# building.measurement.floor_covering_improvement

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

Canonical: https://docs.jmrp.io/phonometry/reference/api/building/floor-covering-improvement/
Language: en
Version: phonometry 4.0.0
Install: pip install "phonometry[full]"
Author: José Manuel Requena Plens (https://orcid.org/0000-0003-1250-6212)
Published: 2026-07-15
Updated: 2026-11-16T21:36:14Z
License: MIT (https://github.com/jmrplens/phonometry/blob/main/LICENSE)
Summary: Impact-sound improvement of floor coverings on a small mock-up (ISO 16251-1:2014).

---

Impact-sound improvement of floor coverings on a small mock-up
(ISO 16251-1:2014).

Laboratory method for the **improvement of impact sound insulation** `ΔL` of a
soft, locally-reacting floor covering (carpet, PVC, linoleum; ISO 10140-1
Annex H category I) laid on a small concrete mock-up plate and excited by a
standard tapping machine. The two rooms of the ISO 10140 series are removed and
the floor is replaced by a softly-supported concrete plate; instead of the
receiving-room sound pressure level, the **structure-borne acceleration level**
on the underside of the plate is measured with and without the covering. For
locally-reacting coverings this acceleration-level difference equals the ISO
10140 impact sound reduction (Clause 4).

**Acceleration level (Formula (1)).**
$L_a = 10 \log_{10}[(1/T_m) \int a(t)^2/a_0^2\,dt]$ dB,
with the reference acceleration $a_0 = 10^{-6}$ m/s², i.e.
$L_a = 10 \log_{10}(\overline{a^2}/a_0^2)$.

**Background correction (Formula (2)).** Each measured level `L'` is
corrected against the background `Lb` per accelerometer position, by the
margin $L' - L_\mathrm{b}$: unchanged for $\ge 15$ dB; energy
subtraction $10 \log_{10}(10^{L'/10} - 10^{L_\mathrm{b}/10})$ for
$6 \le \mathrm{margin} < 15$ dB; and the fixed $L' - 1.3$ dB limit
for $< 6$ dB. Bands hitting the 1.3 dB limit are flagged as the *limit of
measurement* (reported as $> \Delta L$). This differs from the ISO
10140-4 correction ([`phonometry.building.background_correction`](https://docs.jmrp.io/phonometry/reference/api/building/laboratory-insulation/#background_correction)) only at
exactly $\mathrm{margin} = 6$ dB.

**Improvement (Formulae (3)/(4)).** The per-position difference is
$\Delta L_{t,a} = L_{0,t,a} - L_{1,t,a}$ (0 = bare plate, 1 = specimen)
and the improvement is
their arithmetic mean over all tapping-machine (t) and accelerometer (a)
positions, $\Delta L = (1/(t \cdot a)) \sum_t \sum_a \Delta L_{t,a}$.

**Octave bands (Formula (5)).**
$\Delta L_\mathrm{oct} = -10 \log_{10}[(1/3) \sum 10^{-\Delta L_j/10}]$ dB from
the three one-third-octave values in each octave.

**Weighted improvement.** `ΔLw` is the ISO 717-2 weighted reduction of impact
sound pressure level (Clause 6.5), computed with the heavyweight reference
floor via [`phonometry.building.weighted_impact_improvement`](https://docs.jmrp.io/phonometry/reference/api/building/ratings/#weighted_impact_improvement) on the 16
rating bands 100 Hz to 3150 Hz; a wider clause 6.3 spectrum (18 bands 100-5000
Hz, optionally extended to 50 Hz) is rated on that sub-range. The statement of
results (Clause 8 e)) also carries the spectrum adaptation term `CI,Δ` (ISO
717-2:2020 Formula (A.4)) via
[`phonometry.building.impact_improvement_adaptation_term`](https://docs.jmrp.io/phonometry/reference/api/building/ratings/#impact_improvement_adaptation_term).

> Auto-generated from the source docstrings by `scripts/generate_api_docs.py` (`make api-docs`). Do not edit by hand.

## acceleration_level

```python
acceleration_level(
    acceleration: float | np.floating[Any] | np.integer[Any],
    *,
    reference: float = ...,
) -> float

acceleration_level(
    acceleration: Sequence[float] | Sequence[np.floating[Any]] | Sequence[np.integer[Any]] | NDArray[np.bool_ | np.integer[Any] | np.floating[Any]],
    *,
    reference: float = ...,
) -> NDArray[np.float64]

acceleration_level(
    acceleration: NumericArrayLike,
    *,
    reference: float = ...,
) -> float | NDArray[np.float64]
```

Vibratory acceleration level `La` (ISO 16251-1 Formula (1)).

$L_a = 10 \log_{10}(a_\mathrm{rms}^2 / a_0^2) = 20 \log_{10}(a_\mathrm{rms} / a_0)$ dB.

**Parameters**

| Name | Description |
| :--- | :--- |
| `acceleration` | RMS acceleration `a_rms` per band, in m/s² (> 0). |
| `reference` | Reference acceleration `a0`, in m/s² (default 1e-6). |

**Returns:** The acceleration level `La` per band, in dB. A `float` for a number, an array for a sequence or an array, a zero-dimensional one included.

**Raises**

| Exception | When |
| :--- | :--- |
| ValueError | Non-finite or non-positive acceleration, or a non-positive reference. |
| TypeError | if `reference` is not a single number, or `acceleration` holds a text or a value that is not a real number. |

## background_corrected_level

```python
background_corrected_level(
    signal_and_background: NumericArrayLike,
    background: NumericArrayLike,
    *,
    frequencies: NumericArrayLike | None = None,
) -> BackgroundCorrectedLevel
```

Correct a measured level for background noise (ISO 16251-1 F. (2)).

**Parameters**

| Name | Description |
| :--- | :--- |
| `signal_and_background` | Measured level `L'` (with background) per band, in dB; a (positions, bands) array is corrected position by position. |
| `background` | Background-noise level `Lb` per band, in dB, of the same shape. |
| `frequencies` | The band centre frequencies, in Hz, one per band, for the plot to name the bands by; `None` (the default) numbers them. |

**Returns:** The correction, as a [`BackgroundCorrectedLevel`](https://docs.jmrp.io/phonometry/reference/api/building/floor-covering-improvement/#backgroundcorrectedlevel): the corrected levels `L` in its `corrected_levels`, and in its `limited` the bands at the 1.3 dB limit of measurement (report as $> \Delta L$).

**Raises**

| Exception | When |
| :--- | :--- |
| ValueError | Mismatched shapes, non-finite values, or band frequencies that are not one positive value per band. |
| TypeError | if `background`, `frequencies` or `signal_and_background` holds a text or a value that is not a real number. |

## BackgroundCorrectedLevel

```python
BackgroundCorrectedLevel(
    signal_and_background: NDArray[np.float64],
    background: NDArray[np.float64],
    *,
    frequencies: NDArray[np.float64] | None = None,
)
```

A measured level corrected for background noise (ISO 16251-1 F. (2)).

At a margin $L' - L_\mathrm{b}$ of 15 dB or more the level is used
unchanged; between 6 dB and 15 dB the background is subtracted in energy;
under 6 dB the correction is held at 1.3 dB, the limit of measurement,
and the band is reported as $> \Delta L$. The margins, the corrected
levels and the bands at the limit are read from the two readings, so they
are not fields: a record cannot be built to claim a band clear of the
limit that its readings put on it.

**Attributes**

| Name | Description |
| :--- | :--- |
| `signal_and_background` | Measured level `L'` (with background), in dB; one value per band, or a (positions, bands) array. |
| `background` | Background-noise level `Lb`, in dB, of the same shape. |
| `frequencies` | The band centre frequencies, in Hz, one per band (the last axis of the readings), or `None` where the bands were not named; only the plot reads them. |

### BackgroundCorrectedLevel.corrected_levels

*property*

The corrected levels `L`, in dB, in the shape of the readings.

### BackgroundCorrectedLevel.limited

*property*

Where the margin is under 6 dB, so the band is at the limit of measurement.

Judged settled: 32.3 - 26.3 dB is 5.999 999 999 999 996 in binary and
still the 6 dB it is.

### BackgroundCorrectedLevel.margins_db

*property*

$L' - L_\mathrm{b}$, in dB, in the shape of the readings.

### BackgroundCorrectedLevel.plot()

```python
BackgroundCorrectedLevel.plot(
    ax: Axes | None = None,
    *,
    language: Literal['en', 'es'] = 'en',
    **kwargs: Any,
) -> Axes
```

Plot the margin $L' - L_\mathrm{b}$ of each band against the criterion.

The bands at 15 dB or more are used unchanged, those between 6 dB
and 15 dB have the background subtracted in energy, and those under
6 dB are held at the 1.3 dB limit of measurement; the two boundaries
are drawn across the bands and the three regimes are told apart. A
(positions, bands) record draws one line per position.

Requires matplotlib (the `[plot]` extra).

**Parameters**

| Name | Description |
| :--- | :--- |
| `ax` | Existing axes, or `None` to create a figure. |
| `language` | Label language, `"en"` (default) or `"es"`. |
| `kwargs` | Forwarded to the margin bars, or to the line of each position. |

**Returns:** The axes.

## FloorCoveringImprovementResult

```python
FloorCoveringImprovementResult(
    frequencies: NDArray[np.float64],
    improvement: NDArray[np.float64],
    delta_lw: int | None,
    ci_delta: int | None = None,
    background_margin_db: NDArray[np.float64] | None = None,
)
```

Impact-sound improvement of a floor covering (ISO 16251-1).

**Attributes**

| Name | Description |
| :--- | :--- |
| `frequencies` | One-third-octave band centre frequencies, in Hz. |
| `improvement` | Improvement of impact sound insulation `ΔL` per band, in dB. |
| `delta_lw` | Weighted improvement `ΔLw` (ISO 717-2), in dB, or `None` when the spectrum does not contain the 16 one-third-octave rating bands 100-3150 Hz. A wider clause 6.3 spectrum (e.g. the 18 bands 100-5000 Hz, optionally extended down to 50 Hz) is rated on its 100-3150 Hz sub-range. |
| `ci_delta` | Spectrum adaptation term `CI,Δ` (ISO 717-2:2020 Formula (A.4); required in the ISO 16251-1 Clause 8 e) statement of results), in dB, or `None` when `delta_lw` is `None`. |
| `background_margin_db` | Per band, the smallest margin $L' - L_\mathrm{b}$ of a measured level over the background, over every position and both floors (Formula (2)), in dB; `None` when no background correction was applied. `limited` is read from it and the 6 dB of Formula (2), so it is not a field. |

### FloorCoveringImprovementResult.limited

*property*

Per band, whether a position reached the 1.3 dB limit of measurement.

`True` where the smallest margin over the background fell below the
6 dB of Formula (2), so the band is reported as $> \Delta L$; all
`False` when no background correction was applied. The margin is
judged settled, as Formula (2) itself is.

### FloorCoveringImprovementResult.octave_bands()

```python
FloorCoveringImprovementResult.octave_bands() -> OctaveBandSpectrum
```

`ΔLoct` in octave bands, via Formula (5) (needs whole octave triplets).

### FloorCoveringImprovementResult.plot()

```python
FloorCoveringImprovementResult.plot(
    ax: Axes | None = None,
    *,
    language: Literal['en', 'es'] = 'en',
    **kwargs: Any,
) -> Axes
```

Plot the improvement spectrum `ΔL` (with `ΔLw` when available).

Requires matplotlib (the `[plot]` extra); returns the
`Axes`.

**Parameters**

| Name | Description |
| :--- | :--- |
| `language` | The language of the labels: `"en"` or `"es"`. |

### FloorCoveringImprovementResult.report()

```python
FloorCoveringImprovementResult.report(
    path: str | os.PathLike[str],
    *,
    metadata: ReportMetadata | None = None,
    engine: Literal['reportlab'] = 'reportlab',
    verbose: bool = False,
    language: Literal['en', 'es'] = 'en',
) -> str
```

Render an ISO 16251-1 impact-improvement test-report fiche to PDF.

Writes a one-page accredited report (BS EN ISO 16251-1:2014, the
small-mock-up laboratory method for the reduction of transmitted impact
sound by soft floor coverings): the standard-basis line, an optional
metadata header block, a two-panel body with the per-band table
(frequency and the reduction of impact sound pressure level `ΔL`,
bands at the 1.3 dB limit of measurement prefixed `>`) beside the
`ΔL(f)` improvement curve, a boxed single-number weighted improvement
`ΔLw (CI,Δ)` (ISO 717-2, the Clause 8 e) statement of results; a
characterisation headline replaces it when the spectrum lacks the 16
rating bands 100-3150 Hz), an optional verdict row and a footer.

**Parameters**

| Name | Description |
| :--- | :--- |
| `path` | Destination path of the PDF file. |
| `metadata` | Optional [`ReportMetadata`](https://docs.jmrp.io/phonometry/reference/api/building/insulation/#reportmetadata); `None` produces a body-and-disclaimer fiche whose header shows only the measured frequency range. The applicable descriptive fields are `client`, `manufacturer`, `specimen` (the floor covering under test), `mounting`, `mass_per_area`, `test_room`, `test_date`, `temperature_c`, `pressure`, `measurement_standard`, `laboratory`, `operator`, `report_id` and `notes`. The bare reference floor is the standardised heavyweight floor of ISO 717-2:2020 Table 4, fixed by the standard. When `requirement` is set the fiche adds a verdict row (a higher weighted improvement is better, so the result passes at or above the requirement). |
| `engine` | Rendering back end; only `"reportlab"` is supported. |
| `verbose` | When `True`, the per-band table gains the reference-floor-with-covering column $L_\mathrm{n,r} = L_\mathrm{n,r,0} - \Delta L$ (ISO 717-2:2020 Formula (1)), the derivation basis of `ΔLw`, when the spectrum is exactly the 16 rating bands 100-3150 Hz. |
| `language` | Fiche language: `"en"` (default, English, decimal point) or `"es"` (Spanish, decimal comma). |

**Returns:** The written `path` as a `str`.

**Raises**

| Exception | When |
| :--- | :--- |
| ValueError | If `engine` is not `"reportlab"`. |
| ImportError | If reportlab is not installed (the `[report]` extra), or matplotlib is missing for the embedded figure (the `[plot]` extra). |

## impact_improvement

```python
impact_improvement(
    bare: NumericArrayLike,
    with_covering: NumericArrayLike,
    *,
    frequencies: NumericArrayLike,
    background: NumericArrayLike | None = None,
) -> FloorCoveringImprovementResult
```

Improvement of impact sound insulation `ΔL` of a floor covering.

ISO 16251-1. The acceleration levels without (`L0`) and with (`L1`)
the specimen may be
given either as one value per band (already averaged over the
tapping-machine
and accelerometer positions) or, for the raw measurement, as a
`(positions, bands)` array. The standard's order of operations is
followed:
the background correction (Formula (2)) is applied **per position**, then
the
level difference $\Delta L_{t,a} = L_0 - L_1$ (Formula (3)), then
the arithmetic mean
over positions $\Delta L = \mathrm{mean}(\Delta L_{t,a})$
(Formula (4)), necessary because Formula
(2) is non-linear, so correcting must precede averaging.

**Parameters**

| Name | Description |
| :--- | :--- |
| `bare` | Bare-plate acceleration level `L0`, in dB; `(bands,)` or `(positions, bands)`. |
| `with_covering` | Level with the specimen `L1`, in dB; same shape as `bare`. |
| `frequencies` | One-third-octave band centre frequencies, in Hz (`(bands,)`). |
| `background` | Optional background-noise level `Lb`, in dB; `(bands,)` (applied to every position) or `(positions, bands)`. When given, `L0` and `L1` are background-corrected (Formula (2)) per position before the difference, and bands where any position hits the 1.3 dB limit are flagged. |

**Returns:** [`FloorCoveringImprovementResult`](https://docs.jmrp.io/phonometry/reference/api/building/floor-covering-improvement/#floorcoveringimprovementresult) (`ΔL` per band).

**Raises**

| Exception | When |
| :--- | :--- |
| ValueError | Mismatched shapes or non-finite values. |
| TypeError | if `background`, `bare`, `frequencies` or `with_covering` holds a text or a value that is not a real number. |

## improvement_octave_bands

```python
improvement_octave_bands(
    improvement: NumericArrayLike,
    *,
    frequencies: NumericArrayLike,
    quantity: Literal['ΔL', 'ΔR', 'Rs'] = 'ΔL',
) -> OctaveBandSpectrum
```

Convert a one-third-octave improvement spectrum to octaves (F. (5)).

$\Delta L_\mathrm{oct} = -10 \log_{10}[(1/3) \sum_{j=1}^{3} 10^{-\Delta L_j/10}]$
dB over the three thirds of
each octave.

**Parameters**

| Name | Description |
| :--- | :--- |
| `improvement` | Improvement `ΔL` per one-third-octave band, in dB. |
| `frequencies` | The matching one-third-octave centre frequencies, in Hz; must contain whole octave triplets (centre / centre·2^(±1/3)). |
| `quantity` | What the values are, for the record and its plot: `"ΔL"` (Default), `"ΔR"` or `"Rs"`, the three quantities folded into octaves by this average of transmitted power. The rainfall sound intensity levels, the `"LI"` and `"LI,norm"` an [`OctaveBandSpectrum`](https://docs.jmrp.io/phonometry/reference/api/building/floor-covering-improvement/#octavebandspectrum) holds too, are not offered: ISO 10140-1 Formula (K.3) sums their energy, which [`RainfallSoundResult.octave_bands`](https://docs.jmrp.io/phonometry/reference/api/building/rainfall-sound/#rainfallsoundresultoctave_bands) does. |

**Returns:** An [`OctaveBandSpectrum`](https://docs.jmrp.io/phonometry/reference/api/building/floor-covering-improvement/#octavebandspectrum) with the octave centres (`.frequencies`) and the octave values (`.values`).

**Raises**

| Exception | When |
| :--- | :--- |
| ValueError | The thirds do not group into complete octave triplets, or `quantity` is not one of the three, `"LI"` and `"LI,norm"` among them. |
| TypeError | if `quantity` is not a text, or `frequencies` or `improvement` holds a text or a value that is not a real number. |

## OctaveBandSpectrum

```python
OctaveBandSpectrum(
    frequencies: NDArray[np.float64],
    values: NDArray[np.float64],
    quantity: Literal['ΔL', 'ΔR', 'Rs', 'LI', 'LI,norm'],
)
```

A one-third-octave result folded into octave bands, with what it holds.

What [`improvement_octave_bands`](https://docs.jmrp.io/phonometry/reference/api/building/floor-covering-improvement/#improvement_octave_bands) returns, and what the
`octave_bands()` methods of the floor-covering, laboratory lining,
laboratory floor-covering, joint and rainfall results return: the octave
centres and the octave values side by side, named, so that neither can be
read as the other, and the quantity they are values of.

**Attributes**

| Name | Description |
| :--- | :--- |
| `frequencies` | The octave centre frequencies, in hertz. |
| `values` | The octave value of `quantity` in each band, in dB. |
| `quantity` | Which quantity the values are: `"ΔL"` (an impact improvement), `"ΔR"` (a sound reduction improvement), `"Rs"` (the sound reduction index of a joint), `"LI"` or `"LI,norm"` (the rainfall sound intensity level, raw or normalized). |

### OctaveBandSpectrum.plot()

```python
OctaveBandSpectrum.plot(
    ax: Axes | None = None,
    *,
    language: Literal['en', 'es'] = 'en',
    **kwargs: Any,
) -> Axes
```

Plot the octave values against their octave bands.

Requires matplotlib (the `[plot]` extra); returns the
`Axes`.

**Parameters**

| Name | Description |
| :--- | :--- |
| `ax` | Existing axes, or `None` to create a figure. |
| `language` | Label language, `"en"` (default) or `"es"`. |
| `kwargs` | Forwarded to the octave curve `plot` call. |
