Uncertainties of room average
sound pressure levels measured
in the field according to the
draft standard ISO 16283-1
Experiences from a few case studies
Christian Simmons
AkuLite Report 3
SP Report 2012:28
Uncertainties of room average sound pressure
levels measured in the field according to the draft
standard ISO 16283-1
Experiences from a few case studies
Christian Simmons
Abstract
Results according to a new draft ISO standard for measuring sound pressure levels are presented.
The proposed methods are based on various spatial sampling techniques, where a microphone is
moved continuously or kept steady at fixed positions in different parts of the room. The uncertainty of
the average is to a large extent related to the ability of the sampling method to sample sound pressure
levels uniformly from all parts of the room.
The uncertainties of the simplified methods of the draft standard have been estimated empirically by
means of measurements made in five rooms with different acoustic conditions. The result of each type
of simplified method is compared to an average of sound pressures recorded in a dense mesh of microphone positions throughout the permitted space in the same room.
These results are also being submitted to Noise Control Engineering Journal.
Key words: building acoustics, ISO 140, ISO 162 83, measurement uncertainty
SP Sveriges Tekniska Forskningsinstitut
SP Technical Research Institute of Sweden
SP Rapport 2012:28
ISBN 978-9187017-42-1
ISSN 0284-5172
Borås
2
Contents
Preface
4
Summary
5
Sammanfattning
6
Introduction
7
Methods
8
Precision averaging of sound pressure levels in a dense mesh
8
Simplified averaging of the sound pressure levels by moving microphone paths
9
Standard deviation of sound pressure levels
10
Reverberation times
11
The test rooms
11
Example of spatial variation of sound pressure levels (living room)
14
Statistical analyses of sound pressure levels
16
New loudspeaker positions, differences in the mesh average sound pressure level
21
Discussion of results and some practical experiences
22
References
24
Appendix
− Nya fältmätningsmetoder från ISO. Artikel i Bygg & teknik 2012 (in Swedish)
− Impact sound pressure level (diagram)
3
Preface
This report presents results that have been financed within the national Swedish project AkuLite –
Acoustics and vibrations in light weight buildings. AkuLite involves all Swedish research institutions
active in the field, leading industries and consultants. Vinnova and Formas are the public funders. The
project started in late 2009 and will be finalised in early 2013.
The work has been carried out through the company Simmons akustik & utveckling ab.
The main part of this report has been submitted for publication through peer review in Noise Control
Engineering Journal. A diagram and an article in the Swedish technical journal Bygg & teknik has
been appended.
Klas Hagberg
AkuLite project manager
SP Wood Technology and WSP Acoustics
Christian Simmons
SP Acoustics
4
Birgit Östman
Senior advisor
SP Wood Technology
Summary
A draft standard was presented by a working group within ISO in April 2011. It describes six methods
to simplify the measurement of a spatially averaged sound pressure level in a room, in order to determine an airborne sound insulation between two rooms. The draft standard is intended to replace ISO
140 part 4 after revisions and balloting procedures.
The proposed methods are based on various spatial sampling techniques, where a microphone is
moved continuously or kept steady at fixed positions in different parts of the room. The uncertainty of
the average is to a large extent related to the ability of the sampling method to sample sound pressure
levels uniformly from all parts of the room.
The uncertainties of the simplified methods of the draft standard have been estimated empirically by
means of measurements made in five rooms with different acoustic conditions. The result of each type
of simplified method is compared to an average of sound pressures recorded in a dense mesh of microphone positions throughout the permitted space in the same room.
Some results that may be useful when an averaging method is to be decided:
•
the standard deviations may actually be higher above 100 Hz than below, the 160 and 200 Hz
third octave bands may even contain the most uncertain results
•
the fixed positions method is practical and may be used in all types of room
•
the special corner method gave higher average sound pressure levels and lower uncertainties
compared to the other methods
•
moving microphone methods are difficult to apply in small furnished rooms where there is not
enough space to complete several independent microphone paths
•
moving microphone methods may be sensitive to operator generated background noise
•
microphone positions should be spread over the entire room volume
This study has not estimated the uncertainty of the sound pressure level difference between two
adjacent rooms, used for sound reduction index estimates.
5
Sammanfattning
En arbetsgrupp inom ISO för nya fältmätningsmetoder inom byggakustikområdet (ISO TC 43/SC
2/WG 18) har gett ut ett förslag till standard (ISO 16283-1), som ska ersätta dagens ISO 140-standard
för luftljudsisolering. Kommande del -2 beskriver mätning av stegljudsnivå och del -3 för fasadisolering.
En central del i alla mätmetoder är hur man ska bilda medelvärden av ljudtrycksnivån i ett rum. För
stegljudsnivå är det viktigt att veta hur man ska placera hammarapparaten, särskilt på anisotropa lätta
bjälklag som har olika ”mottaglighet” i olika delar. För att undersöka hur metoderna i standardförslaget fungerar i praktiken, samt att bidra med praktisk information till standardiseringskommittén, har en
serie praktiska prover utförts inom det nationella forskningsprojektet AkuLite.
Denna rapport består av två delar. Utvärderingarna beskrivs i detalj i en rapport på engelska, som godkänts för publicering i den vetenskapligt granskade tidskriften Noise Control Engineering Journal
(NCEJ). Efter denna följer en kort svensk artikel i tidskriften Bygg & teknik.
I korthet visar våra försök:
• att standardavvikelserna faktiskt kan vara högre över 100 Hz än vid lägre frekvenser, 160 och 200
Hz tredje oktavbanden kan innehålla de mest osäkra resultaten
• att metoden med fasta mikrofonpositioner är praktisk och kan användas i alla typer av rum
• den speciella hörnmetoden gav högre genomsnittliga ljudtrycksnivåer och lägre slumpmässiga
skillnader jämfört med andra metoder
• metoder med svepande mikrofon är svåra att tillämpa i små möblerade rum där det inte finns tillräckligt utrymme för att utföra flera oberoende mikrofonbanor – man tenderar att mäta i mitten av
rummet vilket ger systematiska fel
• metoder med manuell svepning av mikrofonen kan vara känsliga för att operatören själv genererar
störande bakgrundsljud
• mikrofonpunkter bör spridas över hela rummets volym
I denna studie har vi inte uppskattat osäkerheten i ljudtrycksnivåskillnaden mellan två angränsande
rum, som används för ljud uppskattningar av reduktionstal.
6
Introduction
A draft standard was presented by a working group within ISO in April 2011. The draft standard ISO
16283-1 describes several methods to simplify the measurement of spatially averaged sound pressure
levels for the purpose of determining the airborne sound insulation between two rooms. It is intended
to replace ISO 140 part 4 after revisions and balloting procedures. The same measurement methods
may be used in other parts of this standard, for the measurement of impact sound insulation (part 2)
and façade sound insulation (part 3).
The proposed averaging methods are based on various spatial sampling techniques, where a microphone is moved continuously or placed at fixed positions in each room. The uncertainty of a method is
to a large extent related to its ability to sample sound pressure levels uniformly from all parts of the
room.
There are also extraneous factors that may cause unexpected results, e.g. the influence of furniture,
small room sizes, irregular room shapes, radiating surfaces, stability of the sound source or background noise, most noticeably noise caused by the operator moving the microphone.
In this brief field study, the uncertainties of the simplified averaging methods of the draft standard
have been estimated empirically by means of sound pressure level measurements made in 5 rooms
with different acoustic conditions. The result of each type of simplified method is normalized
(compared) to a ”reference” average of sound pressures recorded in a dense mesh of microphone
positions throughout most of the permitted space of each room, according to the draft standard [1]. A
few modifications of the methods of the draft standard are proposed in this paper, but no other
averageing method has been included in the comparisons.
This study does not examine whether the proposed measurement methods correctly assess the energy
content of a room, nor the sound reduction index between two adjacent rooms.
It was beyond the scope of this empirical study to include a theoretical background of sound pressure
variations in enclosed spaces. However, some earlier studies give a theoretical background.
Uncertainties related to spatial variations of the sound field have been estimated by Hopkins [2, 3]. A
general study on the uncertainty in a spatial averaged sound pressure level was published in 1992 by
Olesen [4], which were later used to write the technical report ISO 140 TR/13 and the standard ISO
140-14. Recently, Pedersen et al compared various techniques to use corner positions in order to
capture the highest sound pressure levels in the room and they compared their results to the room
average [5].
An empirical study similar to the one presented in this report was published by the author in 1996, [6],
where a variety of spatial sampling techniques were simulated and compared with the averages of all
measured positions in each room. References to other studies are given as well in the references [2-6].
7
Methods
Precision averaging of sound pressure levels in a dense mesh
The spatial average of the sound pressure level in each room was determined by measurements with
two microphones mounted in a stand with vibration insulation. The microphone positions were placed
in a mesh 0,7 x 0,7 m (i.e. 0,7 m distance between two neighboring positions), starting 0,5 meter (m)
from two walls according to the provisions of the draft standard [1]. The resulting distance to the two
opposing walls varied between 0,3 and 0,7 m (it was chosen to keep the same resolution of the mesh in
all rooms). One microphone was fixed 1,7 m above the floor, the second 0,7 m above the floor, hence
the distance between the vertical positions was 1,0 m. However, in a few positions, the distances to
beds and chairs were closer than 0,5 m but at least ≥ 0.2 m. In a few positions, the lower microphone
had to be raised to 0.9 m above floor to keep clear from protruding furniture (≥ 0,2 m). The number of
microphone positions of the mesh are stated for each room below, as well as the reverberation times.
The measurement equipment conforms to the provisions of the draft standard which has been verified
by an accredited calibration laboratory.
The spatial average of the squared sound pressures recorded in all positions of the mesh was taken as a
reference later on, when the results obtained by simplified methods (manual scanning and fixed positions) were normalized. The loudspeaker was kept in the same position in each comparison.
This choice of the microphone mesh follows from the provisions of the draft standard and from some
practical limitations. It would have been beneficial to include at least one more height and to keep the
same distance between microphone positions also in the vertical direction. However, the choice of
mesh positions does not affect the comparison of the methods as presented below, since all results
with the simplified methods are normalized to the same mesh average. The standard deviations
presented below have not been normalized.
It has not been examined how well the mesh averages correlate to the subjective impressions of sound
in the test rooms, nor how they correlate with the total energy content of the room including pressures
close to the boundaries. In some studies [2, 5], microphones have been positioned very close to the
walls. If positions very close to the walls would be considered for a standardized method, it should be
studied whether local and non-propagating near fields could influence the microphone signal at
frequencies below the coincidence frequencies. Ceilings and walls with low weight plaster boards
have rather high coincidence frequencies and the effect of near fields should be studied further.
The sound was emitted by a stable MLS pink noise sound generator through a full range loudspeaker,
placed on the floor close to one corner, facing the corner. The measurement time was 16 seconds for
each fixed position. Microphone positions closer to the loudspeaker than 1 m were omitted. It was
verified for each reading that the 50 Hz third octave band was stabilized within 0,3 dB and that the Aweighted level was steady @85 dB +/- 2 dB during each measurement sequence. The S/N ratio was
better than 10 dB (25 Hz – 8 kHz) in all readings. The sound pressure levels were not normalized to a
0,5 seconds reference reverberation time, only to the reference spatial average taken in the mesh of
positions within the same room.
The same microphone positions were used to determine a mesh average a second time, with the loudspeaker in a different position, on a table or on the floor not close to any wall. The purpose was to
study whether the mesh average and the spatial variation (standard deviation) of the sound levels
changed when the loudspeaker was moved. However, these second sets of measurements were not
included in the comparison of the simplified methods of scanning the sound field discussed below.
The results of this study relate to the determination of the spatial average sound pressure level in one
room which is highly relevant e.g. for impact sound pressure level measurements. For airborne sound
insulation, it has not been studied whether the uncertainties presented below reflect the uncertainty of
the sound pressure level difference between two adjacent rooms.
8
Simplified averaging of the sound pressure levels by moving microphone
paths
The spatial averages and standard deviations were calculated for each of several moving microphone
paths as well as sets of fixed positions, as proposed by the draft standard. (Moving microphone paths
are denoted manual scanning paths in the draft standard). The average of each method was normalized
to the spatial average of sound pressure levels recorded in the dense mesh of positions in the same
room, using the same loudspeaker position, as described in the previous paragraph.
The draft standard proposes four moving microphone paths. Methods number 1, 2, 3 and 4 are
illustrated by Figure 1.
Figure 1. Moving microphone paths (1 – 4) used to determine spatial averages of the sound pressure
level according to the draft standard ISO/WD 16283-1 [1]
Method (5) is denoted ”Five fixed positions”. It gives the average of sound pressure level readings
with the microphone in Five fixed positions placed in different parts of the room. This procedure was
repeated 5 times, where all readings were taken in new positions (i.e. all readings used were unique for
each room average).
Method (6) is denoted ”Five fixed positions and Max corner”. It combines the ”Five fixed positions”
room average with a special corner level, determined according to the draft standard. The sound
pressure level is measured in all corners and the highest third octave band level of all corners is then
averaged with the room average. The room average LLF is weighted according to equation (13) of the
draft standard such that the ”Five fixed positions” room average level is weighted by a factor of 2/3
and the highest corner level by 1/3
LLF
10lg
10
,
10
,
(1)
9
In the figures and discussion below, the LLF is denoted ”Five fixed positions and Max corner”.
Figure 2. Corner microphone positions for method (6), proposed in [1] to reduce uncertainty of the
spatial averages of the sound pressure level in small rooms in the three third octave bands 50-63-80
Hz. One microphone position was 0,3 m from the walls and ceiling, the second microphone was at 0.6
m distance to examine whether the choice of distance influences the result.
During the measurements, it was realized that it was not so easy to keep the microphone at the
prescribed distances. In one room, the repeatability was determined by 5 readings, with the distances
varying within realistic tolerances, taken as approximately +/- 0,1 m. The standard deviation between
each corner maximum level tested in all corners was 0,4-0,7 dB.
Standard deviation of sound pressure levels
Since the main purpose of a measurement is to estimate the spatial and time average sound pressure
level of a room, it may be questioned whether the straight forward standard deviation of the ”dBvalues” (denoted s1) is appropriate for the purpose of estimating the uncertainty of the room average,
e.g. as a confidence interval. For the spatial average, the standard uses a logarithmic average, i.e. the
logarithm of the arithmetic average of the squared sound pressures, also denoted the energetic average.
This value is typically higher than the arithmetic average of the ”dB-values” (i.e. of the levels).
In the previous studies [4, 5], it was shown that zones with positive interferences have sound pressure
levels quite close to the room average, but that zones with destructive interferences may vary considerably. Thus, it was examined whether other approaches to estimate the uncertainty would return a more
relevant estimate of the uncertainty of the spatial average sound pressure level.
A simple procedure was used to search for a better descriptor. To serve as a reference, the maximum
sound pressure level in each frequency band of all readings was normalized by the logarithmic
average. This difference was divided by a coverage factor of approximately 3,0, taken from the
Student t’s distribution (for the relevant degrees of freedom and a probability value of 99 %). The
result is a simple estimate of the standard deviation of the logarithmic average (denoted s2), being
representative for sound levels in the range from the average to the maximum.
However, maximum values may be uncertain. A more stable estimate of the standard deviation
(denoted s3), relevant for the levels around or higher than the room average, was obtained from the
difference between the squared sound pressures
10
s3 = 10 lg (p2avg + p2std) – 10 lg (p2avg)
(2)
where
p2avg is the arithmetic time and spatial average of the squared sound pressures
p2std is the standard deviation of the time averaged squared sound pressures
By comparison, it was found that s3 was much closer to s2 than s1. Thus, s3 is used throughout this
report and it is denoted ”Standard deviation’” although this may be incorrect.
Reverberation times
Reverberation times were measured in each room, in 7 positions, using 7 decays in each position, for
the same loudspeaker position in each room. The figures below show that the rooms chosen for this
study have quite different reverberation times.
The sound pressure levels have not been normalized to any reference reverberation time (nor absorption area). The Nordtest report by Olesen [4] gives a method to estimate the spatial variation of sound
pressure levels on the basis of the reverberation times. However, such estimations have not been
included in this study.
The test rooms
Five rooms were used for the measurements, chosen to be representative for real and often rather
complicated conditions in buildings. They have different dimensions, shapes and reverberation times.
The cylinders on the floor mark the mesh of microphone positions.
Another purpose of the photos is to illustrate the difficulties to spread the microphone positions
uniformly throughout a furnished room.
11
Reverberation times (s) vs frequency (Hz)
Length (m)
4,5
Width (m)
50
63
80
100
125
160
200
250
315
400
500
630
800
0,7
0,7
0,6
0,6
0,5
0,4
0,5
0,4
0,5
0,4
0,4
0,5
0,5
4
Height (m)
2,8
Mesh
6x6
1000 1250 1600 2000 2500 3150 4000 5000
0,5
0,5
0,5
0,6
0,5
0,5
0,5
0,4
Figure 3. Photos showing the furniture in the living room. The table shows the reverberation times 505000 Hz and the dimensions of the room. The microphone positions are marked by colored cylinders
on the floor.
Reverberation times (s) vs frequency (Hz)
Length (m)
4,0
Width (m)
50
63
80
100
125
160
200
250
315
400
500
630
800
1,7
0,7
2,6
0,5
0,3
0,5
0,3
0,3
0,3
0,3
0,3
0,3
0,4
2,1
Height (m)
2,5
Mesh
3x5
1000 1250 1600 2000 2500 3150 4000 5000
0,4
0,4
0,4
0,4
0,4
0,4
0,4
0,4
Figure 4. Photos showing the furniture in the small bed room. The table shows the reverberation times
50-5000 Hz and the dimensions of the room. The microphone positions are marked by colored
cylinders on the floor, table and bed.
12
Reverberation times (s) vs frequency (Hz)
Length (m)
3,8
Width (m)
50
63
80
100
125
160
200
250
315
400
500
630
800
0,7
0,4
0,3
0,2
0,2
0,2
0,2
0,2
0,3
0,3
0,3
0,3
0,3
3,3
Height (m)
2,5
Mesh
4x5
1000 1250 1600 2000 2500 3150 4000 5000
0,3
0,3
0,3
0,3
0,3
0,3
0,3
0,3
Figure 5. Photos showing the furniture in the mid-sized bed room. The table shows the reverberation
times 50-5000 Hz and the dimensions of the room. The microphone positions are marked by colored
cylinders on the floor and couch.
Reverberation times (s) vs frequency (Hz)
Length (m)
4,2
Width (m)
50
63
80
100
125
160
200
250
315
400
500
630
800
1,9
1,4
1,1
1,0
0,9
1,0
1,2
1,2
1,1
1,1
1,1
1,0
1,0
3,3
Height (m)
3,5
Mesh
4x5
1000 1250 1600 2000 2500 3150 4000 5000
1,0
1,0
0,9
1,0
1,0
0,9
0,9
0,8
Figure 6. Photos showing the small reverberation chamber and some diffusing items present in the
room (that are removed when the laboratory is used). The table shows the reverberation times 505000 Hz and the dimensions of the room. The microphone positions are marked on the floor but are
not visible on these photos.
13
Reverberation times (s) vs frequency (Hz)
Length (m)
7,7
Width (m)
50
63
80
100
125
160
200
250
315
400
500
630
800
3,5
3,0
3,6
2,5
2,1
2,2
2,5
2,1
2,1
2,1
2,2
3,5
2,5
5,5
Height (m)
3,5
Mesh
7x10
1000 1250 1600 2000 2500 3150 4000 5000
2,3
2,1
2,0
1,9
1,6
1,4
1,4
1,3
Figure 7. Photos showing the large reverberation chamber. Table shows the reverberation times 505000 Hz and the dimensions of the room. There were no diffusing items in the room, but the ceiling
and walls were made of light plaster boards with different weights. The microphone positions are
marked by colored plates on the floor.
Example of spatial variation of sound pressure levels
(living room)
In order to illustrate some typical spatial variations of the sound pressure level in third octave bands,
two 3 dimensional (3D) diagrams from the living room are presented below. The first diagram shows
the distribution of sound pressure levels in the 63 Hz third octave band, the second refers to the 250
Hz third octave band. A total of 36 microphone positions in 6 ’rows’ and 6 ’columns’ are presented by
the horizontal x- and y-axes in Figure 8. The sound pressure level in each position is plotted on the
vertical z-axis. The surface of the 3D-diagrams is interpolated from the discrete positions of the mesh.
The position of the microphone was 1,7 m above the floor. The loudspeaker position (6-6) is indicated
to the upper right in the figures. The minimum distance from the microphone to the loudspeaker is 1.4
meter.
Note: There was no measurement result in position 6-1 because the stove is situated here. These voids
in the 3D-diagrams are indicated as minimum values in lower right hand corner (6-1) of the figures,
since this kind of plots requires a value larger than zero.
14
Sound
pressure
level,
dB
rel
20 μPa
Void
pos.
Sound
pressure
level,
dB
rel
20 μPa
Void
pos.
Figures 8. Living room. Sound pressure levels in a grid of 6x6 microphone positions. Upper figure:
the 63 Hz third octave band, lower figure the 250 Hz third octave band. The microphone was placed
1,7 m above the floor. The loudspeaker position is indicated to the upper right in both figures. There
was no microphone in position 6-1, in the lower right hand corner of both figures (void), because of
the stove (see photos in Figure 3).
15
Statistical analyses of sound pressure levels
The results from the measurements in each room are presented in two types of figures. The first figure
shows the energetic average sound pressure level (SPL) normalized by the mesh average in the same
room. A zero dB difference means the method gave the same result as the mesh average in this room.
The second shows the standard deviation’ (defined as s3 above) in this room, for each simplified
method. The numbers in the legend refer to the averaging methods, where
Circle, Helix, Cylindrical type, Three semicircles, Five fixed positions, Five fixed positions and Max
Corner
3
1
2
4
5
6
Frequency (Hz) / Weighted values
3
1
2
4
5
6
Frequency (Hz) / Weighted values
Figures 9. Living room: For each of the simplified methods: upper; average SPL difference
(compared to the mesh), lower; the standard deviation between the averages (s3).
16
3
1
2
4
5
6
Frequency (Hz) / Weighted values
3
1
2
4
5
6
Frequency (Hz) / Weighted values
Figures 10. Small bed room: For each of the simplified methods: upper; average SPL difference
(compared to the mesh), lower; the standard deviation between the averages (s3).
17
3
1
2
4
5
6
Frequency (Hz) / Weighted values
3
1
2
4
5
6
Frequency (Hz) / Weighted values
Figures 11. Mid sized bed room: For each of the simplified methods: upper; average SPL difference
(compared to the mesh), lower; the standard deviation between the averages (s3).
18
3
1
2
4
5
6
Frequency (Hz) / Weighted values
3
1
2
4
5
6
Frequency (Hz) / Weighted values
Figures 12. Small reverberation chamber: For each of the simplified methods: upper; average SPL
difference (compared to the mesh), lower; the standard deviation between the averages (s3).
19
3
5
Frequency (Hz) / Weighted values
3
5
Frequency (Hz) / Weighted values
Figures 13. Large reverberation chamber: For two of the simplified methods: upper; average SPL
difference (compared to the mesh). Lower; the standard deviations of the average(s3). The dashed line
shows the standard deviation of the sound pressure levels for all positions in the mesh, included to
illustrate the spatial variation of levels in this reverberation chamber.
20
New loudspeaker positions, differences in the mesh
average sound pressure level
The change of average sound pressure level has been measured when the loudspeaker was moved to another
position than was used above for the comparisons of scanning methods. The sound pressure levels were
measured at the same mesh of microphone positions in 4 of the 5 rooms.
The figures 14 show that the resulting mesh average sound pressure level may change considerably
when the position of the loudspeaker is changed. The standard deviation of sound pressure levels
within the mesh seem to be similar after the loudspeaker was moved. Whether these differences
influence the sound pressure levels in a receiving room and hence the sound reduction remains to be
studied.
Figures 14. Effects of changing the loudspeaker position. Upper; change of the average sound
pressure levels. Lower; change of the standard deviations (s3)
21
Discussion of results and some practical experiences
Six simplified methods to obtain a spatial average sound pressure level have been compared. The
methods are (1) Circle, (2) Helix, (3) Cylindrical type, (4) Three semicircles, (5) Five fixed positions
and (6) Five fixed positions and Max Corner. The comparisons presented in the rooms described
above indicate:
•
•
•
•
•
•
•
•
•
Average differences of the spatial average typically deviate in the range of +1 to -2 dB from the
mesh average when the proposed methods are used in the frequency range 50-5000 Hz. Larger
differences may occur at a few frequencies
Method (5) approximates the mesh average within +/-1 dB in the frequency range 50-5000 Hz,
in the rooms used for this study
The moving microphone methods (1 - 4), deviate more from the mesh average at some
frequencies, in comparison with method (5)
Method (5) may give slightly higher standard deviations than some of the moving microphone
methods (1 - 4). However, it is not obvious this is only caused by greater random errors of the
sampling method. The fixed positions were more evenly distributed throughout the permitted
space of each room, compared to a moving microphone path that had to be placed more or less
in the center of the small rooms, which may lead to systematic errors. These relations remain to
be investigated further
The standard deviations may actually be higher above 100 Hz than below, the 160 and 200 Hz
third octave bands may even contain the most uncertain results
Method (6) averages are systematically higher compared to the other methods and to the mesh.
This is consistent with the results found by Hopkins and Turner [2], who also compared their
averages to a mesh of positions not closer than 0,5 m to the walls. However, they also found
that their corner-and-room averages agreed better with another type of room average where
microphone positions very close the walls were included in the mesh average.
The meaning of a mesh reference could be discussed in the context of energy content of the
room and sound insulation prediction models based on energy balances (e.g. EN 12354/ISO
15712). But the choice of mesh reference does not affect the relative difference between results
obtained with different methods
Method (6) reduces the standard deviation below 100 Hz in all rooms used in this study. As
pointed out in the draft standard, it should not be used from 100 Hz and higher
It remains to be investigated whether the systematic differences of method (6) affect the sound
reduction index between the source and receiving rooms. It is not obvious whether these
differences would cancel out if the rooms have different acoustic conditions, since the magnitude
of difference in the third octave bands (50-63-80 Hz) appear to vary in the rooms of this study
Some practical problems with moving microphone techniques (manual scanning) emerged during the
measurements:
• It is difficult to fit wide circles of various shapes into the narrow spaces typical for small and
densely furnished rooms. This leads to increased uncertainties since several paths tend to overlap
or have to be modified to avoid coming too close to protruding furniture. If parts of the
microphone path come close to each other, the result approximates one single fixed position taken
in this part of the room and the result obtains an unknown weight in the room average
• The (2) Helix path performs a low standard deviation but also gives some systematic differences
to the mesh. However, the scanning path proposed in the draft standard was difficult to perform
without causing extraneous noise from moving the body and the feet. Thus, the procedure was
changed such that the microphone was swept in two sequences, where 0-180 degrees were
covered first, then the operator paused the acquisition and turned around, then the remaining 180360 degrees were measured
22
•
•
•
•
The three semicircles path was easy to keep clear from furniture etcetera, but difficult to sweep at
a constant speed. It was found easier to perform 4 semicircles, each 15 seconds, kept 45 degrees
apart. However, this method has a slightly higher uncertainty at high frequency. It was also somewhat prone to causing noise from body and clothes, because the arm has to make rather wide
movements
The background noise caused by the operator may be difficult to keep track of during the actual
measurement. It must be emphasized in the standard, that background noise should be minimized
and the same scanning path should be used during background noise and receiving room
measurements
From earlier experiences, this noise may be kept very low when the operator is trained, concentrates on the scanning and work under favorable conditions. Nevertheless, there is a risk that a
moving foot or a loose microphone cable may cause unexpected noise from time to time that
disturbs the measurement at low sound pressure levels. This has been a topic of discussion among
acousticians using manual scanning methods. Such self-generated noise is difficult to separate
from relevant sounds registered by the microphone. On the other hand, being in the receiving
room facilitates keeping ”an ear” on background noise from the site, which is advantageous
In this study, the instantaneous A-weighed sound pressure level versus time (F time weighting)
and the equivalent sound pressure level in third octave bands were plotted on a large screen with
high resolution, which allowed for a detailed monitoring of the signal during each measurement
session. The sound pressure level was much higher than typical background noises. Still, some
records had to be deleted and replaced by new measurements, e.g. when the A-weighted level
varied more than typical. The reasons for such fluctuations were not examined.
The draft standard actually suggests fixed positions without any operator present in the room to serve
as a reference method in case of dispute. The practical findings of this study indicate that method (5)
Five fixed positions has advantages. It may even be preferred for all types of room. It was quicker than
first anticipated; 5 measurements each of 16 seconds did not take much longer to perform than first
finding out a suitable position and then perform a 60 seconds manual scan. The operator may avoid
causing extraneous noise during measurements and can focus on the signal acquisition during the
measurement.
In order to ensure a representative sampling with fixed microphones; it may be considered to specify that
microphone positions shall be spread over 4 different sectors of the room and the heights be varied between
0.5 – 1.9 m above the floor.
The results of this study are relevant for impact sound pressure measurements (part 2 of the standard).
However, it remains to study the uncertainty of the sound pressure level difference between two
adjacent rooms, used for sound reduction index estimates.
23
References
[1] ISO/WD 16283-1 Acoustics - Measurement of sound insulation in buildings and of building
elements - Field measurements. Part 1: Airborne sound insulation. ISO/TC 43/SC 2 document N18.
Issued 2011-04-29 by ISO/TC 43/SC 2/WG18 AHG6 secretariat (Danish Standards). Available as
ISO/DIS 16283-1 in April 2012.
[2] HOPKINS, C. On the efficacy of spatial sampling using manual scanning paths to determine the
spatial average sound pressure level in rooms. Journal of the Acoustical Society of America, 2011,
129(5), pp. 3027-3034.
[3] HOPKINS, C; TURNER, P. Field measurement of airborne sound insulation between rooms with
non-diffuse sound fields at low frequencies. Applied Acoustics 66 (2005) 1339–1382.
[4] OLESEN, H.S. Measurement of the acoustical properties of buildings - additional guidance.
Espoo 1992. Nordtest. NT Technical Report 203. NT Project No. 963-91. This NT study has been
used for the technical report ISO 140 TR/13 and standard ISO 140-4
[5] PEDERSEN, S; MØLLLER, H; PERSSON WAYE, K: Joint Baltic-Nordic Acoustics Meeting
2008, 17-19 August 2008, Reykjavik, Iceland. This work has also been published in the journal article:
“Indoor Measurements of Noise at Low Frequencies – Problems and Solutions”, Journal of Low
Frequency Noise, Vibration and Active Control, Vol. 26, No. 4, 2007.
[6] SIMMONS, C. Measurement of sound pressure levels at low frequencies in rooms. Comparison
of available methods and standards with respect to microphone positions. Acta Acustica united with
Acustica, 1999, 85, pp.88-100.
24
Nya fältmätningsmetoder
från ISO
ISO:s arbetsgrupp för nya fältmätningsmetoder inom byggakustiken
har gett ut ett förslag till standard,
som ska ersätta dagens ISO 140standard för luftljudsisolering. Kommande delar beskriver mätning av
stegljudsnivå respektive fasadisolering. En central del av alla metoder
är hur man ska medelvärdesbilda
ljudtrycksnivån inom ett rum. För
stegljudsnivå är det viktigt att veta
hur man ska placera hammarapparaten, särskilt på anisotropa lätta
bjälklag som har olika ”mottaglighet” i olika delar. För att undersöka
hur de föreslagna metoderna fungerar i praktiken, samt ge värdefull
information till standardiseringskommittén, har en serie prover utförts inom AkuLite-projektet.
Förslaget prISO 16283-1 för luftljudsisolering (ersättare till ISO 140-4) anger flera
metoder för att medelvärdesbilda ljudtrycksnivån i sändarrum och mottagarrum. Antingen låter man mätmikrofonen
röra sig kontinuerligt genom rummet med
en motordriven bom eller genom att svepa
den runt med sin egen arm, eller så mäter
man i fixa positioner och räknar fram ett
energimedelvärde efteråt. Metodens repeterbarhet, det vill säga hur nära ett antal
oberoende försök ligger varandra, beror
till stor del på om fördelningen av mikrofonpositioner är jämn och täcker in hela
rummet. Metoder som bara mäter mitt i
rummet ger en systematisk underskattning av medelljudnivån i rummet, vilket
man vill undvika.
För att prova de sex metoder som
föreslagits gjorde vi upprepade mätningar i fem rum för att få med inverkan av
rumsstorlek och mängden diffuserande
och absorberande material. För att få en
referens i respektive rum gjordes i tillägg
detaljerade ljudtrycksnivåmätningar i ett
Artikelförfattare är
tekn dr Christian
Simmons,
Simmons akustik
och utveckling AB,
Göteborg.
28
Figur 1: Metoder 1 till 4 med manuella svep av mätmikrofonen för att bestämma
rumsmedelvärdet av ljudtrycksnivå, enligt förslag till ISO-standard 16283-1.
Metod 3 är den svenska metoden i SS 25267 bilaga H (tidigare SIS TR-8).
rutnät av mikrofonpositioner med 0,7
meters mellanrum. Två höjder mättes,
0,7 och 1,7 m över golvet. Energimedelvärdet av mätningarna i detta rutnät användes som normalisering, vilket gör att
man kan studera både de systematiska
skillnaderna mellan metoderna i standarden och de slumpmässiga variationerna.
Högtalaren hade samma plats i alla delprov, normalt på golvet i ena hörnet, riktad mot väggen.
Metoderna 1 till 4 med svepning av
mikrofonen visas i figur 1. Metod 5 är att
helt enkelt placera ut mikrofonen i fem
olika positioner i rummet och ta energimedelvärdet. Denna metod har använts
länge, med kända för- respektive nackdelar.
Metod 6 är tänkt enbart för små rum
(rumsvolym högst 25 m³) och bara i tredjedelsoktavbanden 50-63-80 Hz. Principen är att kombinera resultat från metod 5
med fyra extra mätpositioner nära (0,3 till
0,5 m) hörnen. Hörnmätningarna jämförs
först med varandra, och den högsta ljudtrycksnivån i något av hörnen används sedan vid medelvärdesbildningen, enligt
följande formel 1:
LLF = 10log[13–100,1Lhörn + 23–100,1Lrum]
(1)
Det här var inte så enkelt att få till
praktiskt. När hörnmätningarna upprepades fem gånger i ett rum visade det sig
att standardavvikelsen blev cirka 0,4 till
0,7 dB. Om metoden ska användas bör
man precisera mätpositionen tydligare.
Figur 2.
Resultat
Diagrammen 1 och 2 visar hur energimedelvärdena för respektive provmetod avviker från det noggranna medelvärdet (av
mätningar i ett tätt rutnät), respektive
standardavvikelsen från upprepade försök
inom ett vardagsrum (25 m²).
Några observationer och tankegångar
från dessa försök är:
● Skillnader om ± 1 till 2 dB mot rutnätsmätningarna är normala i frekvensområdet 50 till 5000 Hz, men större skillnader
kan förekomma vid frekvenser under 200
Hz. Hörnmetoden ger systematiskt förhöjda nivåer, se kommentar nedan
● Metod 5 med fem fasta mikrofonpositioner fördelade över rummet ligger inom
±1 dB från rutnätet i alla de rum som studerats
● Metod 5 har något högre standardavvikelse än de manuella svepmetoderna 1 till
4. Då ska man beakta, att de fasta positioBygg & teknik 3/12
tid i samma del av en rumsmod och rumsmedelvärdet påverkas för mycket av denna mätpunkt
● De manuella svepmetoderna är känsliga
för att man alstrar störande ljud. Man har
också sämre kontroll på bakgrundsljud
när man ska hantera instrumentet samtidigt som man sveper med mikrofonen
● Slutsatsen blev, att för denna tekniska
metod (ISO 16283) bör man använda fasta mikrofonpositioner. Manuella svep kan
ligga i en översiktsmetod typ ISO 10052.
Stegljudsmätningar på träbjälklag
Ytterligare en fråga som är aktuell i standardiseringsdiskussionerna och även
inom AkuLite, är hur man ska placera
hammarapparaten på ett skivgolv med bärande bjälkar av trä (eller stål). För att
undersöka känsligheten för olika placeringar genomfördes en serie tester i SP:s
stegljudslaboratorium. Ett provgolv byggdes upp enligt ISO 10140, med en normal
golvspånskiva (22 mm) som skruvades
och limmades till 220 x 45 mm träbjälkar
cc 600 mm, se bild 1. Ett undertak monterades med två lag 12,5 mm gipsskivor på
akustikprofiler. I mellanrummet placerades lätt mineralull.
Hammarapparaten placerades i det mest
extrema fallet antingen helt ovanpå bjälkarna, så att alla hammare träffar på bjälken. I nästa mätning placerades den mitt
mellan bjälkarna och parallellt med dessa.
Diagram 1.
Diagram 2.
nerna var mer utspridda än de manuella
svepen, som tenderar att hamna i rummens mitt. Det kan medföra att upprepningarna blev mer lika varandra och gav
en skenbar ”stabilitet” i metoderna
● Standardavvikelsen kan vara högre
över 100 Hz än vid lägre frekvenser
● Hörnmetoden ger systematiskt högre
nivåer än rutnätsmätningarna. Detta resultat har andra forskare också kommit
fram till, frågan är hur man ska hantera
skillnaderna mot övriga metoder. Fördelen med hörnmetoden är att den ger avsevärt lägre standardavvikelse, det vill säga
repeterbarheten blir bättre
● Metoderna ”Helix” och ”Cylinder” ger
bra repeterbarhet men också en del märkliga avvikelser från rutnätsmätningarna.
Det visade sig vara svårt att genomföra
den här typen av mätningar i små rum, det
blir lätt att man rundar utstickande möbler
etcetera. Då får man en längre uppehållsBygg & teknik 3/12
Bild 1.
Diagram 3.
29
Figur 3.
För varje hammarapparatsposition mättes
ljudtrycksnivån i mottagarrummet under
bjälklaget som ett medelvärde av ett varv
med 1,1 m radie och 32 sekunders integrationstid. Delar av resultaten framgår av figur 3 ovan och diagram 3 på sidan 29 (där
30
kurva J = joist avser placering på bjälke
och W = web av ser placering mellan bjälkarna).
Det är tydligen så, att man får avsevärt
högre stegljudsnivå på detta bjälklag då
hammarapparaten placeras på bjälkarna
jämfört med då den placeras mitt emellan
dessa. Kompletterande försök med hammarapparaten vriden 45 grader relativt bjälkarna gav något mindre skillnader men
trenden var densamma. Därmed blir valet
av positioner för hammarapparaten högst
väsentlig för vilket medelvärde som erhålls. Mätningarna visar att det finns möjlighet att förenkla mätmetoden, eller att
minska standardavvikelsen i fältmätningar
jämfört med befintlig standard genom noggrant val av antal hammarpositioner och
placering av hammarapparaten relativt bjälkar. I laboratoriemetoden för mätning av
stegljudsförbättring hos golvbeläggningar
på lätta bjälklag specificeras placeringen av
hammarapparaten i detalj. Så kunde vi göra
även för fältmetiden. Man kan också fundera över vad som är mest relevant för den
subjektivt upplevda störningen, medelvärdet eller en förhöjd nivå som inträffar så
snart någon går, flyttar möbler, dammsuger
eller leker ovanpå en bjälke.
En fråga som man också måste fundera
över i sammanhanget, är hur resultat med
den nya metoden kan jämföras med den
äldre metoden, det vill säga om det går att
skapa en ”bakåtkompatibilitet”. Hur mätmetoden ska se ut är alltså långt ifrån givet, här finns utrymme för diskussioner!
Om någon har egna erfarenheter av den
här typen av variationer eller förslag på
hur metoden borde se ut, så ta gärna kontakt med författaren.
■
Bygg & teknik 3/12
Appendix
Diagram on Impact sound Pressure level
Complementary to figure 3 in the Swedish article
Uncertainties of room average
sound pressure levels measured
in the field according to the
draft standard ISO 16283-1
Experiences from a few case studies
Results according to a new draft ISO standard for
measuring sound pressure levels are presented.
Stockholm • Borås • Skellefteå • Växjö
Tel: +46 10 516 50 00 • www.sp.se
SP Rapport 2012:28
These results are also being submitted no Noise
Control Engineering Jounal.