, Annex B terms in the title (needs matplotlib)

The rating itself is still evaluated on the 16 core bands (100–3150 Hz); the enlarged bands only enter the Annex B adaptation terms, so the shifted reference curve stops at the core-range edges while the measured curve continues into the shaded enlarged range.

Show the code for this figure

import matplotlib.pyplot as plt
from phonometry import building

# Single-number rating from a measured 16-band R spectrum (ISO 717-1 Annex C)
R = [20.4, 16.3, 17.7, 22.6, 22.4, 22.7, 24.8, 26.6,
     28.0, 30.5, 31.8, 32.5, 33.4, 33.0, 31.0, 25.5]
freqs = [50, 63, 80, 100, 125, 160, 200, 250, 315, 400, 500,
         630, 800, 1000, 1250, 1600, 2000, 2500, 3150, 4000, 5000]
r_ext = [18.7, 19.2, 20.0, *R, 26.8, 29.2]     # ISO 717-1 Annex C, Table C.2
ext = building.weighted_rating_extended(r_ext, freqs)

# One line — the enlarged-range curve vs the shifted core reference:
ext.plot()
plt.show()

# By hand, from the band curves the result carries (the full enlarged-range
# curve on ext, the core-band reference on ext.core):
fig, ax = plt.subplots()
ax.semilogx(ext.band_centers, ext.measured, "o-", label="Measured R")
ax.semilogx(ext.core.band_centers, ext.core.shifted_reference, "s--",
            label="Shifted reference (core bands)")
ax.fill_between(ext.core.band_centers, ext.core.measured,
                ext.core.shifted_reference,
                where=ext.core.measured < ext.core.shifted_reference,
                interpolate=True, alpha=0.3, label="Unfavourable deviations")
ax.set_xlabel("Frequency [Hz]")
ax.set_ylabel("Sound reduction index [dB]")
ax.set_title(f"Rw = {ext.rating} dB  (C50-5000={ext.c_50_5000:+g}; "
             f"Ctr,50-5000={ext.ctr_50_5000:+g})")
ax.legend()
plt.show()

ISO 717 report (.report())

Both rating results render a one-page PDF fiche laid out like an accredited-laboratory test report through a report(path) method: a standard-basis line (measurement standard plus the ISO 717 rating part), an optional metadata header block, the one-third-octave table beside the measured-versus-shifted-reference plot (the result’s own .plot()), the boxed single-number result, an optional verdict row and a footer with the fixed disclaimer. WeightedRatingResult.report() labels the airborne ISO 717-1 fiche ($R_w(C;C_{tr})$, deviations where the reference is above the measurement); ImpactRatingResult.report() labels the impact ISO 717-2 fiche ($L_{n,w}(C_I)$, deviations the opposite way). SoundReductionResult.report() is a convenience that rates the predicted $R(f)$ and writes its fiche in one call; that result comes from the EN 12354 panel models of Panel sound insulation, so it is the one entry point here whose input was never measured.

The report metadata is supplied as a ReportMetadata frozen dataclass (every field optional; only the supplied fields are rendered, and the numeric fields must satisfy their field-specific physical ranges, detailed in the table below). Passing metadata=None produces a lightweight prediction fiche (body, result and disclaimer only). When metadata.requirement is set, a verdict row is added: an airborne rating passes when it is at or above the requirement, an impact rating when it is at or below it (a lower impact level is better). Setting verbose=True swaps the two-column f | value table for the ISO 717 Annex C columns (frequency, measured value, shifted reference, unfavourable deviation).

Rendering needs reportlab, kept out of the runtime dependencies as the optional phonometry[report] extra (pip install phonometry[report]); a missing reportlab raises a clear ImportError with the install command, and the plot still needs matplotlib (phonometry[plot]). Only engine="reportlab" is supported; any other engine raises ValueError. The fiche renders in English by default; pass language="es" for a Spanish fiche (translated fixed strings and a comma decimal separator), e.g. building.weighted_rating(R).report("Rw_fiche_es.pdf", language="es").

from phonometry import building, ReportMetadata

# Airborne rating from a measured 16-band R spectrum (ISO 717-1)
R = [20.4, 16.3, 17.7, 22.6, 22.4, 22.7, 24.8, 26.6,
     28.0, 30.5, 31.8, 32.5, 33.4, 33.0, 31.0, 25.5]
metadata = ReportMetadata(
    specimen="200 mm reinforced-concrete wall",
    client="Acoustic Test Client Ltd.",
    area=10.0, mass_per_area=460.0,
    source_volume=53.0, receiving_volume=51.0,
    temperature=21.5, relative_humidity=45.0, pressure=101.3,
    test_room="Transmission suite T1",
    measurement_standard="ISO 10140-2",
    test_date="2026-07-18",
    laboratory="Phonometry Reference Laboratory",
    operator="José Manuel Requena Plens",
    report_id="PHN-2026-0042",
    requirement=42.0,          # adds the PASS/FAIL verdict row
)
building.weighted_rating(R).report(
    "Rw_fiche.pdf", metadata=metadata
)                                                           # Rw (C; Ctr)

# Impact rating from a measured 16-band L'nT spectrum (ISO 717-2)
l_nt = [45.0, 47.0, 48.0, 49.0, 51.0, 52.0, 53.0, 54.0,
        55.0, 56.0, 57.0, 58.0, 55.0, 52.0, 49.0, 46.0]
building.weighted_impact_rating(l_nt).report("Lnw_fiche.pdf")  # Ln,w (CI)

Both example fiches are regenerated with make reports and kept rendered in the repository; click either preview to open the PDF.

One-page airborne sound insulation fiche: a metadata header (client, specimen, sample area, room volumes, per-room temperature and humidity, ambient pressure and mounting), the one-third-octave R table beside the measured-versus-shifted-reference plot, the boxed Rw (C; Ctr) single-number result and a PASS verdict against the 30 dB requirement.

Download the report (PDF)

Airborne rating fiche (WeightedRatingResult.report), Rw (C; Ctr).

One-page impact sound insulation fiche in the same accredited layout for the normalized impact sound pressure level Ln: the metadata header, the one-third-octave Ln table beside the measured-versus-shifted-reference plot with the 500 Hz read-off, the boxed Ln,w (CI) single-number result and a FAIL verdict against the 53 dB requirement (a lower impact level is better).

Download the report (PDF)

Impact rating fiche (ImpactRatingResult.report), Ln,w (CI).

Report metadata (ReportMetadata)

Every field is optional and only the supplied ones are rendered, so the same object drives a full accredited fiche and a lightweight prediction fiche. The numeric fields are validated on construction by physical range.

Field	Type	Rendered as
specimen, client, mounted_by, manufacturer	str	Header identity of the tested element and who it was tested for / mounted by
area, mass_per_area	float > 0	Sample area $S$ (m²) and measured mass per unit area (kg/m²)
source_volume, receiving_volume	float > 0	Room volumes (m³)
temperature, relative_humidity	float	Single representative climate: air temperature (°C, any sign), relative humidity (0–100 %)
pressure	float > 0	Ambient (static) air pressure during the test (kPa)
source_temperature, source_relative_humidity, receiving_temperature, receiving_relative_humidity	float	Per-room climate when source and receiving rooms are reported separately (same ranges as above)
test_room, mounting, measurement_standard, test_date	str	Facility, mounting condition, the measurement standard (forms the standard-basis line) and the test date
laboratory, operator, report_id, notes	str	Footer: institute, operator signature line, report number and free-form remarks
requirement	float	Target single number; adds the verdict row (airborne passes at or above it, impact at or below it)

What this guide covers

• Covered

  ISO 717-1:2020’s reference-curve airborne rating (the 1 dB shifting rule with the 32.0 dB / 10.0 dB unfavourable-deviation caps, read at 500 Hz) with the Annex A spectrum adaptation terms $C$ and $C_{tr}$, via building.weighted_rating; ISO 717-2:2020’s impact rating with the opposite deviation sign, the octave −5 dB reduction and the $C_I$ term, via building.weighted_impact_rating; the Annex B / A.2.1 enlarged-range terms and the 0.1 dB-step one-decimal variants via building.weighted_rating_extended and building.weighted_impact_rating_extended; and the ISO 717 fiche through .report() with the shared ReportMetadata container. Conformance is anchored on the standards’ own Annex C worked examples.

• Not covered

  Nothing on this page measures: every function takes an already-measured (or predicted) band spectrum from the field, laboratory, intensity, survey or flanking guides. The façade single number of ISO 16283-3 Annex F and the flanking $D_{n,f,w}$ reuse these engines from their own guides rather than duplicating them. One member of the ISO 717 family is deliberately absent: the A-weighted maximum impact level of ISO 717-2:2020 Annex D, which rates the rubber ball and the bang machine. It shifts no curve at all — it is an energy sum of A-weighted band levels — so it is not produced by either engine on this page and lives with the sources it rates, in Heavy and Soft Impact Sources.

See also

• Field Insulation Measurement (ISO 16283): the airborne, impact and façade spectra these engines rate in the field.
• Laboratory Insulation Measurement: the laboratory $R$ and $L_n$ behind $R_w$ and $L_{n,w}$.
• Sound Insulation by Intensity (ISO 15186): the intensity indices rated with the same airborne engine.
• Sound Insulation Survey Method (ISO 10052): the survey quantities and their automatic ratings.
• Laboratory Flanking Transmission (ISO 10848): the flanking descriptors $D_{n,f,w}$ and $L_{n,f,w}$.
• Floor-Covering Impact Improvement (ISO 16251-1): the weighted improvement $\Delta L_w$ built on the ISO 717-2 reference floor.
• Predicting Sound Insulation (EN 12354): the single-number model that consumes these ratings.
• Theory: the reference-curve derivation behind the weighted single-number ratings.
• API reference: building.measurement.insulation.

Quick answers

What is the difference between the C and Ctr spectrum adaptation terms?

Both re-rate the same measured curve against the two source spectra of ISO 717-1 Annex A: $C$ against A-weighted pink noise (living activities: speech, music, radio, television) and $C_{tr}$ against A-weighted urban road traffic, whose energy sits at low frequency. $C$ is typically 0 to −2 dB, while a lightweight double leaf with a mass-air-mass resonance near 100 Hz can carry a $C_{tr}$ of −5 to −10 dB.

References

• Hopkins, C. (2007). Sound insulation. Butterworth-Heinemann. https://doi.org/10.4324/9780080550473The interpretation of the single-number ratings and what they hide about the underlying band spectra. ISBN 978-0-7506-6526-1.
• International Organization for Standardization. (2020). Acoustics — Rating of sound insulation in buildings and of building elements — Part 1: Airborne sound insulation (ISO 717-1:2020). The reference-curve rating and the spectrum adaptation terms C and Ctr interpreted above; also the single-number engine for the façade quantity (Annex F).
• International Organization for Standardization. (2020). Acoustics — Rating of sound insulation in buildings and of building elements — Part 2: Impact sound insulation (ISO 717-2:2020). The impact reference-curve rating and the spectrum adaptation term CI, with the enlarged-range CI,50-2500 term.