Applied Acoustics 127 (2017) 126–132

Contents lists available at ScienceDirect

Applied Acoustics
journal homepage: www.elsevier.com/locate/apacoust

Estimation of the minimum measurement time interval in acoustic noise

Consolatina Liguori, Alessandro Ruggiero, Domenico Russo ⇑, Paolo Sommella
Department of Industrial Engineering (DIIn), University of Salerno, Via Giovanni Paolo II 132, I - 84084 Fisciano (SA), Italy

a r t i c l e i n f o a b s t r a c t

Article history: The appropriate choice of the minimum measurement time interval is introduced for an accurate estima-
Received 11 October 2016 tion of environmental noise indicators. The proposal is based on a bootstrap approach for the continuous
Received in revised form 24 May 2017 estimation of measurement uncertainty in order to determine the statistical variability of the acquired
Accepted 25 May 2017
sound pressure levels. Experimental results concerning the adoption of the proposed method regarding
environmental noise from three different sources (road traffic, outdoor air conditioner fan motor and con-
struction site) confirm the reliability of the proposal and its feasibility in evaluating the equivalent sound
Keywords:
pressure level of an acoustic phenomenon using short-term indicators.
Noise control
Measurement uncertainty
Ó 2017 Elsevier Ltd. All rights reserved.
Measurement time interval
Statistical analysis
Bootstrap method

1. Introduction periods. If continuous measurements over such a period cannot be

One of today’s most important environmental problems is noise
pollution. In order to assess the acoustic environmental impact of
new infrastructure or the compatibility of the existing noise with
the maximum limits established by legislation [1], it is necessary
to arrive at an appropriate calculation of the environmental noise
equivalent continuous A-weighted sound pressure level. According
to [2], the measurement should optimally be carried out continu-
ously over the entire observation period. Such an approach is
prohibitively costly in terms of time and money. In practice there-
fore, environmental noise indicators are estimated from several
measurement time windows (episodes) randomly distributed
throughout the observation period, during day, evening and night
reference time.
Each measurement time interval has to contain representative
values of noise pressure levels to accurately calculate equivalent
sound pressure level LAeq. This is obtained by integrating and aver-
aging the squared A-weighted sound pressure of fluctuating noise
[3–6]. The actual form of the technical standards do not provide
practical criteria for the choice of the temporal distribution of
the samples, i.e. the number and duration of the measurement epi-
sodes. For example, ISO 1996-Part 2 states only ‘‘to select the mea-
surement time interval to cover all significant variations in sound
emission and propagation. If the noise shows periodicity, the measure-
ment time interval should cover an integer number of at least three
⇑ Corresponding author at: University of Salerno, Department of Industrial
Engineering Via Giovanni Paolo II, Nr. 132, 84084 Fisciano, Italy.
E-mail address: drusso@unisa.it (D. Russo).
made, measurement time intervals shall be chosen so that each one
constitutes a part of the cycle, so that, together, they represent the
complete cycle’’. Unfortunately, this condition is generally not
applicable for environmental noise where the acoustic signal is
random.
Some surveys [7–9] show that the reliability of the estimate of
environmental noise indicators depends largely on the time vari-
ability of the noise, so a sampling technique has to be chosen
according to address this parameter.
Drawing on previous studies, the authors introduce an original
method for choosing the minimum measurement time interval
that takes into account the statistical variability of the acoustic
phenomenon. The model may be adopted for the automatic
determination of the measurement episodes, by lowering the sam-
pling window necessary to estimate the short and/or long term
indicators.
2. State of the art
The road traffic noise is one of the main sources of pollution, so
the estimation of minimum measurement time interval represents
an important research topic, to which many studies have been
devoted.
In [10], with reference to 5 years of continuous noise measure-
ments of LAeq,24h carried out in Valencia (Spain), the appropriate
period of measurement over a 24-h noise intervals in order to cal-
culate the corresponding annual equivalent level has been investi-
gated. The findings offer very useful information on traffic noise

http://dx.doi.org/10.1016/j.apacoust.2017.05.032
0003-682X/Ó 2017 Elsevier Ltd. All rights reserved.
 C. Liguori et al. / Applied Acoustics 127 (2017) 126–132
measurement techniques. In particular, the sampling strategy with
a selection of at least 6 randomly chosen days provides an accurate
representation of the annual equivalent noise level.
In [11] long term equivalent levels of road traffic noise acquired
in Ladenburg, Germany, have been considered and different mea-
surement approaches have been discussed. The author demon-
strate that a sampling strategy of at least one week leads to the
uncertainty of less than ±1 dB in noise equivalent levels.
Jagniatinskis and Fiks in [12] focus their attention on a one-year
duration noise monitoring experiment in the town of Vilnius
(Lithuania) near an arterial road with intensive road traffic. They
observe that, under normal weather and source emission condi-
tions, the lowest uncertainty values of Lden occur when a total mea-
surement time of 7 consecutive days is considered.
In [13] the author determines the minimum time necessary for
accurately measuring the hourly equivalent level of road traffic
noise according to preset measurement uncertainty [14,15] on LAeq.
Analogously, in [16–18] the authors find that the minimum mea-
surement time interval should be calculated from the error associ-
ated with LAeq on the basis of various vehicle distributions. More
specifically, by measuring over intervals of less than an hour, they
find it is possible to have an accurate measurement of road traffic
noise, within a predetermined uncertainty range for the hourly val-
ues of LAeq.
A different approach based on probability studies focuses on the
determination of the minimum time for estimating a reliable value
of LAeq drawing cases in which the probability density function of
noise level is known theoretically or experimentally [19–21].
Some authors, such as Maruyama et al., address the issue of the
minimum measurement time interval T to evaluate the equivalent
sound pressure level LAeqT of road traffic noise from the viewpoint
of the reliability required for the estimation. In particular, in [22]
they consider the influence of four kinds of traffic variables (traffic
volume, average vehicle speed, percentage of heavy vehicles and
the number of vehicle transits). In [23] they introduce two types
of dynamic statistics: the mean time interval between two succes-
sive maximum A-weighted sound pressure levels observed during
the reference measurement time interval and the mean recurrence
time of the maximum A-weighted sound pressure level.
In [24], for the estimation of the traffic noise minimum mea-
surement time interval in the city of Hamadan (West of Iran), the
main roads are divided into 54 segments and 94 measuring sta-
tions are fixed. Field data obtained from 282 measurements,
including 2 daily-hour and one nightly-hour measurements, show
that 10-min interval measurement of equivalent sound pressure
level is able to forecast the hourly values of LAeq in each station.
Jagniatinskis et al. in [25] suggest the replacement of full year
measurement by choosing a shorter time interval using the idea
of representative time interval that contains an appropriate
amount of transportation noise events. In particular, they analyse
the representative sample definition for the cases of road, aircraft
and rail transport. In Lithuania, for example, a representative mea-
surement time interval for annual urban traffic noise assessment,
under normal weather conditions, is one week.
In [26] the authors focus their attention on the effectiveness of
using short time span measurements to monitor or assess the
acoustical environment. On the basis of data acquired in the
high-rise residential areas of Hong Kong, they observe that short
time interval results are much lower than the worst scenario of a
site, but the energy-based Day–night level and Day–evening–night
level are acceptable.
As a summary of the reported literature, most of the proposed
approaches refer to the estimation of the minimum measurement
time about long-term traffic noise indicator. The topic of the mea-
surement of the short term equivalent sound pressure level, which
may be independent from the previous knowledge (i.e. statistical
127
distribution and/or main influence parameters) of the acoustic
phenomenon, has not been sufficiently explored
3. The proposal
In order to determine the minimum measurement time interval
for the estimation of the environmental noise, a data-driven sam-
pling strategy is proposed, which takes into account the observed
variability associated with measured sound pressure levels. In par-
ticular, the data variability is estimated by adopting the popular
technique of the bootstrap [27], i.e. a statistical nonparametric
resampling method which replicates the initial dataset, without
any restrictions in terms of shape and properties of the statistical
distributions under consideration.
In this proposal, the bootstrap technique has been adopted to
determine the Confidence Interval (CI) of the short time statistic
LAeq, once the desired Confidence Level (CL) has been chosen. In lit-
erature, some rules of thumb have been suggested both for the
number of bootstrap samples and type of algorithm for CI calcula-
tion, such as the normal approximation (NORM), the t-student (T-
STUD), the basic percentile (PER), the bias corrected percentile
(CPER), and the bias corrected and accelerated percentile (BCA)
methods [28–32].
Typically, values (m) at least equal to 1000 should be considered
for CI calculation [33,34]. Moreover, as reported in [35], the CPER
method has been revealed the most repeatable algorithm for deter-
mining the CI of the short-term statistic and consequently it has
been implemented in this proposal. In particular, the CPER tech-
nique is a slight modification of the PER algorithm and it allows
the mean of the transformed estimate to differ from the population
mean: interval endpoints depend on the bias correction (z0), that is
calculated from the bootstrap sampling distribution.
As presented in Fig. 1, the proposed strategy hypothesizes a
minimum acquisition time corresponding to the number (N) of
sound pressure levels for assuring the statistical significance of
the starting dataset. Consequently, the minimum measurement
time Tmeas_min (resulting from the proposed algorithm) is forced
to be an integer multiple of the minimum acquisition time (as well
as dependent from the frequency 1/ Tconst of the recorded data by
the sound level meter).
The N A-weighted sound pressure levels Lt are considered to cal-
culate the corresponding equivalent level according to the
equation:
!
1 X
Nmin Lt
LAeq ¼ 10Log 1010 ð1Þ
Nmin t¼1
The CI of the above short term statistic (once the CL is fixed) is
determined by applying the CPER bootstrap method and consider-
ing m bootstrap samples (resampling the Lt dataset). In order to
take into account the random variability introduced by the boot-
strap method, k repetitions of the CI calculation are suggested to
determine the (mean) values for the interval width (DCI) and
extremes (CIlower and CIupper).
According to the strategy, the introduced steps are repeated for
calculating the CI information for each next acquisition time win-
dow as long as both the actual interval width and extremes show
a data variability lower than the one observed in the previous win-
dows. Furthermore, satisfaction of the following conditions (Eq.
(2)) are considered as a stopping rule:
     
DCIiþ1 < DCIi;updated AND CIlower
iþ1 i;updated
> CIlower AND CIiþ1 i;updated
upper < CI upper
¼ true; i P 1
ð2Þ
128 C. Liguori et al. / Applied Acoustics 127 (2017) 126–132

dB (A)

me

i≥1
Li 1 L2
i … Li Li+11 Li+1
2 … Li+1

k repeons k repeons

Bootstrap CPER CPER Bootstrap

sampling bootstrap bootstrap sampling
method method

Li1,1 Li1,2 … Li Li+11,1 Li+11,2 … Li+1

Average on k: Average on k:
CIi CIi CIi+1 CIi+1
Li2,1 Li2,2 … Li
(1- ) Li+12,1 Li+12,2 … Li+1
m
(1- ) CIiLower Cii+1Lower
m

CIiUpper Cii+1Upper

Lim,1 Lim,2 … Li Li+1m,1 Li+1m,2 … Li+1

Stop/Update
Rules

Minimum
measurement
me interval

Fig. 1. Algorithm used to obtain the minimum measurement time.

once the update rule (Eq. (3)) is applied to the (i + 1) time win- (a) For the measurement of road traffic noise, the sound level
dow next to the acquisition starting point. meter was placed in close proximity to the motorway (at a distance
of 1 m from the roadside at a height of 4 m).
8   Main parameters of the noise source, recorded during the mea-
  >
< CIlower
iþ1;updated
¼ min CIiþ1 i;updated
lower ; CI lower surement campaign, are reported in Table 1.
DCIiþ1 < DCIi;updated !  
>
: CIupper
iþ1;updated
¼ max CIiþ1 i;updated Data collection was performed during one day (wind speed less
upper ; CI upper
than 5 m/s and no rain) over the course of 3 consecutive hours
ð3Þ (starting time: 10 a.m.). With regard to the instrument
setting, the time history logging step (Tconst) was fixed at one
Finally, the minimum measurement time Tmeas_min is set to the
second, resulting in a dataset of 10,800 consecutive sound pressure
extent of the consecutive acquisition time windows before the stop
levels.
rule is satisfied.
(b) For the noise measurement of the air conditioner outdoor
fan motor (heat pump and liquid chiller) WSAN-XIN 141 of CLIVET,
the sound level meter was placed at a distance of 1 m from the
4. Experimental results nearest building facade and a height of 4 m.
Main parameters of the noise source, recorded during the mea-
In order to characterize the proposed strategy [36,37], a statis- surement campaign, are reported in Table 2.
tical analysis has been carried out on the measurements of envi- Data collection was carried out during one day (wind speed less
ronmental noise from three sources: (a) road traffic, (b) the than 5 m/s and no rain) with measurement taking place over
outdoor fan motor of an air conditioner and (c) a construction site. 50 min covering two different working conditions (namely start-
Measurement data were collected by using the Sound Level up and modulation state) of the air conditioner fan motor. With
Meter Larson Davis 831, Class 1 Environmental Noise & Building regard to the instrument setting, the time history logging step
Acoustics Analyzer.

Table 1
Main parameters of road traffic.

Average hourly traffic volume Mean speed

Light vehicles Heavy vehicles Light vehicles Heavy vehicles
Diurnal reference time (6–22 h) 172 veh/h 62 veh/h 113 km/h 93 km/h
Nocturnal reference time (22–6 h) 30 veh/h 25 veh/h 109 km/h 91 km/h
 C. Liguori et al. / Applied Acoustics 127 (2017) 126–132 129

Table 2
Main parameters of the air conditioner outdoor fan motor.

WSAN-XIN 141 Heating Capacity [kW] Cooling Capacity [kW] Sound Power Level (dB)
Octave bands (Hz)
63 125 250 500 1000 2000 4000 8000 dB(A)
32.36 33.19
85.0 83.0 80.0 79.0 83.0 75.0 72.0 61.0 85.0

Table 3
Main noise sources at the construction site.

Truck (Eurotrakker 410 – Iveco) Excavator (318BLN – Caterpillar) Concrete mixer (FG3 – Unimec) Circular saw (Master 03CMF-Edilsider)
Lwmax [dB(A)] 103 104 80 113

Fig. 2. Normal probability plot for collected dataset of environmental noise of different sources: (a) road traffic; (b) outdoor air conditioner fan motor; (c) construction site.
130 C. Liguori et al. / Applied Acoustics 127 (2017) 126–132

Table 4  k = 30 repetitions of the bootstrap method for each time

Application of Anderson-Darling Test to the sound pressure levels dataset of interest. window.
Noise Source p-value Test result
Road Traffic 0.0005 H0 refused The minimum number N of sound pressure levels has been var-
Air Conditioner 0.0005 H0 refused ied in the range [10500] to analyse the repeatability of the algo-
Construction site 0.0786 H0 not refused rithm. To this aim, the proposed strategy has been executed L
multiple times (L = 100) for an acquisition starting point randomly
was fixed at 0.1 s, resulting in a dataset of 30,000 consecutive chosen within the observation period of interest. For each type of
sound pressure levels. the noise source, the corresponding measurement time (Tmeas_min)
(c) For the noise measurement of construction site, the sound is not influenced by the statistical variability introduced by the
level meter was placed at a distance of 2 m from the noise source bootstrap method: the observed repeatability of the estimated
and a height of 4 m. Main noise sources at the considered construc- minimum measurement time was equal to 100%.
tion site with corresponding sound power levels are reported in Experimental results take into account the mean value for:
Table 3:
Data collection was performed during two days (wind speed i. minimum measurement time Tmeas_min;
less than 5 m/s and no rain) over the course of 24 consecutive ii. the 95%-CI for the equivalent sound pressure level LAeq,
hours (starting time: 10 p.m.). In terms of instrument setting, the Tmeas_min corresponding to Tmeas_min;

time history logging step was fixed at one minute, resulting in a
dataset of 1440 consecutive sound pressure levels.
The normal probability plot corresponding to the different data-
sets are reported in Fig. 2, whereas the Anderson-Darling Test with
5% significance level (see Table 4) leads to refuse the null hypoth-
esis (H0: ‘‘dataset is Gaussian”), showing the examined populations
are not strictly Gaussian. Thus, for the CI estimation, use had been
made of the bootstrap method (CPER) [38–41].
The proposed algorithm for determining a significant measure-
ment time has been applied to each dataset according to the fol-
lowing parameters:
 application of the CPER bootstrap method with m = 1000 boot-
strap samples;
 a = 0.05 for determining the CI corresponding to 95% confidence
level;
iii. the maximum observation period Tobs (from the starting
point) for which the 95%-CI for the equivalent sound pres-
sure level LAeq,Tobs, (by taking into account Nlevels = Tmeas_min
/ Tconst sound pressure levels randomly distributed) is com-
patible with the results achieved for Tmeas_min.
In order to analyse the stability of the proposed algorithm (in
terms of the resulting Tmeas_min) with respect to the noise source
 
variability, a sliding window  N4 samples has been considered
around the starting point. Then, the relative occurrence OTmode of
the observed mode for Tmeas_min achieved correspondingly to
(N/2 + 1) starting points within the sliding window has been deter-
mined for all the noise sources. As an example, the relative occur-
rence for different values of N is reported in Fig. 3 about the road
traffic noise. The better stability at N increasing may be observed
as a general behaviour. Thus, a threshold about the relative occur-

Fig. 3. Relative occurrence of the mode for Tmeas_min observed for different sliding window width (road traffice noise).

Table 5
Minimum measurement time and 95% confidence intervals for the equivalent sound pressure level.

Noise Source Tconst [s] N Tmeas_min [s] 95%-CI (LAeq,Tmeas_min) 95%-CI (LAeq,Tobs)
Lower limit Upper limit Width [dB] Tobs [s] Lower limit Upper limit Width
LAeq,lower [dB] LAeq,upper [dB] LAeq,lower [dB] LAeq,upper [dB] [dB]
Road Traffic 1.0 300 1200 64.1 65.0 0.9 10800 64.3 65.2 0.9
Air Conditioner 0.1 300 120 61.4 61.6 0.2 600 61.2 61.4 0.2
(Start-up)
Air Conditioner 0.1 400 360 57.7 58.0 0.3 1080 57.2 57.8 0.6
(Modulation State)
Construction site 60 10 3600 69.7 70.3 0.6 28800 69.0 69.7 0.7
 C. Liguori et al. / Applied Acoustics 127 (2017) 126–132 131

rence may be adopted as criterion for choosing the parameter N of
the proposed approach.
The experimental results are summarized in Table 5, where the
reported N is the minimum value which assures a relative occur-
rence OTmode greater than 95% for the corresponding noise source.
You may note that the proposed algorithm has been revealed
effective in the estimation of the acoustic noise from the observa-
tion of short episodes. For each noise source (see Fig. 4), the short-
term indicator achieved from the minimum measurement time
Tmeas_min (highlighted as red line) is able to predict the behaviour
of the acoustic noise observed during the time window Tobs (high-
lighted as green line), that is longer from 3 to 9 times, respectively
for the air conditioner working in modulation state and the road
traffic.
More in details, independently from the noise source, the mea-
surement results (95%-CI for the equivalent sound pressure level)
corresponding respectively to the minimum measurement time
Tmeas_min and the observation window Tobs are compatible, i.e. the

Fig. 4. Sequence of sound pressure levels during observation period (solid green) and resulting minimum measurement time (solid red), LAeq 95% CI (dash black) for
observation period, LAeq 95% CI (dash red) for measurement time about different noise sources: a) traffic road; b) outdoor air conditioner fan motor; c) construction site.
132 C. Liguori et al. / Applied Acoustics 127 (2017) 126–132
interval intersection is different from the empty set. Thus, the mea-
surement episode detected through the proposed algorithm is well
representative of the inner variability of the measurand.
5. Conclusion
The paper describes a novel statistical methodology for the
evaluation of the minimum measurement time interval during
which the estimation of environmental noise should be assessed
using short-term indicators and the corresponding Confidence
Interval.
More in details, the statistical analysis about the main parame-
ters of the 95% confidence intervals is continuously carried out
with respect to the acquisition time windows within the measure-
ment episode, as long as both the actual interval width and
extremes show a data variability greater than the one observed
in the previous windows.
The experimental verification of the proposal has been per-
formed against real data acquired in three non Gaussian scenarios
(road traffic, outdoor air conditioner fan motor and construction
site as noise sources).
As main result, a very good stability in determining the mini-
mum measurement time interval is showed with respect to the
starting point of the measurement episode and a criterion for the
optimum choice of the minimum acquisition time is suggested in
terms of the observed mode for multiple repetitions.
Moreover, the proposed model reveals to be effective in the pre-
diction of the short-term noise indicator corresponding to longer
observation period at measurement and influence parameters
(equivalent sound pressure level, observation period, time history
logging step) varying in a broad range. Thus, the suggested data-
driven approach may be adopted in the field of the environmental
acoustics as guideline in the choice of short measurement episodes
during the acquisition campaign.
References
[1] Ruggiero A, Russo D, Sommella P. Determining environmental noise
measurement uncertainty in the context of the italian legislative framework.
Measurement 2016;93:74–9. http://dx.doi.org/10.1016/j.
measurement.2016.07.007.
[2] ISO 1996–2 Acoustics - Description, assessment and measurement of
environmental noise - Part 2: Determination of sound pressure levels.
[3] Ohta M, Mitani Y. Time series prediction method for stochastic acoustic signals
by the use of an adaptive function model. Measurement 1989;7(2):68–76.
[4] Heimann D, Schady A. Derivation of sound-level characteristics to assess traffic
development scenarios. Appl Acoust 2016;103:1–9.
[5] Deželak F, Čurović L, Čudina M. Determination of the sound energy level of a
gunshot and its applications in room acoustics. Appl Acoust 2016;105:99–109.
[6] Ruggiero A, Russo D. Acoustical design and experimental verification of school
music rooms: A case study. Appl Acoust 2016;107:1–9. http://dx.doi.org/
10.1016/j.apacoust.2015.12.023.
[7] Liguori C, Paolillo A, Ruggiero A, Russo D. Towards the Evaluation of the
Measurement Uncertainty of Environmental Acoustic Noise. In: Proceedings of
2015 IEEE international instrumentation and measurement technology
conference (I2MTC 2015), Pisa, Italy, May 11–14, 2015, p. 1238–42. http://
dx.doi.org/10.1109/I2MTC.2015.7151450.
[8] Liguori C, Paolillo A, Ruggiero A, Russo D. Outlier detection for the evaluation of
the measurement uncertainty of environmental acoustic noise. IEEE Trans
Instrum Meas 2015;65(2):234–42. http://dx.doi.org/10.1109/
TIM.2015.2491038.
[9] Liguori C, Paolillo A, Ruggiero A, Russo D. A Preliminary Study on the
Estimation of the Uncertainty of Traffic Noise Measurements. In: Proceedings
of 2016 IEEE international instrumentation and measurement technology
conference (I2MTC 2016), Taipei, Taiwan, May 23–26, 2016, p. 246–51. http://
dx.doi.org/10.1109/I2MTC.2016.7520368.
[10] Gaja E, Gimenez A, Sancho S, Reih A. Sampling techniques for the estimation of
the annual equivalent noise level under urban traffic conditions. Appl Acoust
2003;64(1):45–53.
[11] Kuehner d. Long-term Leq errors expected and how long to measure
(uncertainty and noise monitoring). Forum Acusticum 2005, Budapest
(Hungary), Aug 29 – Sep 2; 2005.
[12] Jagniatinskis A, Fiks B. Assessment of environmental noise from long-term
window microphone measurements. Appl Acoust 2014;76:377–85.
[13] De Donato SR. Estimation of the minimum integration time for determining
the equivalent continuous sound level with a given level of uncertainty
considering some statistical hypotheses for road traffic noise. Noise Control
Eng J 2007;55(6):526–31.
[14] Joint Committee for Guides in Metrology. Evaluation of Measurement Data:
The Role of Measurement Uncertainty in Conformity Assessment, JCGM
106:2012:2012 BIPM, Sèvres, 2012.
[15] JCGM 100:2008, Evaluation of measurement data – Guide to the expression of
uncertainty in measurement.
[16] Garg N, Kumar A, Sanjid MA, Chaudhary KP, Maji S. Reaffirmation of
measurement uncertainty in pressure sensitivity determination of LS2P
microphones by reciprocity method. Measurement 2014;51:281–8.
[17] Černetič J, Čudina M. Estimating uncertainty of measurements for cavitation
detection in a centrifugal pump. Measurement 2011;44(7):1293–9.
[18] Khodabandeh M, Mohammad-Shahri A. Uncertainty evaluation for an
ultrasonic data fusion Based target differentiation problem using generalized
aggregated uncertainty measure 2. Measurement 2015;59:139–44.
[19] Skarlatos D, Manatakis E. Noise probability density function for poisson type
traffic flow. Appl Acoust 1989;27:47–55.
[20] Skarlatos D, Drakatos P. On selecting the minimum observation time for
determining the Leq of a random noise with a given level of confidence. J
Sound Vib 1992;152(1):141–8.
[21] Skarlatos D. A numerical method for calculation of probability density function
of equivalent level in the case of traffic noise. Appl Acoust 1993;38:37–50.
[22] Maruyama M, Kuno K, Sone T. The minimum measurement time for estimating
LAeqT of road traffic noise from the number of vehicle pass-bys. Appl Acoust
2013;74(3):317–24.
[23] Maruyama M, Kuno K, Sone T. Minimum measurement time interval to
estimate a reliable Sound pressure level of road traffic noise using two types of
dynamic statistics. Appl Acoust 2014;83:150–5.
[24] Abbaspour M, Golmohammadi R, Nassiri P, Mahjub H. An investigation on
time-interval optimisation of traffic noise measurement. J Low Freq Noise, Vib
Active Contr 2006;25(4):267–73.
[25] Jagniatinskis A, Fiks B, Zaporozhets O. Annual assessment of transport noise
using representative time measurements. Procedia Eng 2016;134:301–8.
[26] Ng CH, Tang SK. On monitoring community noise using arbitrarily chosen
measurement periods. Appl Acoust 2008;69(7):649–61.
[27] Wehrens R, Putter H, Buydens LMC. The bootstrap: a tutorial. Chemometr
Intell Lab Syst 2000;54(1):35–52.
[28] Betta G, Capriglione D, Pietrosanto A, Sommella P. A statistical approach for
improving the performance of a testing methodology for measurement
software. IEEE Trans Instrum Meas 2008;57(6):1118–26.
[29] Williams J, MacKinnon DP. Resampling and distribution of the product
methods for testing indirect effects in complex models. Struct Equation
Model: a Multidiscip J 2008;15(1):23–51.
[30] Ste˛pień B. Bootstrap confidence intervals for noise indicators. Acta Acustica
United with Acustica 2016;102(2):389–97.
[31] Burbrink FT, Pyron RA. The taming of the skew: estimating proper confidence
intervals for divergence dates. Syst Biol 2008;57(2):317–28.
[32] Polansky AM. Stabilizing bootstrap-t confidence intervals for small samples.
Canadian J Stat 2000;28(3):501–16.
[33] Davison AC, Hinkley DV. Bootstrap methods and their application, vol.
1. Cambridge University Press; 1997.
[34] Efron B, Tibshirani RJ. An introduction to the Bootstrap. London: Chapman and
Hall; 1993.
[35] Liguori C, Ruggiero A, Sommella P, Russo D. Choosing bootstrap method for the
estimation of the uncertainty of traffic noise measurements. IEEE Trans
Instrum Meas 2017;66(5):869–78. http://dx.doi.org/10.1109/
TIM.2016.2627260.
[36] Liguori C, Ruggiero A, Russo D, Sommella P. Innovative bootstrap approach for
the estimation of minimum measurement time interval in road traffic noise
evaluation. Measurement 2017;98:237–42. http://dx.doi.org/10.1016/j.
measurement.2016.12.008.
[37] Liguori C, Ruggiero A, Russo D, Sommella P. Influence of measurement time on
acoustic noise uncertainty. In: Proceedings of 2017 IEEE international
instrumentation and measurement technology conference (I2MTC 2017),
Torino, Italy, May 22–25, 2017, p. 1173–78.
[38] Tang SK, Au WH. Statistical structures of indoor traffic noise in a high-rise city.
J Acous Soc America 1999;106(6):3415–23.
[39] Batko W, Ste˛pień B. Application of the Bayesian inference for estimation of the
long-term noise indicators and their uncertainty. Acta Phys Pol, A 2011;119
(6A):916–20.
[40] Batko W, Ste˛pień B. Type A standard uncertainty of long-term noise indicators.
Arch Acous 2014;39(1):25–36.
[41] De la Rosa JI, Fleury GA. Bootstrap methods for a measurement estimation
problem. IEEE Trans Instrum Meas 2006;55(3):820–7.