19

th
INTERNATIONAL CONGRESS ON ACOUSTICS
MADRID, 2-7 SEPTEMBER 2007
CRITERIA FOR LOW FREQUENCY NOISE
PACS: 43.28.Dm
Vercammen, Martijn
Peutz bv, P.O. Box 66, NL-6585 ZH MOOK, Netherlands; m.vercammen@mook.peutz.nl
SUMMARY
Some specific cases of low frequency noise annoyance have been the driving force for an
investigation commissioned by the Dutch ministry of environment and housing into the
occurrence of low frequency noise and possibilities to impose criteria. The result was a proposal
for limiting levels. But then it stopped. There were no legal enforcements. After a number of
fruitless attempts to apply these criteria in environmental permits, now, after 18 years, it
suddenly happened. In a single case the criteria were allowed. Since this can be important for
next cases, the once formulated criteria are now relevant again.
In this paper the reasons behind the criteria will be given, an overview will be given of other
criteria and some remarks will be made on the limitations of the method.
INTRODUCTION
Some 18 years ago we have done some research on limiting levels for low frequency noise
(LFN). The results were published in [1] and [2]. Since then there is not much done with the
proposals, at least not in the Netherlands. The client, the Dutch Ministry of Housing and
Environment did not want to impose additional limiting levels for LFN. There has been some
attempts to incorporate the LFN criteria in environmental permits with no success, until recently.
A recent verdict suddenly makes these criteria relevant again. For this reason the background
of the criteria is explained here and some remarks are made.
LEGAL ASPECTS
In the Netherlands there is legislation to prevent high noise levels in residential areas. The
limiting levels are formulated in dB(A). There is no specific regulation for low frequency noise.
Industrial companies need an environmental permit. Regulations for noise are part of this permit
and are based on (national) legislation on noise. However local government is to some extent
free to set the conditions for this permit. In the past several attempts were made to incorporate
low frequency noise limits into the conditions of the environmental permit. The court of appeal
however decided these conditions had to be skipped, because of the lack of hard evidence that
annoyance by low frequency noise could objectively be determined.
Recently however, in a new case, the court of appeal has made a judgment that is not in line
with these earlier judgments. In this case there has been an investigation into the occurrence of
low frequency noise, using the criteria described below, and the judgment states that:
- there is sufficient, reproducible, information on the dose-response relationship of low
frequency noise available
- low frequency noise can be characterized as a cause for annoyance, and it can be
determined in an objective way
- there are several international investigations into the criteria for low frequency noise and
most of them are based on the hearing threshold
- the method used corresponds to systems adopted in other countries
- the choice for a 3-10% threshold can be justified and it substantially reduces the risk for
severe noise annoyance
So in this case the judgment was made that the investigation into low frequency noise had a
solid base and was accepted. Lower noise criteria proposed by the complainant were rejected.
Although of little importance in this particular case, it might have as a consequence that
application of the method will be also acceptable in other cases.
19
th
INTERNATIONAL CONGRESS ON ACOUSTICS ICA2007MADRID
2
And although its only one verdict now and this verdict is not the same (yet) as enforcing low
frequency noise criteria for all future environmental permits, the route seems to be open now.
WHATS SPECIFIC ABOUT LFN?
There are some specific characteristics about LFN where it may differ from normal noise. They
are mentioned briefly here:
- Low frequencies usually means frequencies form 0 to 125 Hz. The range from 0-20 Hz
is called infrasound, from 20-125 Hz low frequency audio range. Although frequencies
below 20 Hz may be generated and can be heard, almost all cases of LFN annoyance
fell in the low audio range, a few cases at 16 Hz.
- Low frequencies means large wavelengths. At these frequencies mainly large radiating
structures with very high amplitudes can generate LFN at sufficient levels. Absorption is
very low, so travel distances are large. Infrasound is captured in atmosphere.
- There is a high threshold for LFN that makes that normally LFN is hardly perceived, its
either not present (below threshold) or masked by higher frequencies. In nature LFN is
produced by thunder, storm, waterfalls, avalanches. So this may be an explanation why
people react more intense to LFN then to normal noise. It also means that LFN is not
experienced as part of our natural surroundings. It gives a feeling that it should not be
there.
- LFN is always tonal noise, since either at the source or in the transmission it is amplified
by resonances (e.g. room resonances or a window resonating on the volume of the
room)
- If more resonances are involved narrow band peaks with small frequency difference
may occur, resulting in a throbbing noise, extremely annoying.
- Equal loudness and equal annoyance lines show that for low frequencies, especially for
infrasound, there is a steep increase in loudness and annoyance of LFN, once it is
perceived
- There is a significant variation in individual threshold for LFN
- This leads to a situation where some people, or just one, can hear it and is annoyed by
it and other people just dont hear it
- By focussing on the sound, one can better hear it, especially since it is always tonal
noise. This tuning-in effect will further increase the annoyance for the one or few
individuals that hear it.
- The complaints of LFN-sufferers are stress-related symptoms. They might fear for
collapse of their dwelling, consider to move house etc.
- There are also cases of tinnitus with similar sensation and complaints. It is difficult to
distinguish between LFN and tinnitus, just on the bases of complaints. Measurements
are necessary.
PROPOSED LFN CRITERIA
For the criteria proposed two frequency ranges are considered: the infrasound range (0-20 Hz)
and the low frequency audio range (20-125 Hz)
For the infrasound range the consideration is that, since the annoyance is so strong once heard,
it should not be heard at all. That means the threshold is the limiting value. However there is a
large fluctuation between individual thresholds. From a literature review it was established that
the average hearing threshold for infrasound is 96 dB at 10 Hz and has a slope of 12 dB/octave
between 4 and 16 Hz,. This threshold corresponds to 96 dB(G), see [1]. The G-weighting curve
[3] is a line with a slope of 12 dB/octave, from 1 to 20 Hz. The reference value is at 10 Hz.
However the standard variation between individuals found in these laboratory experiments was
5 dB on average. Setting the limit on the average (50%) threshold would mean that 50% of the
people can hear the noise and will be annoyed if the SPL equals this threshold. To lower this
percentage to 3-10% it was proposed to set the limit to average-2s=86 dB(G).
For the low frequency audio range setting the limit to the threshold would of course be sufficient
to prevent possible annoyance, but it would be a disproportionate severe criterion, not realistic
since many sources exist that do not give unacceptable annoyance. A proposal for limiting
levels had to be based on a good connection with the 86 dB(G) and the mid/high frequency
regulations and had to be based on experience with cases of LFN annoyance.
19
th
INTERNATIONAL CONGRESS ON ACOUSTICS ICA2007MADRID
3
Since all cases of LFN have a strong tonal component that is responsible for the annoyance, a
single point with a level and frequency can characterize each case. A number of cases from our
own (consultancy) experience and of the German LIS (Landesanstalt fr Immisionsschutz) in
Essen are shown in figure 1. The cases where people were annoyed are shown in dots, the
situations people did not complain (any more) are shown in crosses.
To make the connection between infrasound (criterion=hearing threshold) and normal audio
range (criterion is well above hearing threshold) it makes sense to use the A-weighting curve,
since it is based on the 40 phon, that has a lower slope than the hearing threshold. In the figure
two lines are shown: the 20 dB(A) and 25 dB(A) line. Most of the cases that people do not
complain any more are around that 20 dB(A). There are some cases of annoyance around 20
dB(A), but hardly no cases under that level. So it seems that a level of 20 dB(A), limited in
frequency range between 20 and 125 Hz, seems a reasonable limit. This level is 5 to 10 dB(A)
lower than the overall dB(A) criterion in the dwelling.
To make it an easy to apply system the dB(A) approach for a limited low frequency range
seemed to be confusing, so a system with limiting levels for third octave bands, based on the 20
dB(A) curve was proposed. In the 125 Hz octave band some adaptations might be needed to
connect to the overall dB(A).
Since it is not practical to have limiting levels inside dwellings, there might be a need to set
limits outdoors. Due to resonances there is a strong frequency dependency for the transmission
outside-inside. To be on the safe side no more than 10 dB should be added [2].
Figure 1. Cases of LFN annoyance. Left: cases from consultancy work of Peutz[1], Right: cases
from LIS (Landesanstalt fr Immissionsschutz des Landes Nordrhein-Westfalen, Essen, see
also [8]). Dots and circles are the tonal peaks of the indoor SPL in cases the inhabitants where
annoyed. The crosses indicate the situation after reduction. The lines are the 86 dB(G) for
f<16Hz and the inverted 20 and 25 dB(A) curves.
19
th
INTERNATIONAL CONGRESS ON ACOUSTICS ICA2007MADRID
4
CRITERIA IN OTHER COUNTRIES
There are 4 European countries where noise criteria are set for LFN (see also [4]):
In all countries the criteria apply to limiting levels indoor for third octave bands.
Sweden
The criteria are from 31 Hz to 200 Hz. For 31 to 50 Hz these correspond to the hearing
threshold, for frequencies 100-200 Hz the levels correspond to the 20 dB(A) curve. In between
there is a low slope (almost horizontal).[5]
Denmark
In Denmark the method from [1] is followed, with a slight difference: 85 dB(A) and 20 dB(A).
During daytime 25 dB(A) is allowed.[6]
Germany
In Germany the criteria are based on the average hearing threshold from 8 to 100 Hz. In
daytime these levels may be exceeded by 5 dB (for 80Hz: 10 and 100 Hz:15 dB), during night
time the levels are 5 dB lower. There are 2 additional conditions: it has to be tonal noise (1/3
octave level 5 dB higher than its neighboring 1/3 octave bands) and it has to be LFN (dB(C)-
dB(A)>20 dB).
Poland
In Poland 10 dB(A) is used as a reference. The tonal noise also has to be 10 dB above
background level.
The criteria are summarized in figure 2.
DISCUSSION
The annoyance of a sound is a complex matter and difficult to assess with just a limiting level
(see also [10]). Criteria and regulations have to be simple however.
For infrasound there is sufficient evidence that the noise is annoying once its perceived. So
applying the hearing threshold, taking into account the individual variation of the threshold,
seems to be a clear approach. This approach is followed in Denmark. In Germany the average
Figure 2. Overview of hearing thresholds and criteria for low frequency noise in several
European countries
10
20
30
40
50
60
70
80
90
100
45
6
,
18
1
0
1
2
,
1
6
2
0
2
5
3
1
,
4
0
5
0
6
3
8
0
1
0
0
1
2
5
1
6
0
2
0
0
2
5
0
frequency [Hz]
S
P
L

i
n
s
i
d
e

[
d
B
]
ISO226
Average Infrasound
hearing Threshold
Vercammen [1]
Poland
Germany
Sweden
Denmark
19
th
INTERNATIONAL CONGRESS ON ACOUSTICS ICA2007MADRID
5
threshold is used in this frequency range, this seems to be inadequate to prevent annoyance. In
Sweden there is no limit under 31 Hz.
The complexity of annoyance is far more relevant in the low frequency audio range.
Fluctuations of the noise level and protrusion above background level are difficult to incorporate.
In Poland the background level is taken into account. Denmark and Germany have the influence
of the background noise implicitly taken into account by differentiating between day and night. In
Germany a check is made for the tonality of the sound.
In Denmark a 5 dB penalty is given for impulsive sounds.
Another aspect is the tuning-in effect that might be important for tonal noise.
So generally these complex factors are not incorporated.
The Swedisch (31-50 Hz) and German standards are based on the average hearing threshold,
which is especially in the 63 Hz octave much lower than the Danisch standards that is based on
the 20 dB(A). This is relevant since many cases of LFN are actually in that 63 Hz octave band.
The Polish standard is for frequencies under 50 Hz lower than the others and lower than the
average hearing threshold.
It is difficult to obtain good scientific evidence what the appropriate limit should be. For now the
pratical approach has been to look for cases with LFN annoyance. It has to be realized however
that there are many cases where these levels are exceeded but there is hardly any annoyance,
e.g. LFN due to passing busses or trucks.
CONCLUSION
In this paper the LFN criteria proposed in the Netherlands are presented. The method
corresponds to criteria in Denmark. Other criteria in Germany and Sweden are higher for
infrasound, which might result in unacceptable annoyance, or more strict for higher frequencies,
mainly in the 63 Hz octave band.
The influence of background noise and time fluctuations should be incorporated in the system.
REFERENCES
1. M.L.S.Vercammen,. Setting Limits for Low Frequency Noise, Journal of Low frequency
noise and vibration, vol. 8, no 4, 1989.
2. M.L.S.Vercammen, Low-Frequency noise limits, Journal of Low frequency noise and
vibration, vol. 11, no 1, 1992.
3. ISO7196, Frequency weighting characteristics for infrasound measurements, 1995.
4. G.Leventhall,P.Pelmear,S.Benton, A Review of Published Research on Low Frequency
Noise and its Effects, report for the Department for Environment, Food and Rural
Affairs, 2003
5. Socialstyrelsen-Sweden , SOSFS 1996:7/E Indoor Noise and High Sound Levels, 1996.
6. J.Jakobsen, Danish guidlines on environmental low frequency noise, infrasound and
vibration. Jnl Low Freq Noise Vibn & Active Control 20, 141-148, 2001.
7. DIN45680, Measurement and evaluation of low-frequency environmental noise, 1997.
8. D.Piorr, K.H.Wietlake, Assesment of low frequency noise in the vicinity of industrial
noise sources. Jnl Low Freq Noise Vibn 9, 116-119,1990
9. M. Mirowska, Evaluation of low frequency noise in dwellings. New Polish
recommendations. Jnl Low Freq Noise Vibn 20, 67-74, 2001.
10. G.Lassche,F.Koopmans, Low frequency noise criteria: foundation, applicability and
consequences, Internoise 2001