Sage Journals: Discover world-class research

Abstract

Acoustic comfort has been used in engineering to refer to conditions of low noise levels or annoyance, while current standardized methods for airborne and impact sound reduction are used to assess acoustic comfort in dwellings. However, the results and descriptors acquired from acoustic measurements do not represent the human perception of sound or comfort levels. This article is a review of laboratory studies concerning airborne sound in dwellings. Specifically, this review presents studies that approach acoustic comfort via the association of objective and subjective data in laboratory listening tests, combining airborne sound acoustic data, and subjective ratings. The presented studies are tabulated and evaluated using Bradford Hill’s criteria. Many of them attempt to predict subjective noise annoyance and find the best single number quantity for that reason. The results indicate that subjective response to airborne sound is complicated and varies according to different sound stimuli. It can be associated sufficiently with airborne sound in general but different descriptors relate best to music sounds or speech stimuli. The inclusion of low frequencies down to 50 Hz in the measurements seems to weaken the association of self-reported responses to airborne sound types except for the cases of music stimuli.

Keywords

Acoustic comfort airborne sound laboratory subjective responses association evaluation

Introduction

This is the third and final part of a review of acoustic comfort evaluation in dwellings. It is accompanying part I, which reviewed subjective responses to field data from building surveys¹ and part II, which reviewed subjective responses to impact sound data in laboratory tests.² This article is focused on subjective responses relevant to airborne sound data used in laboratory tests.

The presented studies of this review approach acoustic comfort through the association of acoustic data and subjective responses: they analyze laboratory listening tests that utilize airborne sound reduction data from measurements and sometimes involve recorded sounds of various noise types. The acoustic data are utilized in controlled listening experiments where the subjects (participants) provide their self-reported responses.^3–13 In some cases, the acoustic data come from in situ measurements or sound recordings in test buildings. However, the data are still processed and used for listening experiments within a laboratory setup under controlled conditions in some reviewed studies.

The laboratory studies of airborne sound concern mainly the perception of annoyance or loudness of noise within living environments, the evaluation of existing standardized indicators, and the rating of building elements. Fewer studies were found concerning airborne sound reduction than the ones concerning impact sound. That is mainly because impact sound types have been reported as the most disturbing in residential environments.¹⁴

The concept of acoustic comfort is hardly defined in the literature, despite being an important concept in engineering. It is typically used to consider a state of low or no noise and therefore lack of annoyance for the residents. A complete definition is provided in Rasmussen and Rindel,¹⁴ as: “a concept that can be characterized by absence of unwanted sound, desired sounds with the right level and quality, opportunities for acoustic activities without annoying other people.”

Standardized measurements and indicators are used to assess acoustic conditions in buildings;¹⁴ they are also used as a measure of acoustic comfort. But they do not always represent the perception of people in living sound environments. For instance, there are different types of building components, such as walls that offer various insulation and acoustic conditions in residencies.^3–6 In some other cases the characteristics of noise types might influence in various ways the perception of subjects: thus different standardized descriptors work better for various sound sources.^10–13

Therefore, the relation of the measured acoustic data to self-reported responses is important to study. The level of association is explored with statistical analyses comparing objective and subjective data. If a strong association is found between a descriptor and the subjective responses, then that descriptor could be used to predict the response of residents to a living environment based on acoustic data. Alternative versions of standardized descriptors with new adaptation terms are introduced many times in order to achieve stronger association of acoustic data with subjective responses.^10–13 Consequently, the study of acoustic comfort and the development of prediction models constitute an essential tool for building design with proper acoustic conditions.

Methods

A wide search for peer-reviewed publications and conference proceedings, which include examination between acoustic data and subjective responses relevant to airborne sound, has been done in the following databases: ScienceDirect, AIP Scitation, Ingenta Connect, ResearchGate, PubMed, Scopus, and Google Scholar. The search method included numerous searches in the databases using relevant keywords, such as objective, subjective, acoustic, psychoacoustic, self-report, rating, score, comfort, quality, airborne, sound, insulation, noise, annoyance, assessment, association, correlation. Several publications were subsequently found as references of the first selected papers.

This review article includes 11 studies: eight European studies,^3–8,12,13 one Asian study,⁹ and two Canadian studies.^10,11 In overall, 37 papers were found during the search in databases or relevant references and were evaluated by title name, abstract reading, and full reading. Only 11 papers met the requirements of this review: they offered comparison of results between airborne sound data and subjective responses, which is the subject of focus in this review. The other papers found were excluded because they concerned impact sound laboratory studies^15–30 or field studies.^31–40 Other exclusion criteria were the year of publication and language: only articles published after 1980 in English were included. The bibliographic research took place between April 2015 and September 2017.

Summary of methods, metrics and quantities in the reviewed studies

Many different indicators (or descriptors) are used to represent different quantities in acoustic measurements. They are all standardized in international ISO standards or other compliable national standards. Many variations of them exist as well, since experimental research has been done to acquire better indicators than the standardized ones. A description of all indicators involved in this review is presented in Table 1. For the detailed methods to acquire and calculate the indicators, please see the relevant standards.^41–48

Table 1.

Acoustic indicators used in the review studies.

Indicator	Description	Standards	References
$R_{w}$	Airborne weighted sound reduction index characterizing a building element (laboratory measurements)	ISO 717-1, ISO 140-3, ISO 16283-1, EN 12354-1	3,4,6,8 –11,41 –44
${R^{'}}_{w}$	Apparent airborne sound reduction index (same as $R_{w}$ , for in situ measurements)	ISO 717-1, ISO 140-4, ISO 10140-2, EN 12354-1	7,41,43,45,46
$R_{l i v i n g}$	Airborne weighted sound reduction index used in certain studies, calculated as $R_{w} + C_{50 - 5000}$	Similar to $R_{w}$	3,4,6,41 –43
$R_{A (50 - 5000)}$	Airborne A-weighted sound reduction index used in certain studies with freq. range 50–5000 Hz	Similar to $R_{w}$	5,41 –43
$R_{s p e e c h}$	Modified version of $R_{w} + C_{o p t}$	Similar to $R_{w}$	12
$L_{A e q}$	Sound pressure level equivalent to the total A-weighted levels measured over a stated period of time.	ISO 1996-1:2016	7,9,47
${\bar{L}}_{63 - 4000}$	Arithmetic mean value of sound pressure levels in octave bands of 63–4000 Hz		9
${\bar{L}}_{125 - 4000}$	Arithmetic mean value of sound pressure levels in octave bands of 125–4000 Hz		9
$L L$	Loudness level		48
$P L$	Perceived level		49
$N C$	Noise criteria		50
$P N C$	Preferred noise criteria		51
$S T C$	Airborne sound transmission class, calculated similar to $R_{w}$	ASTM E413	10 –12,52
$S T C_{n o 8}$	Modification of the airborne sound transmission class, ignoring the 8 dB rule (maximum allowed deviation from the rating contour)		10,11
$T L_{200 - 2500}$	Transmission loss, calculated similar to $R_{w}$ , with freq. range 200–2500 Hz	ASTM E413	10,52
$C$	C is an A-weighted pink noise spectrum adaptation term	ISO 717-1, EN ISO 12354-1,	3,4,6,10 –13,41 –43
$C_{50 - 3150}$	$C$ adaptation terms, freq. range 50–3150 Hz	Same as $C$	10,41,42
$C_{50 - 5000}$	$C$ adaptation terms, freq. range 50–5000 Hz	Same as $C$	3,4,6,41,42
$C_{t r}$	$C_{t r}$ is similar to $C$ but represents urban traffic noise spectra; it can be added to $D_{n T, w}$ or $R_{w}$ to include low-frequency noise influence	Same as $C$	10 –12,41,42
$C_{t r, 100 - 3150}$	$C_{t r}$ adaptation terms, freq. range 100–3150 Hz	Same as $C$	10,11
$C_{t r, 200 - 2500}$	$C_{t r}$ adaptation terms, freq. range 200–2500 Hz	Same as $C$	10
$C_{t r, m o d}$	Modified suggested spectrum based on $C_{t r}$	Not standardized	11
$C_{o p t}$	Optimal spectrum adaptation term calculated in order to adapt sound reduction index curves to associate better to subjective responses	Not standardized	9,13

Several statistical methods are also applied such as analysis of variance (ANOVA) and regression analysis which associate airborne sound data to subjective responses. The quality of statistical association is usually described with typical parameters such as the correlation coefficient, denoted as $r, ρ$ , or $R$ , the coefficient of determination, denoted as $R^{2}$ . The p-values and the confidence intervals (CIs) are measures of statistical significance. Details on the statistical methods can be found in relevant literature.⁵³

Some acronyms are used in this manuscript as abbreviations, namely SPL for sound pressure level, SNQ for single number quantity, SRI for sound reduction index (measured frequency spectra), STC for sound transmission class, and TL for transmission loss. The latter two terms are defined in the US standards:⁵² they are similar to the airborne SRI $R_{w}$ .

Evaluation of included studies

The quality of evidence for studies in this review was evaluated by means of Bradford Hill’s criteria^54,55 which is an evidence classification method often used in epidemiology and health review studies. The fulfillment of the criteria is rated in this review in a scale of high (+++), moderate (++), low (+), as happened in the previous parts.^1,2 The results are tabulated in Table 3 while the Bradford Hill’s criteria are as follows:

Strength of association. It refers to the causality proven by the association between the studied variables (cause, effect size, confounding factors).

Consistency. It indicates the degree of certainty when similar results are observed by different studies in different tests.

Specificity. Specific factors and effects on a specific population lead to a more likely causal relationship.

Temporality. It is based on temporal relations between effects, and used as an indicator for causality, meaning one effect happening after an exposure.

Biological gradient. It refers to the relation between exposure and effect; usually bigger exposure leads to greater effect, but not always, while the opposite outcome can happen as well.

Plausibility. It means that a biological explanation of why a cause leads to a certain effect supports a reasonable causality.

Coherence. It is a condition meaning that a stated causal relationship should not contradict with other accepted results or knowledge.

Experiment. It refers to the study design parameters that guarantee a reasonable causation, like randomization.

Analogy. The possibility of having or predicting analogous effects from similar factors without total evidence.

Publication type. An additional criterion was used to rank the reviewed studies. Scientific journal papers are thoroughly peer reviewed, while conference papers are usually less well reviewed. There are study reports from research organizations that may be scientifically well conducted but not reviewed at all. There are others, for example, unofficial reports, which are excluded. Thus, publications were evaluated as scientific journal (+++), conference paper (++), and report (+).

The included studies were evaluated by the authors of this article. In Table 2, the evidence evaluation rating of the studies is presented according to the criteria analyzed above. In Table 3, an overview of all the selected studies can be found, which are tabulated with summary of results, study design, methods, and conclusion. Readers who require a deeper insight in specific study results or conclusions may use the references and read the original publications. Sometimes, essential information are missing from this review article if they are not reported in the publications.

Table 2.

Evaluation of the presented studies according to selected criteria.

Reference	Publication type^a	Strength of association	Consistency	Specificity	Temporality	Biological gradient	Plausibility	Coherence	Experiment design	Analogy
Rychtáriková et al.³	++	+	+	+	+	+	+	+	+	+
Rychtáriková et al.⁴	+++	++	+	++	++	+	+	+	+	+
Monteiro et al.⁵	+++	++	++	++	++	+	++	++	++	++
Rychtáriková et al.⁶	+++	+	+	++	++	++	++	++	++	++
Pedersen et al.⁷	++	++	+	++	++	+	++	++	+	+
Vian et al.⁸	+++	++	+	++	++	++	++	++	++	++
Tachibana et al.⁹	+++	+	+	+	+	+	+	+	+	+
Park et al.¹⁰	+++	+++	++	+++	++	++	++	++	++	++
Park and Bradley¹¹	+++	+++	++	+++	++	++	++	++	++	++
Hongisto et al.¹²	+++	+++	++	+++	++	+++	++	++	+++	++
Virjonen et al.¹³	+++	+++	++	+++	++	+++	++	++	+++	++

Types: scientific journal (+++), conference paper (++), and report (+).

Table 3.

Overview of studies’ summaries.

Reference	Samples and experiment details	Variables as defined in studies	Models and results	Parameters for evaluation	Summary of results
Rychtáriková et al.^3,4	8 participants in a listening test (age span: 22–63)30 sound samples tested	Independent:15 different stimuli (neighbor noise, of sounds filtered through 2 types of wall airborne SRI spectra:a heavyweight wall (HW)a lightweight wall (LW)of the same index value $R_{l i v i n g} = R_{W} + C_{50 - 5000} = 52 dB$ Dependent:V1: perceived loudness of participants	Pairwise comparison test of sound stimuli	No details provided	LW sound samples were rated as less loud than HW filtered samples, except a case of movie sound reproduction $R_{W} + C_{50 - 5000}$ does not represent subjective perception of living noise through walls; this descriptor emphasizes on low frequencies
Monteiro et al.⁵	33 participants, mostly students in a listening test (12 females, 21 males)900 sound samples tested	Independent:90 pairs of sound samples (pink noise signals) filtered through different types of airborne sound reduction index (SRI) spectra:5 heavyweight walls (HW)5 lightweight walls (LW)with the same $R_{A, 50 - 5000}$ index but $R_{W}$ curvesDependent:V1: perceived loudness of participants	Pairwise comparison test(1HW–1LW with the same $R_{A, 50 - 5000}$ but different $R_{w}$ )Ranking by mode (RbM)Ranking by Q Score (RbQS)	Ranking and cumulative cumulative matrix outputs presentedNull hypothesis $H_{0}$ : LW sounds rated louder than HW sounds proven false within 95% CI using t-test	LW sound samples were rated as less loud than HW filtered samplesWall types representative of European structures $R_{A, 50 - 5000}$ does not reflect subjective perception and offers worse correlation than $R_{W} + C$ to the rankings examined
Rychtáriková et al.⁶	39 subjects in a listening test (14 females, 25 males)128 sound samples tested	Independent:64 different stimuli (everyday neighbor noise, 5 s each) of sounds filtered through 2 types of wall airborne SRI spectra:a heavyweight wall (HW)a lightweight wall (LW)of the same index value $R_{W} + C_{50 - 5000} = 52 dB$ Dependent:V1: perceived loudness of participants with interview	Two-alternative forced-choice task (2AFC) of sound stimuli in random pairwise comparisons	No details provided	LW sound samples were rated as less loud than HW filtered samples, except a case of movie sound reproduction $R_{W} + C_{50 - 5000}$ does not represent subjective perception of living noise through walls; this descriptor emphasizes on low frequencies
Pedersen et al.⁷	22 participants in a home listening test with self-adjustment of volume as calibration (6 female, 18 male, age span: 26–62)24 sound samples tested	Independent:4 neighbor noise types (20 s) sounds filtered through six types of wall airborne SRI spectraDependent:V1: annoyance response of participants	Dose–response regression curves of: $L_{A e q}$ to V1 ${R^{'}}_{w}$ to V1	Coefficients: $R^{2} = 0.95$ $R^{2} = 0.98$	Online survey test is feasibleAverage annoyance response highly associated with $L_{A e q}$ sound pressure levels and greatly associated with ${R^{'}}_{w}$
Vian et al.⁸	24 participants in a listening test (14 females, 10 males, age: 18–43)144 sound samples tested	Independent:12 different stimuli (music sound samples) of sounds filtered through 12 types of electronically synthesized wall airborne SRI spectraDependent:V1: annoyance response of participants	Analysis of variance (ANOVA) to test H₀: differences insignificant among test samplesCorrelation analysis for V1 and:The slopes of the SRI curvesThe A-weighted band spectra (40 Hz–10 kHz)The A-weighted band spectra (125 Hz–4 kHz)	Coefficients: $H_{0}$ was rejected (p < 0.01)95% CI (α = 0.05) $r = - 0.84, R^{2} = 0.71$ $r = 0.48, R^{2} = 0.23$ $r = 0.58, R^{2} = 0.33$	Subjective annoyance is better associated with samples of A-weighted spectra 125 Hz–4 kHz than 40 Hz–10 kHzThe slope of the SRI spectra and the dips in the curves found related to different responsesMusic samples with semantic context (words in a language that listeners can understand) were more annoying than without
Tachibana et al.⁹	8 male students in a listening test33 sound samples tested	Independent:3 different stimuli (white or pink noise filtered with different slopes) filtered through 11 types of electrically synthesized wall airborne SRI spectraMeasures: $L_{A}, L L, P L, N C, P N C, {\bar{L}}_{63 - 4000}$ , ${\bar{L}}_{125 - 4000}$	Method of adjustment by subject	No details provided	${\bar{L}}_{63 - 4000}$ showed the best correlation with loudness adjustmentsPerceived level $P L$ had good correlation as well
Park et al.¹⁰	15 participants in a listening test100 sound samples tested	Independent:5 Harvard speech sentence samples filtered through 12 types of electronically synthesized wall airborne SRI spectra based on real measured characteristicsDependent:V1: rated speech intelligibility of participants	Regression models with Boltzmann equations: V1 to:averaged $T L_{200 - 2500}$ $R_{w, 200 - 2500}$ $R_{w} + C_{t r, 100 - 3150}$ $R_{w} + C_{t r, 200 - 2500}$ $R_{w} + C_{o p t}$	Coefficients:All results statistically significant (p < 0.05) $r = - 0.98, R^{2} = 0.959$ $r = - 0.96, R^{2} = 0.922$ $r = - 0.74, R^{2} = 0.542$ $r = - 0.92, R^{2} = 0.842$ $r = - 0.98, R^{2} = 0.957$	Low association for most common indices $S T C$ , $S T C_{n o 8}$ , $R_{w}$ , $R_{w} + C_{100 - 3150}$ , $R_{w} + C_{t r, 100 - 3150}$ , butvery high for mean $T L_{200 - 2500}$ , $R_{w, 200 - 2500}$ and a tested $R_{w} + C_{o p t}$ The subjective speech intelligibility decreases when low frequencies included
Park and Bradley¹¹	30 subjects in 2 listening tests (10 subjects for the annoyance test; 20 for the loudness test)120 sound samples tested	Independent:3 music and 3 speech samplesFiltered through 20 types of electronically synthesized wall airborne SRI spectra based on real measured characteristicsDependent:V1: Mean annoyance responseV2: Mean loudness responseV3: Mean audibility response	Regression models with Boltzmann equations:V1 (annoyance) to: $S T C$ $R_{w}$ $R_{w} + C_{t r (100 - 3150)}$ $R_{w} + C_{t r, m o d}$ $S T C$ $R_{w}$ $R_{w} + C_{t r (100 - 3150)}$ $R_{w} + C_{t r, m o d}$ $S T C$ $R_{w}$ $R_{w} + C_{t r (100 - 3150)}$ $R_{w} + C_{t r, m o d}$	Coefficients:All results statistically significant (p < 0.01)(s: speech, m: music)s: R² = 0.856, m: R² = 0.728s: R² = 0.890, m: R² = 0.798s: R² = 0.566, m: R² = 0.950s: R² = 0.541, m: R² = 0.983s: R² = 0.886, m: R² = 0.734s: R² = 0.933, m: R² = 0.779s: R² = 0.676, m: R² = 0.970s: R² = 0.634, m: R² = 0.991s: R² = 0.968, m: R² = 0.452s: R² = 0.971, m: R² = 0.526s: R² = 0.903, m: R² = 0.757s: R² = 0.853, m: R² = 0.920	Both and predictors were associated more with speech sounds than music soundsOverall trends for the regression equations between annoyance and loudness responses similarFew descriptors were strongly related to all response variables (V1–V3) while there were differentiations
Hongisto et al.¹²	59 subjects in a listening test(19 males, 40 females, age 20–43)54 sound samples tested	Independent:6 sound samples (living sounds) filtered through 9 types of wall airborne SRI spectra.Dependent:V1: mean loudness responseV2: mean disturbance responseV3: mean acceptability response	Regression models:V1 (loudness) to: $R_{w}$ $S T C_{n o 8}$ $R_{w} + C_{100 - 3150}$ $R_{w} + C_{100 - 5000}$ $R_{w} + C_{50 - 3150}$ $R_{w} + C_{50 - 5000}$ $R_{s p e e c h}$ V2 (disturbance) to: $R_{w}$ $S T C_{n o 8}$ $R_{s p e e c h}$ $R_{w}$ $S T C_{n o 8}$ $R_{s p e e c h}$	Coefficients:All results statistically significant (p < 0.05)R² = 0.83R² = 0.85R² = 0.92 (music only)R² = 0.93 (music only)R² = 0.93 (music only)R² = 0.93 (music only)R² = 0.80R² = 0.84R² = 0.87R² = 0.85R² = 0.83R² = 0.87R² = 0.88	Single number quantities (SNQs) including the extended range at low frequencies of 50–80 Hz performed worse than the SNQs without itSubjective responses for living sound types with flat spectrum seems to be predicted best by, and SNQs which ignore frequencies below 100 HzSubjective responses for music (bass sounds) were best associated with spectra including low frequencies
Virjonen et al.¹³	Reused data from Jeon et al.²¹ 59 subjects in a listening test (19 males, 40 females, age 20–43)	Independent:same data as [R8]Dependent:mean disturbance responses	Regression models: $R_{w} + C_{o p t}$ to disturbance from:S1: guitarS2: music (traffic)S3: music (living)S4: baby cryS5: loud speechS6: dog bark	Coefficients:All results statistically significant (p < 0.05)R² = 0.96R² = 0.66R² = 0.92R² = 0.57R² = 0.94R² = 0.74	$R_{w} + C_{o p t}$ is developed as a SNQ that would explain the disturbance in the range of 50–5000 Hz better than any SNQFinally, it provides better association for most sound types tested than the usual standardized descriptors

Results: associations of airborne sound data to self-reported responses in laboratory tests

In Rychtáriková et al.,^3–6 a group of studies regarding the subjective perception of loudness in living environments are presented. The studies are based on rating noise types transmitted through lightweight and heavyweight wall structures, from now on denoted as LW and HW, respectively. Several wall structures were measured according to ISO 717-1, ISO 717-2 and the acquired airborne SRI spectra were used to filter different recorded noise types. Then the created samples were used in listening tests and they were randomly sorted in pairs of a heavyweight and lightweight wall case, both having the same single value but different spectra for their SRI. The test samples were presented in random order for pairwise comparisons. The participants were asked in all cases to rate the sound that was perceived as the loudest but without knowing that they actually rate noise transmission through different types of walls. In some cases, the participants reported that they would probably reply differently if the question of the test was to address the most annoying sound instead of the loudest one.⁴

Specifically, in Rychtáriková et al.,^3,4 a small sample of eight people rated 15 different sound stimuli of typical neighbor noise types (5 s each) filtered through a heavyweight wall (masonry) spectrum and a lightweight wall spectrum (gypsum boards on metal studs). Both cases of wall structures had the same single value for $R_{l i v i n g} = R_{w} + C_{50 - 5000} = 52 dB$ but different airborne SRI curves. The tested sound stimuli were not only recorded noise samples but also the inverted versions of them, that is, the reversed signals, so the sound samples were played backward. Additional noise signals were used too, which had the same spectra as the original sounds, but without semantic information (i.e. speech, music, etc.). The listeners rated as louder the sound transmitted through the heavy wall structure in most cases, except some cases including bass sounds with many present dynamics and much modulation, as commented in the article. The main objective was to evaluate the effect of frequency and time variations in loudness perception. It was indicated that modulation and semantic context play an important role in loudness perception, especially when there is dominant low-frequency content. Time variations with several minima and maxima were found important to make people perceive a sound as loud. The metric of loudness is commented to overestimate sometimes the expected perceived loudness, due to high influence at the dominant low-frequency bands.

In Monteiro et al.,⁵ a bigger listening test was conducted in Belgium and Spain with 33 participants (21 females, 12 males) where 90 pairs of sound samples were used to test the descriptor $R_{A, 50 - 5000}$ and compare it to perceived loudness. Pink noise signals filtered through 10 different types of walls were used, which were presented to the participants in five pairs: each pair included a HW case and a LW. In every pair comparison, both wall types had the same $R_{A, 50 - 5000}$ single value but different airborne sound reduction spectra. All 10 cases of walls were compared to each other twice in randomly formed pairs. The stimuli filtered through the five LW cases were rated as less loud than the five HW cases and the wall types used are reported as representative of European wall structures in the study. A t-test was performed indicating that the listeners perceived a difference between the test pairs of sounds within 95% CI. In some cases, noise sounds transmitted through the LW walls were considered less loud even while compared to the noise sounds through HW with higher single value $R_{w}$ but the same $R_{A, 50 - 5000}$ . It is concluded that $R_{A, 50 - 5000}$ does not associate well with subjective loudness perception. That descriptor is also reported to offer worse correlation than standardized $R_{w} + C$ (frequency range 100–3150 Hz) but no test parameters are provided for this.

Then, in Rychtáriková et al.,⁶ another listening test was performed using 64 typical everyday sounds, recorded live during 2 weeks in 10 selected living rooms of apartments in Austrian buildings. The sounds were again filtered through a heavyweight and a lightweight wall sound reduction spectrum, forming pairwise comparisons for 39 participants (14 women and 25 men). Most responses considered louder the transmission through the heavyweight wall type as before in Rychtáriková et al.³ In few cases, where sounds through lightweight wall types were considered louder, they included low-frequency content extremely amplified by electronic devices as reported. The LW sound reduction was better than the HW in the middle frequency range of 100–3150 Hz. Then 12 of the participants were deployed, to test the hypothesis whether A-weighting in SPL is adequate when evaluating everyday living noise types, due to low weighted sound levels. The subjects reported to perceive low frequencies less loud in low SPL compared to high SPL. Calibration and background noise levels are reported as crucial parameters for the reliability of a listening test. Temporal amplitude modulations in the test sounds are stressed as important as well. Summing up, it is indicated in all the above studies^3–6 that $R_{l i v i n g} = R_{W} + C_{50 - 5000}$ is not an adequate descriptor of airborne sound performance of walls regarding the subjective perception of loudness. The explanation provided is due to the high influence of low-frequency content in the frequency adaptation terms.

In Pedersen et al.,⁷ 22 persons from the COST TU0901 action took part in an online listening test, testing 24 sound stimuli at their home computer setup with headphones. Four typical neighbor noise types filtered through the airborne SRI curves of six usual various types of walls were assessed. The results indicated a high association (R² = 0.95) between the average annoyance response and the $L_{A e q}$ SPL of the sound samples after filtering. Great association with R² = 0.98 was reported also between the average annoyance response and the ${R^{'}}_{w}$ apparent SRI levels. However, the conditions of this test might seriously deviate from controlled laboratory conditions, since there is no calibration except from a self-adjustment in volume of the users. In addition, the study reports that some sound samples were radically amplified up to 14 dB so as to be definitely audible in the online test.

In Vian et al.,⁸ a listening experiment for the evaluation of French regulation toward airborne sound insulation levels was conducted. Twenty-four participants took part in a laboratory test (14 females, 10 males, age span: 18–43). They reported their annoyance on 12 music sound stimuli which were filtered through 12 electronically synthesized wall SRI curves. The experiment was based on an incomplete factorial design, so from the whole 144 test samples, 1/3 was assessed by every subject. ANOVA was used to test the distinguishability between the samples (significant differences with p < 0.01). Newman–Keuls multiple-means comparison was used for grouping and ranking the annoyance responses. That comparison showed that slope and dips in the SRI curves, as well as the bandwidth and character of the sounds, have statistically important effects on the self-reported annoyance. It is concluded that an increasing slope of the insulation curve (i.e. more reduction in higher frequencies) leads to less annoyance. Then, a correlation analysis proved that there is a strong relation between the slopes of the SRI curves and the reported annoyance (r = –0.85, R² = 0.72, 95% CI). It is reported that subjective annoyance is better associated with samples with A-weighted spectra of 125 Hz–4 kHz (r = 0.58, R² = 0.33) instead of 40 Hz–10 kHz (r = 0.48, R² = 0.23) within 95% CI. In addition, noise from neighbors’ speech in both French and English was assessed by French subjects in this test: the intelligibility of the sounds was found important, meaning that when there is a semantic context in the noise, the annoyance is bigger.

In Tachibana et al.,⁹ a listening test took place for setting up a method for the evaluation of airborne sound insulation testing different measures. A limited sample of eight university students tested three different types of artificial noise sounds. Specifically, white or pink noise was filtered through various artificial frequency spectra of walls, based on real frequency spectra. The sound stimuli were evaluated by self-adjustment, meaning that the subjects used reference sounds to adjust the amplitude of test stimuli until they perceived every test sample as equally loud to the reference. The point of subjective equality (PSE) was used in that test for the adjustments. Many loudness measures were mentioned to have been tested such as the A-weighted SPL $L_{A}$ , the loudness level $L L$ , the perceived level $P L$ , the noise criteria $N C$ , and the preferred noise criteria $P N C .$ As concluded in that article, the SPL weighted from 63 Hz: ${\bar{L}}_{63 - 4000}$ showed the best correlation with loudness adjustments. It is also mentioned that perceived level $P L$ had good correlation as well. However, lack of statistical test parameters in the paper and the small sample size make the outcome of this study unreliable.

In Park et al.,¹⁰ a listening experiment for the evaluation of airborne sound insulation SNQs regarding speech intelligibility was conducted with 15 subjects (participants). A total of 100 sound samples was tested consisting of five Harvard speech test sentences filtered through 20 different types of wall airborne sound reduction spectra; the spectra were synthesized but based on real measured characteristics. Different measures were explored in terms of best-fitting regression curves (using Boltzmann’s equations) to the self-reported speech intelligibility. The comparison is a bit different in this study: good speech intelligibility corresponds to bad sound insulation performance and vice versa. Thus, a low intelligibility in the test would predict a sufficient airborne sound reduction of the test walls. The statistical associations of the most common standardized measures: $S T C$ , $S T C_{n o 8}$ , $R_{w}$ , $R_{w} + C_{100 - 3150}$ , $R_{w} + C_{t r, 100 - 3150}$ with the subjective intelligibility ratings were found weak in overall, having acquired determination coefficients $R^{2}$ of 0.510, 0.661, 0.542, 0.359, and 0.205, respectively.

It was concluded that the examined descriptors are influenced plenty by the frequency range. Thus when low frequencies are included in the calculations correlation with speech intelligibility decreases because they do not contain useful information on the transmitted speech; the low frequencies do not contain information on the transmitted speech. In a parametric analysis included, the authors demonstrate that the highest correlation can be acquired when using the arithmetic average TL with the restricted frequency range of 200–2500 Hz: $T L_{200 - 2500}$ (r = –0.98, R² = 0.959). For the standardized $R_{w, 200 - 2500}$ (with the same restricted range), a similarly great association was acquired (R² = 0.922), while they are not so good for descriptors with wider frequency range: $R_{w} + C_{t r, 100 - 3150}$ (R² = 0.542) and $R_{w} + C_{t r, 200 - 2500}$ (R² = 0.842). Finally, a spectrum adaptation based on a speech rating contour a (band pass filter for speech) was suggested and tested: $C_{o p t}$ . When added to $R_{w}$ , there was a great association between the self-reported speech intelligibility and the indicator: $R_{w} + C_{o p t}$ (r =−0.98, R² = 0.957). The descriptors which involve arithmetic averages in frequency bands relevant to speech (200 Hz–2.5 kHz) were reported as the best associated with the intelligibility responses. All the presented results were found statistically significant (p < 0.05).

A continuation of this research is presented in Park and Bradley¹¹ for the evaluation of the existing standardized airborne sound insulation measures for annoyance, loudness, and audibility. Another listening test was conducted with the same methodology as Park et al.,¹⁰ this time using three speech samples and three music samples, filtered through 20 various wall SRI spectra. The measured walls had a spectrum of $S T C$ values of 34–58 dB. The total of 120 speech and music samples were presented in random orders and rated in two occasions: one test regarding the annoyance degree (10 participants, rating scale 1 (Not at all) to 7 (Extremely)) and a second test rating loudness perception (another 20 participants, rating scale 0 (Not audible) to 7 (Extremely)). The results were again evaluated in terms of average annoyance or loudness response regression fit to standard SNQs (using Boltzmann’s equations) again.

The associations observed for self-reported mean annoyance were strong in relation to $S T C$ (speech: R² = 0.856, music: R² = 0.728) and a bit higher for the mean annoyance and $R_{w}$ (speech: R² = 0.890, music: R² = 0.798). As for the loudness test, the associations were just good between the self-reported mean loudness ratings and $S T C$ (speech: R² = 0.886, music: R² = 0.734) and better for the association of loudness and $R_{w}$ (speech: R² = 0.933, music: R² = 0.779). Different spectrum adaptation terms were added to $R_{w}$ offering some improvements for the association with self-reported annoyance: $R_{w} + C_{t r (100 - 3150)}$ (speech: R² = 0.566, music: R² = 0.950) and loudness: $R_{w} + C_{t r (100 - 3150)}$ (speech: R² = 0.676, music: R² = 0.970) which worked well only for the music cases. The best association in the study was reported with the use of a new suggested adaptation term $C_{t r, m o d}$ , after modifying the $C_{t r}$ spectrum to emphasize more on high frequencies around 1 kHz. Again the association was sufficient only for the music sound cases in relation to annoyance: $R_{w} + C_{t r, m o d}$ (speech: R² = 0.541, music: R² = 0.983) and loudness $R_{w} + C_{t r, m o d}$ (speech: R² = 0.634, music: R² = 0.991).

Finally, for the audibility test, the previous values of loudness responses equal to 0 were used to define the state of “not audible.” The association of audibility to $S T C$ was good for speech only (speech: R² = 0.968, music: R² = 0.452); the same applies to the relation to $R_{w}$ with better values (speech: R² = 0.971, music: R² = 0.526). Improvements were made for $R_{w}$ results for music only when adding the spectrum adaptation terms (opposite results for speech): $R_{w} + C$ (speech: R² = 0.903, music: R² = 0.757) and $R_{w} + C_{t r (100 - 3150)}$ (speech: R² = 0.853, music: R² = 0.920). All results for the determination coefficients were reported statistically significant (p < 0.01) in the study. The overall trends for the regression equations between annoyance and loudness responses were reported very similar indicating very small differences between the two measures. Summing up, few descriptors were strongly associated with all tested variables of subjective responses. Consequently, it is concluded that different descriptors work better for different kinds of noise.

A similar study was conducted in Finland¹² regarding the subjective evaluation of standardized SNQs characterizing airborne sound insulation of building elements, as stated in ISO 717-1⁴¹ and ASTM E413.⁵² A listening test took place including 59 subjects (19 males, 40 females, age 20–43; mean age: 27 years) who rated a set of six recorded sounds of typical noise types found in residential buildings: guitar sound, two music samples, speech, baby cry and barking dog. The music samples were modified to correspond to the traffic and the living spectrum of the relevant adaptation terms according to ISO 717-1. The test sound samples were filtered according to airborne SRI spectra of nine different wall structures measured in laboratory conditions (54 test samples in total, $R_{w}$ values range 48–75 dB). The subjective measures used were loudness, disturbance and acceptability ratings (scale 0 (Not at all) to 10 (Extremely)) formulated in simple questions. Linear regression analysis was performed between the averaged subjective responses and the SNQs: the determination coefficients (average $R^{2}$ for all six sound samples) were reported, as individual or averaged values. Mostly, the latter are presented in this review.

An initial conclusion was that SNQs including the extended frequency range at low frequencies of 50–80 Hz performed worse than the SNQs without it, which was derived in previously presented studies as well.^10,11 The SNQ of $S T C_{n o 8}$ predicted loudness (R² = 0.85) and disturbance (R² = 0.87) response better than every other, while $R_{s p e e c h}$ (R² = 0.88) was the best metric to predict acceptability. Another conclusion is that $R_{w} + C_{t r, 50 - 3150}$ and $R_{w} + C_{t r, 50 - 5000}$ were the most inefficient predictors in general with coefficients R² less than 0.7. $C_{t r, 50 - 5000}$ is the A-weighted urban traffic noise adaptation spectrum of ISO 717-1. Different SNQs had different prediction efficiencies for the various sound types tested, due to emphasis on different frequencies. For the case of music sounds (which included traffic noise adaptation spectrum) with dominant bass frequency context, the descriptors $R_{w} + C_{100 - 3150}$ (R² = 0.92), $R_{w} + C_{100 - 5000}$ (R² = 0.93), $R_{w} + C_{50 - 3150}$ (R² = 0.93) and $R_{w} + C_{50 - 5000}$ (R² = 0.93) predicted the loudness response best, whereas the last two SNQs predicted best also the disturbance and acceptability ratings contrary to some previous studies.^12,13 Overall, the response for living sound types with flat spectrum seems to be predicted best by $R_{w}$ , $S T C$ , and SNQs which ignore frequencies below 100 Hz showing $R^{2}$ values around 0.9, while for noise types with dominant higher frequencies, such as baby cry and dog bark the best predictors were $R_{s p e e c h}$ , $S T C$ , and $S T C_{n o 8}$ . All the presented results were found statistically significant (p < 0.05).

Additionally, in Virjonen et al.¹³ which is a continuation of the previous study,¹² the same data of the six sound samples and the subjective ratings were utilized in order to create a SNQ measure which would predict and explain the disturbance in the frequency range of 50–5000 Hz better than all other standardized SNQs. A certain algorithm was developed for optimal fit between mean subjective ratings and the optimized averaged reference spectra. The outcome was another descriptor with a new adaptation spectrum: $R_{w} + C_{o p t}$ . That finally works as expected in general, specifically well in association with guitar sound with R² = 0.96, music (living spectrum as stated in Hongisto et al.¹²) with R² = 0.92, and loud speech with R² = 0.94. The adaptation term $C_{o p t}$ is utilized for the averaged spectra of noise types tested, while $C_{S i}$ was tested for the individual noise types offering for every specific sound optimization a determination coefficient of R² = 0.95 in all individual cases. All the presented results were found again statistically significant (p < 0.05).

Discussion

There can be strong associations between airborne sound data and self-reported responses of annoyance and loudness^10–13 in general. However, the statistical associations were weak in overall when intelligibility was used as a measure of perception.¹⁰ The measures of self-reported audibility and acceptability were also used once, in Park and Bradley¹¹ and Virjonen et al.,¹³ respectively. It is also highlighted in a study that few differences were found between loudness and annoyance in the subjective responses.¹¹

Most laboratory studies based on airborne sound data examine the horizontal sound transmission, that is, the $R_{w}$ of walls. When it comes to vertical sound transmission, most studies deal with impact sound data as a priority, which is critical for propagation through floors and it has been found to be the most disturbing noise type in residential buildings.^14,27–35 Therefore, it seems that airborne sound studies remain supplementary to impact sound for researching vertical noise transmission.

The airborne sound studies testing transmission through walls conclude that subjects perceive noise through heavyweight walls as louder than the ones transmitted through lightweight walls.^3–6 The airborne sound reduction frequency spectra of lightweight walls are usually better than heavyweight ones, except the low-frequency range. The effect of frequency and time variations in subjective perception was also highlighted in Rychtáriková et al.:⁶ it was indicated that modulation and semantic context (e.g. speech) affect loudness perception. The latter observation is also supported in Vian et al.⁸ where the intelligibility of the sounds was found important: annoyance perception is affected by semantic context in the noise sounds.

The inclusion of low frequencies (down to 50 Hz) in the measurements and derivation of descriptors seems to be an important issue in some of the reviewed studies. In the studies,^3–6 the descriptors with extended spectrum adaptation terms like $R_{W} + C_{50 - 5000}$ do not associate sufficiently with the subjective annoyance of residents because such descriptors emphasize a lot on low frequencies. The same conclusion is supported in literature,^8–12 while in Park et al.¹⁰ a limited frequency range of 200–2500 Hz is suggested to be optimal for association of subjective intelligibility of speech sounds.

The previous conclusions from studies^3–6,8–12 are contradictory to the general trend in impact sound measurements and descriptors: inclusion of low-frequency spectra down to 50 Hz (even down to 20–25 Hz) in impact sound data is considered necessary to achieve sufficient association with self-reported responses.^27–35 However, sometimes the same is stated in studies for airborne sound: in Park and Bradley,¹¹ many associations were very good between subjective annoyance or loudness and SRI descriptors for various stimuli with low-frequency content too. In addition, in Hongisto et al.,¹² descriptors such as $R_{w}$ and $S T C$ with extended frequency range down to 50 Hz are suggested perform best for prediction of subjective annoyance and loudness to music stimuli (but not the other sounds tested).

Furthermore, the different types of sound stimuli are highlighted as an important parameter in some studies^10,13 because speech sounds have more high-frequency content while low frequencies are dominant in music sound stimuli. There are variations of course, for the cases of different sounds tested and responses. Therefore, in many studies, some of the descriptors work for certain types of sounds only, for example, in Park and Bradley¹¹ and Virjonen et al.,¹³ the descriptors with extended spectra to low frequencies associate very well with music but not with speech in most cases. In Virjonen et al.¹³ the descriptors that relate well to speech do not relate that well for music or other living sounds.

In overall, many studies attempt to find a descriptor for airborne sound that would predict well the general noise annoyance perception. But there is not a certain descriptor that seems to associate great with every type of sound. This review demonstrates that different descriptors work better for various stimuli. However, in Virjonen et al.¹³ and Rasmussen and Rindel,¹⁴ a spectrum adaptation term is suggested to formulate the measure: $R_{w} + C_{o p t}$ which predicts well most airborne sound types (except baby cry) within the full measurement spectrum of 50–5000 Hz and it is different to the standardized adaptation term $C_{50 - 5000}$ .

In some cases, the study design parameters are problematic: the sample size of listening tests can be small as in Pedersen et al.⁷ and Vian et al.⁸ or extremely small as in literature.^3,4,9,10 However more subjects would be essential for a sufficient sample size and thus further inference of the results from such experiments; conclusions are weak when they are based on a small sample. Most of the studies have a good level of presentation and evaluation of evidence as can be seen in Table 2. Many statistical evaluations took place to compare and relate results between objective and subjective data, some were incomplete with missing parameters, and tests of significance or some study details were not reported at all.^3,4,9 It would be essential for every study to have a method presentation, study design, and sufficient tabulation of relevant parameters.

Conclusion

This review shows that subjective response to airborne sound in dwellings is complicated: it can be predicted well in some cases but not always. Standardized SNQs and alternative descriptors for airborne sound have been evaluated and associated sufficiently with subjective responses collected in laboratory listening tests. The type of sound stimulus is significant because different stimuli with various frequency spectra correspond better to different descriptors. Inclusion of low frequencies down to 50 Hz in airborne sound measurements seems to be problematic: it leads only certain metrics to associate better with self-reported responses.

Consequently, all descriptors do not associate well with all kinds of airborne sound stimuli in living environments. The descriptors with low-frequency adaptation spectra relate better to music sound sources with dominant low-frequency content while the opposite applies to sound stimuli of speech. Finally, there is no overall indicator to work best for all kinds of airborne sound types in dwellings, but few suggestions of frequency adaptation spectra work very well for that reason. Additional properties of noise signals such as frequency and time modulation and semantic context are indicated to play a role in subjective perception of annoyance or loudness.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This article was written within the research project “Acoustic Comfort in Building Apartments” funded by Saint-Gobain Weber.

References

Vardaxis

N-G

Bard

Waye

KP.

Review of acoustic comfort evaluation in dwellings—part I: associations of acoustic field data to subjective responses from building surveys. Build Acoust 2018; 25: 151–170.

Vardaxis

N-G

Bard

Review of acoustic comfort evaluation in dwellings—part II: impact sound data associated to subjective responses in laboratory tests. Build Acoust 2018; 25: 171–192.

Rychtáriková

Müllner

Urbán

et al . Influence of temporal and spectral features of neighbour’s noise on perception of its loudness. In: Proceedings of the Euronoise 2013, Innsbruck, 15–18 September 2013.

Rychtáriková

Roozen

Müllner

et al . Listening test experiments for comparisons of sound transmitted through light weight and heavy weight walls. Akustika 2013; 19: 8–13.

Monteiro

Machimbarrena

De la Prida

et al . Subjective and objective acoustic performance ranking of heavy and light weight walls. Appl Acoust 2016; 110: 268–279.

Rychtáriková

Muellner

Chmelík

et al . Perceived loudness of neighbour sounds heard through heavy and light-weight walls with equal R_w + C_{50 – 5000}. Acta Acust United Ac 2016; 102(1): 58–66.

Pedersen

Antunes

Rasmussen

. Online listening tests on sound insulation of walls—a feasibility study. In: Proceedings of the Euronoise 2012, Prague, 10–13 June 2012.

Vian

J-P

Danner

Bauer

JW.

Assessment of significant acoustical parameters for rating sound insulation of party walls. J Acoust Soc Am 1983; 73(4): 1236–1243.

Tachibana

Hamado

Sato

Loudness evaluation of sounds transmitted through walls—basic experiments with artificial sounds. J Sound Vib 1988; 127(3): 499–506.

10.

Park

Bradley

Gover

BN.

Evaluating airborne sound insulation in terms of speech intelligibility. J Acoust Soc Am 2008; 123(3): 1458–1471.

11.

Park

Bradley

JS.

Evaluating standard airborne sound insulation measures in terms of annoyance, loudness and audibility ratings. J Acoust Soc Am 2009; 126(1): 208–219.

12.

Hongisto

Oliva

Keränen

Subjective and objective rating of airborne sound insulation—living sounds. Acta Acust United Ac 2014; 100: 848–864.

13.

Virjonen

Hongisto

Oliva

Optimized single-number quantity for rating the airborne sound insulation of constructions: living sounds. J Acoust Soc Am 2016; 160(4): 4428.

14.

Rasmussen

Rindel

JH.

Concepts for evaluation of sound insulation of dwellings—from chaos to consensus? In: Proceedings of the forum acusticum 2005, Budapest, 29 August–2 September 2005.

15.

Yeon

Subjective evaluation of floor impact noise based on the model of ACF/IACF. J Sound Vib 2001; 241(1): 147–155.

16.

Yeon

Jeong

JH.

Objective and subjective evaluation of floor impact noise. J Temp Des Archit Environ 2002; 2(1): 20–28.

17.

Jeon

Jeong

Vorländer

et al . Evaluation of floor impact sound insulation in reinforced concrete buildings. Acta Acust United Ac 2004; 90: 313–318.

18.

Jeon

Ryu

Jeong

et al . Review of impact ball in evaluating floor impact sound. Acta Acust United Ac 2006; 92: 777–786.

19.

Lee

Kim

Jeon

JY.

Psychoacoustical characteristics of impact ball sounds on concrete floors. Acta Acust United Ac 2009; 95: 707–717.

20.

Jeon

Sato

Annoyance caused by heavyweight floor impact sounds in relation to the autocorrelation function and sound quality metrics. J Sound Vib 2008; 311: 767–785.

21.

Jeon

Lee

Kim

et al . Subjective evaluation of heavy-weight floor impact sounds in relation to spatial characteristics. J Acoust Soc Am 2009; 125(5): 2987–2994.

22.

Kim

Ryu

Jeon

JY.

Effect of temporal decay on perception of heavy-weight floor impact sounds. J Acoust Soc Am 2013; 134(4): 2730–2738.

23.

Jeon

. Psychoacoustical evaluation of heavyweight floor impact sounds in apartment buildings. In: Proceedings of the inter-noise 2014, Melbourne, VIC, Australia, 16–19 November 2014.

24.

Ryu

Sato

Kurakata

et al . Relation between annoyance and single-number quantities for rating heavy-weight floor impact sound insulation in wooden houses. J Acoust Soc Am 2011; 129(5): 3047–3055.

25.

Gover

Bradley

Schoenwald

et al . Subjective ranking of footstep and low-frequency impact sounds on lightweight wood-framed floor assemblies. In: Proceedings of the forum Acusticum 2011, Aalborg, 27 June–1 July 2011.

26.

Gover

Bradley

Zeitler

et al . Objective and subjective assessment of lightweight wood-framed floor assemblies in response to footstep and low-frequency impact sounds. In: Proceedings of the inter-noise 2011, Osaka, Japan, 5–7 September 2011.

27.

Späh

Hagberg

Bartlomé

et al . Subjective and objective evaluation of impact noise sources in wooden buildings. Build Acoust 2013; 20(3): 193–214.

28.

Kylliäinen

Hongisto

Oliva

et al . A laboratory listening experiment on subjective and objective rating of impact sound insulation of concrete floors. In: Proceedings of the inter-noise 2016, Hamburg, 21–24 August 2016.

29.

Kylliäinen

Hongisto

Oliva

et al . Subjective and objective rating of impact sound insulation of a concrete floor with various coverings. Acta Acust United Ac 2017; 103: 236–250.

30.

Öqvist

Ljunggren

Johnsson

. Listening test of walking noise from 20 Hz in dwellings. In: Proceedings of the inter-noise 2017, Hong Kong, China, 7 December 2017, pp. 3935–3942. Reston, VA: Institute of Noise Control Engineering.

31.

Ljunggren

Simmons

Hagberg

Correlation between sound insulation and occupants’ perception—proposal of alternative single number rating of impact sound. Appl Acoust 2014; 85: 57–68.

32.

Ljunggren

Simmons

Hagberg

. Findings from the AkuLite project: correlation between measured vibro-acoustic parameters and subjective perception in lightweight buildings. In: Proceedings of the Inter-noise 2013, Innsbruck, 15–18 September 2013.

33.

Simmons

Ljunggren

Aku20—searching for optimal single number quantities in EN ISO 717-2 correlating field measurements 20—5000 Hz to occupant’s ratings of impact sounds—new findings for concrete floors. In: Proceedings of the EuroNoise 2015, Maastricht, 31 May–3 June 2015.

34.

Ljunggren

Simmons

Öqvist

Correlation between sound insulation and occupants’ perception—proposal of alternative single number rating of impact sound-part II. Appl Acoust 2017; 123: 143–151.

35.

Ljunggren

Simmons

Öqvist

. Evaluation of impact sound insulation from 20 Hz. In: Proceedings of the 24th international congress on sound and vibration (ICSV24), London, 23–24 July 2017.

36.

Guigou-Carter

Balanant

Villenave

Acoustic comfort evaluation in lightweight wood-based buildings. In: Proceedings of the forum acusticum, Krakow, 7–12 September 2014.

37.

Milford

Høsøien

Løvstad

et al . Socio-acoustic survey of sound quality in dwellings in Norway. In: Proceedings of the inter-noise 2016, Hamburg, 21 August 2016.

38.

Høsøien

Rindel

Løvstad

et al . Impact sound insulation and perceived sound quality. In: Proceedings of the inter-noise 2016, Hamburg, 21 August 2016.

39.

Hagberg

Evaluating field measurements of impact sound. J Build Acoust 2010; 7: 105–128.

40.

Hongisto

Mäkilä

Suokas

Satisfaction with sound insulation in residential dwellings—the effect of wall construction. J Build Environ 2015; 86: 309–320.

41.

ISO 717-1:1996. Acoustics—rating of sound insulation in buildings and of buildings elements—part 1: airborne sound insulation.

42.

ISO 140-3:1995. Acoustics—measurement of sound insulation in buildings and of building elements—part 3: laboratory measurements of airborne sound insulation of building elements.

43.

EN 12354-1:2000. Building acoustics—estimation of acoustic performance of buildings from the performance of elements—part 1: airborne sound insulation between rooms.

44.

ISO 16283-1:1998. Acoustics—field measurement of sound insulation in buildings and of building elements—part 1: airborne sound insulation.

45.

ISO 140-4:1998. Acoustics—measurement of sound insulation in buildings and of building elements—part 4: field measurements of airborne sound insulation between rooms.

46.

ISO 10140-2:2010. Acoustics—laboratory measurement of sound insulation of building elements—part 2: measurement of airborne sound insulation.

47.

ISO 1996-1:2016. Acoustics—description, measurement and assessment of environmental noise—part 1: basic quantities and assessment procedures.

48.

Stevens

SS.

Procedure for calculating loudness: Mark VI. J Acoust Soc Am 1961; 33(11): 1577–1585.

49.

Stevens

SS.

Perceived level of noise by Mark VII and decibels (E). J Acoust Soc Am 1972; 51(2): 575–601.

50.

Schultz

TJ.

Noise-criterion curves for use with the USASI preferred frequencies. J Acoust Soc Am 1968; 43(3): 637–638.

51.

Beranek

Blazier

Figwer

JJ.

Preferred Noise Criterion (PNC) curves and their application to rooms. J Acoust Soc Am 1971; 50(5): 1223–1228.

52.

ASTM E413-87:1999. Classification for rating sound insulation.

53.

Rice

JA.

Mathematical statistics and data analysis. 2nd ed. Boston, MA: Cengage Learning, 2007.

54.

Höfler

The Bradford hill considerations on causality: a counterfactual perspective. Emerg Themes Epidemiol 2005; 2: 11.

55.

Schünemann

Hill

Guyatt

et al . The GRADE approach and Bradford hill’s criteria for causation. J Epidemiol Community Health 2011; 65: 392–395.

Review of acoustic comfort evaluation in dwellings: Part III—airborne sound data associated with subjective responses in laboratory tests

Abstract

Keywords

Introduction

Methods

Summary of methods, metrics and quantities in the reviewed studies

Evaluation of included studies

Results: associations of airborne sound data to self-reported responses in laboratory tests

Discussion

Conclusion

Footnotes

Declaration of conflicting interests

Funding

References