Int. J. Metrol. Qual. Eng.

4, 97–107 (2013)
c EDP Sciences 2013
DOI: 10.1051/ijmqe/2013045

Acoustic metrology – an overview of calibration methods

and their uncertainties
E. Frederiksen
ef-consult – retired 2010 from Brüel & Kjaer, Denmark

Received: 22 April 2013 / Accepted: 12 July 2013

Abstract. The paper gives a brief description of the principles and the uncertainty of the acoustic cali-
bration methods that today are applied by National Metrology Institutes and calibration service centers.
Even if some of the calibration principles have been applied over more than half a century, the methods
and the instrumentation are still being refined in order to minimize their uncertainty, to extend their fre-
quency ranges, to include extra parameters and to speed up slow processes. In addition to the traditional
methods for microphone sensitivity and frequency response calibration, new development areas, like for
example wind power, has created needs for low-frequency and infra-sound calibration, down to 0.1 Hz.
Other high-tech areas have lead to the development of methods for phase response comparison calibration
of microphones for large arrays, for sound intensity measurement and for verification of dynamic linearity
of microphones at very high sound pressure levels, up to about 174 dB that corresponds to 10 kPa.

Keywords: Microphone calibration; pressure reciprocity; free-field reciprocity; coupler comparison

calibration; electrostatic actuator; phase response calibration; high-pressure calibration; infrasound
calibration; calibration uncertainty

1 Introduction or 0 dB, and goes up to more than 20 kPa or 180 dB. These
wide ranges and the different types of sound field, which
During the later decades measurement of sound has be- occur, cause a need for many different models of micro-
come increasingly important. More and more people are phone and several different calibration and test methods.
being disturbed by noise at home, during transportation This paper describes principles and uncertainties of the
and at their workplaces. People are getting stressed by basic methods that are standardized by the International
noise and some may even, in severe cases, lose their hear- Electrotechnical Commission (IEC), Technical Commit-
ing due to the noise. Many manufactures have become tee No. 29 [1]. Also some supplementary, less commonly
aware of the drawbacks related to the noise of their prod- applied, methods are briefly described.
ucts. If too high, the noise may even harm their own busi-
nesses – they may over time lose their ability to compete.
Noise, or rather low noise, has become an important com-
petition parameter for many types of products, ranging
3 Measurement and reference microphones
from consumer products like vacuum cleaners and refrig-
erators to advanced industrial machinery and to means Essentially all commonly used measurement [2] and refer-
of transportation like ships, high speed trains and large ence standard [3] microphones are condenser microphones.
passenger aircrafts. For these and other reasons, acoustic Condenser microphones are selected, because of their es-
measurement and acoustic metrology have become very sentially flat frequency responses and their high mechani-
important activities for the modern society. cal stability. Also their simple shape and design is impor-
tant, as this makes it possible not only to measure, but
also to calculate many of their properties, including their
interaction with the principal types of sound field.
2 Measurement and calibration ranges IEC/TC29 has worked out a standard (IEC61094-4)
Sound measurements in air are performed over a wide for the commonly used types of measurement micro-
range of frequency that goes from infrasound to ultra- phone, which are called working standard microphones.
sound or, say, from a tenth of a Hertz to about 200 kHz. Depending on their diameters, 23.77 mm, 12.7 mm and
Sound is also measured over a wide dynamic range that be- 6.35 mm they are designated WS1, WS2 and WS3 (see
gins below the threshold of human hearing, below 20 μPa Fig. 1). Another standard (IEC61094-1) specifies Labora-
tory Standard Microphones, named LS1 (23.77 mm) and

Correspondence: erlingfred@gmail.com LS2 (12.7 mm). They are designed to fit into calibration

Article published by EDP Sciences

98 International Journal of Metrology and Quality Engineering

12
dB
8
Free-field Response

0
Diffuse-field
Response
-4 Pressure-field
Response
Diffuse-field Microphone
-8

-12
10 100 1000 10000 Hz 100000
Fig. 1. Working standard microphones. IEC61094-4 types
WS1 (23.77 mm), WS2 (12.7 mm) and WS3 (6.35 mm) are Fig. 3. Free-field, diffuse-field and pressure responses of a
shown with and without diaphragm protection grids. microphone (WS2) optimized for use in diffuse sound fields.

depends on the type of sound field. In order not to influ-

ence the pressure that occurs in a cavity, the applied mi-
crophone should ideally have a very stiff diaphragm (high
acoustic impedance). Correspondingly for a free-field, in
order not to influence the pressure in the selected mea-
surement point, the diameter of the inserted microphone
should be small, ideally less than say 5–7% of the wave-
length. In practice, this is rarely achievable; therefore, it
is necessary to account for the occurring change of pres-
Fig. 2. Laboratory standard microphones. LS2 Ø 12.7 mm sure. This should be done both while measuring with the
(left) and LS1 Ø 23.77 mm (right) may be calibrated by both microphone and while calibrating the microphone. A cor-
the pressure and the free-field reciprocity methods and serve responding, but smaller, pressure change occurs, when a
as National Reference Standards. microphone is placed in a diffuse sound field.
It is very important to notice that the pressure changes
occurring in free and diffuse sound fields, in practice,
couplers and are the types of microphone that by na- depend only on the dimensions of the microphone body.
tional metrology institutes (NMIs) are used as national It is, therefore, essentially the same for all microphones
standards, Reference Microphones (see Fig. 2). The two units of same model and, therefore, if measured once,
standards specify mechanical dimensions, sensitivity, fre- the obtained values are valid for and may be used as
quency response, acoustic impedance, dynamic range, correction values for any other microphone of the given
ambient influence and stability. model. Figure 3 shows frequency responses of a micro-
A condenser microphone is a reciprocal transducer [4]. phone (Ø 12.7 mm) that is designed to have a flat response
It can, of course, work as a microphone, by converting an in a diffuse sound field. There, is a fixed ratio between its
acoustic signal to an electrical signal, but it can also work free-field, diffuse-field and pressure responses (fixed differ-
as a sound source – it can convert an electrical input to ence in dB). Most microphone manufacturers supply cor-
an acoustical output. Such reciprocal microphones may be rection data, so, if one of the responses is measured and
calibrated by the reciprocity calibration principle, which thus known, the two other ones can be calculated. Es-
is described in the following. sentially all calibrations of working standard microphones
make use of this fact.

4 Sound fields and dedicated microphones

5 Principle of reciprocity calibration
Sound is measured at many different places and sound
fields can be very different, but there are three basic and Microphone calibration by the reciprocity technique was
principal types. Within cavities, whose dimensions are invented in the 1940s [7–9]1 . Since then the method has
smaller than say a quarter of the sound’s wavelength, become refined and standardised and is now the dominat-
the field is called a pressure-field. Typically such a field ing primary calibration technique for determining both
occurs in couplers for calibration of microphones, tele- pressure-field [10] and free-field [11] responses. The meth-
phones and hearing aids. The sound field inside an-echoic ods that are quite complex and time consuming to work
rooms or out of doors, where sound may propagate freely with are mainly applied by national metrology institutes
without disturbance from reflecting objects, is called a and by leading microphone manufacturers.
free-field [5], while the type of field in rooms with hard re- Reciprocity calibration is based on the measurement
flecting walls and/or between many sound sources is called of the transfer function between two coupled microphones
a diffuse-field [6]. that are operated as a source and a sensor respectively.
Any measurement probe influences the quantity it
1
measures – also the condenser microphone. The influence Reference [8] is abbreviated form of reference [7].
 E. Frederiksen: Acoustic metrology – an overview of calibration methods and their uncertainties 99

uout The pressure-field responses are thus obtained by cou-

pling the microphones with the air (or gas) contained in
a small closed cavity, while the free-field responses are
obtained with an open-air coupling in a room with no dis-
turbing reflecting surfaces (an anechoic room) (see Figs. 4
mic(B) and 5).

air-filled coupler 6 Primary pressure reciprocity calibration

mic(A) Today several national metrology institutes around the

world perform pressure reciprocity calibration of labora-
tory standard microphones [12]. Typically the frequency
range is from 20 Hz to 10 kHz for LS1 and 20 Hz to 20 kHz
iin for LS2 microphones, but some institutes have experience
with calibration to lower and to higher frequencies. Gen-
erally, standing waves inside the applied acoustic couplers
Fig. 4. Microphones coupled by air-filled cavity for pressure
determine the upper frequency limit of the calibration.
response calibration.
Such wave problems may, in practice, be reduced either
by filling the coupler cavity with hydrogen, which has a
higher speed of sound than air, or by shaping the cav-
ity in a way that leads to a simpler wave pattern and
thereby facilitates a better determination of the acoustic
mic(B)
transfer impedance. The latter is the case for cylindrical
mic(A)
iin uout
cavities having same diameter as diaphragms of the cali-
brated microphones. Typically the volume of such couplers
is 4–8 cm3 for LS1 and 0.3–0.7 cm3 for LS2 microphones.
The calculation of transfer impedance for this type of cou-
pler, called a plane-wave coupler, is based on the trans-
200 mm
typical with
mission line theory. IEC 61094-2 describes the method and
12.7 mm microphones dedicated software is available for the calculations.
Fig. 5. Microphones coupled by air in the open space for Today hydrogen filled couplers are rarely used. The
free-field reciprocity calibration. filling with the hydrogen is very time consuming to work
with. The pressure and concentration of the hydrogen that
must be precisely known are difficult to determine and to
The microphones are coupled together in an acoustically keep constant during the measurements. Therefore, the
well-defined way, while the over-all transfer function, the much more practical plane-wave couplers have taken over,
ratio between sensor output voltage and source input even if they due to their smaller size, have the disadvan-
current, is measured (see Figs. 4 and 5). From this ra- tage that they require individual corrections for the load-
tio, called the electrical transfer impedance (Ze ), and ing made up by the connected microphones. Table 1 shows
from the acoustic coupling represented by the acoustic the typical calibration uncertainty that today is achieved
transfer impedance (Za ), the product of the microphone by an experienced NMI.
sensitivities can be determined:
 
Ze
M1 M2 = , Table 1. Typical uncertainty of pressure reciprocity calibra-
Za A tion for experienced NMI.
where M1 and M2 are the sensitivities of microphones 1
and 2, Ze /Za is the ratio of electrical and acoustical Uncertainty 20 32 63 4 10 20 25
dB (k = 2) Hz Hz Hz kHz kHz kHz kHz
transfer impedance.
LS1 0.06 0.04 0.03 0.03 0.08 − −
By using three microphones (1, 2, 3) and by mak-
LS2 0.08 0.05 0.04 0.04 0.04 0.08 0.12
ing impedance ratio determinations (A, B, C) for the
three possible combinations of microphone (1-2, 1-3, 2-3),
the sensitivities of all microphones can be calculated by
solving the equations below: Pressure Reciprocity calibration requires no specific
      acoustic laboratory facilities, but it is practical to work
Ze Ze Ze
M1 M2 = , M1 M3 = , M2 M3 = . with a small measurement chamber, which can isolate
Za A Za B Za C
the microphone and coupler set-up from the noise of
Free-field or pressure sensitivity can be obtained by proper the ambient (see Fig. 6). The chamber improves the sig-
selection of the microphone coupling method. nal to noise ratio, which may shorten the measurement
100 International Journal of Metrology and Quality Engineering

Another disturbing factor of free-field reciprocity cal-

ibration that may highly influence the calibration uncer-
tainty is room reflections. For many years this problem
could only be reduced by applying large an-echoic rooms of
high quality, which is of course a very costly solution. But
even in such rooms the reflections can be too strong for
obtaining the uncertainty that today is required. There-
fore, some NMIs have worked with newer signal processing
methods that can separate the part of the signal that goes
directly from the source to the receiver microphone from
the delayed parts of the signal that are due to reflected
sound arriving via the walls or other reflecting objects of
the room.
Among others, former staff members of the Danish
Technical University (Lyngby) have developed such a
method [15]. The method effectively suppresses the reflec-
tions and implies even that much smaller and acoustically
“bad” rooms are applicable for obtaining some of the very
Fig. 6. Measurement chamber designed for reduction of noise best results. Small rooms or chambers of few cubic meters
and for pressurisation. are in fact also advantageous in other respects. They may
be placed within a normal laboratory room (see Fig. 7).
This eases the work and gives shorter cables that help to
time and improve the measurement repeatability. This reduce possible problems related to cross-talk. The com-
may especially be necessary, when making calibrations pany Brüel & Kjaer has developed a turn-key system [16]
in the range between 2 Hz and 20 Hz. Other advan- applying the DTU method and has also built such systems
tages are gained by connecting a pressurisation system for customers. The calibration uncertainty that can be
to the chamber. This makes it possible to calibrate at obtained with this system is shown in Table 2.
various static pressures and, thus, to determine micro-
phone pressure coefficients [13] and to calibrate directly
at the standard pressure, 101,325 kPa. For laboratories Table 2. Typical uncertainty of free-field reciprocity
that are located far above sea level this is especially calibration for experienced NMI.
important when participating in inter-laboratory and
key-comparison calibrations. Uncertainty 1 2 4 8 10 20 25
dB (k = 2) kHz kHz kHz kHz kHz kHz kHz
LS1 0.10 0.08 0.07 0.07 0 08 0.10 −
7 Primary free-field reciprocity calibration LS2 − 0.10 0.08 0.07 0.07 0.09 0.12

The standard IEC 61094-3, books [14] and articles de-

scribe the free-field reciprocity method. But, quite few
NMIs are able to perform free-field reciprocity calibra- 8 Free-field corrections for LS microphones
tion. The theory behind the determination of the acous-
tic transfer impedance is simpler to deal with than that Even if new methods and measurement systems are now
of the pressure reciprocity calibration, but the practical available, the facilities, the equipment and the experience
measurement of the electrical transfer impedance is tech- that are needed for performing free-field reciprocity cal-
nically more difficult. This is due to the very weak sound ibration may for some laboratories be too costly to es-
pressure that in the open space is produced by the source tablish and maintain. Therefore, IEC has issued a stan-
microphone and due to the very small receiver output dard IEC61094-7 with values for the difference between
voltages that, therefore, have to be measured. This may free-field and pressure responses of Laboratory Standard
lead to serious measurement problems that are related to Microphones, LS1 and LS2. It is thus possible to obtain
ambient or inherent noise and to electrical cross-talk in individually valid free-field frequency responses of these
the measurement system. As these problems especially oc- microphones by making pressure reciprocity calibrations
cur at lower frequencies, the lower limit of free-field reci- and by adding the corrections stated in the standard [17].
procity calibrations is typically between 0.8 kHz (LS1) and The free-field responses are valid for perpendicular sound
2 kHz (LS2). Free-field calibration is made up to about incidence (0o ) on the diaphragms. The uncertainty of
25 kHz and 50 kHz and for LS1 and LS2 microphones, responses obtained this way is higher than that, which
respectively. It should be added that the mentioned, rel- may be achieved by making direct free-field reciprocity
atively high, lower limits give no serious problems, as calibration (see Tab. 3).
there are only small differences between the pressure and The above method is valid, because the difference be-
the free-field responses below the limits, and because the tween the free-field and the pressure responses depends
missing lower part of the responses can, therefore, be dominantly on the dimensions of the microphones, which
determined by pressure reciprocity calibration. are the same within the model, and only to a small degree
 E. Frederiksen: Acoustic metrology – an overview of calibration methods and their uncertainties 101

12
dB B&K Type 4191 0o Incidence
8 Sound Field Corrections
10 deg. steps
4

-4
Diffuse-field
-8
Valid without protection grid
-12

-16
1000 1000 0 Hz 100000

Fig. 8. Free-field corrections of WS2 microphone measured

in steps of 10◦ sound incidence and the diffuse-field correction
calculated in accordance with IEC61183.

Reference
Microphone

Fig. 7. Small an-echoic chamber designed for free-field reci-

procity calibration by the Danish Primary Laboratory of
Acoustics (DPLA). Total volume: 6 m3 , inner volume: 1.7 m3
density of glass wool: 90 kg/m3 (not critical). The microphones
being calibrated are mounted on displaceable rods to allow se-
lection of the distance between the microphones. Prof. Knud Unknown
Rasmussen in front of the room. Microphone

Table 3. Typical uncertainty of a free-field response obtained

by adding corrections to a pressure reciprocity calibration Fig. 9. Diffuse-field calibration by comparing a large unknown
made by an experienced NMI. The numbers include correction microphone with a reference microphone that is several times
uncertainties stated by IEC61094-7. smaller. By moving the microphones in the room the possible
effects of dominant standing waves at specific frequencies is
Uncertainty 1 2 4 8 10 20 minimized.
dB (k = 2) kHz kHz kHz kHz kHz kHz
LS1 0.09 0.12 0.14 0.18 0.28 0.20
times smaller pressure-field calibrated reference micro-
phone. If the reference microphone is sufficiently small, the
on the acoustic impedance of the diaphragm that varies a difference between its pressure and diffuse-field responses
little from unit to unit. may be so small that it can be neglected. The measure-
ment may be made with a practically obtainable, yet not
fully perfect, diffuse sound field by interchanging the mi-
9 Determination of diffuse-field responses crophones between fixed positions or better, by mounting
them on moving arms that slowly pass them through the
Diffuse-field responses may in principle be measured by same circular orbit within the sound field of the reverber-
the reciprocity calibration method like pressure and free- ant room [20] (see Fig. 9).
field responses. Some researchers have made reciprocity
calibration experiments during the time [18], but the
method needs further analysis to be practically applied for 10 Coupler comparison calibration
metrology. One difficulty is to design a room, within which
diffuse sound fields of required quality can be generated. The sensitivity and frequency response calibration of a
Therefore, other methods are applied. measurement microphone may either be performed by
The diffuse-field response or correction may be calcu- making comparison measurements with a reference micro-
lated of a number of free-field responses that are mea- phone at a series of frequencies, or by using a combined
sured for different angles of sound incidence and that to- technique that implies a comparison at one frequency
gether make a good representation of all possible angles. only and a relative frequency response measurement.
The standard IEC61183 [19] describes the calculation and This response can be measured by excitation with an
gives details about choice of the measurement angles and Electrostatic Actuator (see description below). In the lat-
the associated weighting factors for the calculation (see ter case, the sensitivity is generally measured at a fre-
Fig. 8). quency that is so low that the pressure-, free- and diffuse-
The diffuse-field response of a larger microphone might field sensitivities of the microphone are essentially equal.
also be measured by comparing its response with a several The frequency 250 Hz is most commonly used, as this
102 International Journal of Metrology and Quality Engineering

Microphone
Interfaces

0.5
8.0 1.8

9.3
Source Source
16
Cavity Diaphragm
18

Fig. 10. Typical comparison coupler with built-in sound

source for calibration of WS2 microphones. It is shown beside
the cross-sectional view of its air-filled cavity.

Fig. 11. Electrostatic actuator used for frequency response

is also the operation frequency of some common type of calibration of WS1 microphones. Actuators are standardised
acoustic calibrator (pistonphone). by IEC and are widely used for calibration of working standard
Comparison calibration of pressure sensitivity is cov- microphones. They can excite the microphone diaphragms with
ered by IEC61094-5 [21]. The standard describes different constant electrostatic pressure over very wide frequency ranges.
ways of mounting the two microphones. In all cases, they
are lined up along their axes of symmetry with their di-
aphragms or protection grids facing each other in a dis-
tance of typically 2–3 mm. One might either use couplers
with built-in sound sources or microphone holding jigs
with separate, nearby mounted sound sources. A typical
coupler and its cross-sectional view are shown in Figure 10.
Such mountings are useable over wide frequency ranges,
as the pressure exciting the two microphones is essentially
the same. The comparison uncertainty for f < 1000 Hz is
typically <0.03 dB (k = 2), while it is about 0.20 dB for
good couplers and equally large microphones up to 16 kHz.
The resulting uncertainty of the method is obtained by
combining these uncertainties with those of the reference
microphone calibration. Typically this becomes <0.12 dB
(k = 2) for WS1 and WS2 microphones if f < 1000 Hz. Fig. 12. Principle of electrostatic actuator voltage supply.
Free-field and diffuse-field frequency responses may be
obtained by adding corrections to measured responses,
but it should be noticed that the sound field corrections close to (0.5 mm) and in parallel with the microphone
that should be applied depend on the type of the coupler diaphragm. It generally stands on its small glass legs on
or jig – and also that the corrections differ from corre- top of the supported part of the microphone diaphragm.
sponding free- and diffuse-field corrections valid for other Typically 800 V DC and 30 V AC are applied between
measurement methods. the actuator and the diaphragm (see Fig. 12). These volt-
ages will with a typical actuator generate an equivalent
sound pressure of about 1 Pa (94 dB). The pressure may
be calculated by the equation below:
11 Electrostatic actuator calibration
εa 2
p (t) = − (U0 + u sin (ωt)) (1)
The reciprocity calibration methods described above are 2d2
absolutely necessary for primary calibration of reference
where
standard microphones, but these methods are quite slow,
and generally too slow for calibration of working standard p(t) equivalent instantaneous pressure (Pa),
microphones that are calibrated in large numbers. In order ε dielectric constant of the surrounding air (8.85 ×
to measure their frequency responses, it would be natural 10−12 F/m),
to look for a sound source that within narrow tolerances a ratio between the areas of the perforated actuator
could supply a constant sound pressure over a wide fre- and the diaphragm,
quency range. Unfortunately, such sources do not exist, d distance (m) between actuator and microphone
but the electrostatic actuator is a very good substitute. It diaphragm,
can generate a constant electrostatic pressure (a uniformly U0 DC voltage (V) applied between actuator and
distributed force) on the flat metallic (electrically conduct- diaphragm,
ing) diaphragms that are used with all WS microphones. u peak value of sinusoidal AC voltage (V) applied
The actuator is a plane and very stiff metallic plate between actuator and diaphragm,
with holes or with slits that make it acoustically trans- ω angular frequency (rad/s),
parent (see Fig. 11). During the operation, it is placed t time.
 E. Frederiksen: Acoustic metrology – an overview of calibration methods and their uncertainties 103

Table 4. Typical uncertainties of actuator response (A) and

of free-field response (B) obtained by adding corrections. The
reference frequency is 250 Hz.
Uncertainty 0.02 0.1–5 10 20 40
dB (k = 2) kHz kHz kHz kHz kHz
WS2–A 0.12 0.10 0.15 0.20 0.30
WS2–B 0.14 0.14 0.20 0.35 0.60

In principle, the actuator method is a primary method,

but the method is not used for absolute sensitivity
calibration, as the uncertainty of some parameters (a, d)
and, therefore, of the resulting pressure is generally too
large.
The pressure that the actuator produces on the mi- Fig. 13. Different sound exposure of microphones and related
crophone diaphragm is frequency independent, if the sup- responses. Diaphragm only (left) gives upper response (A).
plied AC and DC voltages are kept constant during the Diaphragm and vent (right) gives lower response (B).
calibration. One should, therefore, expect to measure the
pressure response of the microphone, but the measured
response deviates from this. The deviation is due to a that are calibrated by specific methods accounting for the
small additional pressure that is generated, because the influence of the microphone vents that equalize changes of
diaphragm moves, when this is excited by the actuator. static pressure.
The magnitude of this pressure is smaller for less sensi- It is important to be aware of that the microphone
tive microphones, having stiffer diaphragms, than for more low-frequency response depends very much on, if the vent
sensitive microphones. However, there are essentially fixed – during measurements and calibration – is exposed to the
ratios (differences in dB) between the actuator response sound pressure or not (see Fig. 13).
and each of the three principal responses, the pressure, Different methods have been and are still used, but no
the free-field and the diffuse-field responses. Therefore, if method has yet been standardized. One simple method
these ratios are known, all three responses may simply that is, however, not very precise, is to calculate the cut-off
be determined by applying corrections to the easily mea- frequency (the –3 dB point) from the decaying response to
sured actuator response. The corrections, which depend on a step-change of the ambient pressure. A more advanced
the microphone model, are available from the leading mi- method is to make the calibration with a device that is
crophone manufacturers. The actuator method is widely called a laser pistonphone. Such devises are designed and
used by calibration service centres. Typical uncertainties described by several NMIs. Other calibration devices that
of the actuator calibration itself and of a free-field response are used have an electro-dynamic sound source of low
are shown in Table 4 for a Working Standard Microphone acoustic impedance that generates the calibration pres-
(WS2). The method is generally used between 20 Hz and sure inside a very small cavity. By the low source and the
100 kHz, but it can be used at both higher and lower fre- high load impedance an essentially flat frequency response
quencies. But by low-frequency calibration, say f < 20 Hz, may be obtained.
attention should be paid to the facts that any measure- Furthermore, some five years ago Brüel & Kjaer pre-
ment microphone has a static pressure equalization vent sented a new infrasound calibration method [24] and have
that influences its low-frequency response and that the now developed a system that is based on this. It measures
effect of this cannot, in all cases, be measured with an frequency responses from 250 Hz down to 0.1 Hz with an
actuator. Therefore, when low-frequency calibration is re- uncertainty less than 0.1 dB. The system calibrates its
quired, it may be necessary to apply specific calibration own reference microphones by a method that is called the
methods. Related Microphones Method. It can also calibrate work-
The actuator method is relatively cheap to estab- ing standard microphone and expose them to the sound in
lish, as no specific acoustic laboratory facilities are re- both of the above mentioned ways (see Figs. 13 and 14).
quired. The method is described in detail in the standard
IEC61094-6 [22] and in reference [23].
13 High- and low-level testing
12 Infrasound calibration As previously mentioned sound measurements are made
in a range from say 0 dB (20 μPa) to 180 dB (20 000 Pa).
Presently relatively few NMIs deal with infrasound cal- There is thus a ratio of 1:1.000.000.000 between the weak-
ibration, but low-frequency measurements and calibra- est and the strongest measured sound. However, most cal-
tion are fields of growing interest. Windmill parks, power ibration and test methods work within a narrow range –
stations, aircrafts and road traffic are some sources that say between 0.2 Pa and 30 Pa. This is, of course, fine if
may create disturbing infrasound noise. This will have to the microphones are linear over their dynamic application
be measured with dedicated low-frequency microphones range. With respect to this, it is common to rely on the
104 International Journal of Metrology and Quality Engineering

Fig. 16. The B&K high-pressure system for testing of dy-

namic linearity and distortion of measurement microphones
Fig. 14. Infrasound calibration Unit with built-in sound and sound level meters. The system is optimized for low dis-
source. The unit and its associated measurement system work tortion of the generated sound and for low vibration excitation
from 250 Hz down to 0.1 Hz and can calibrate its own reference of the microphones.
microphones as well as working standard microphones. They
may be exposed only on the diaphragm or on both diaphragm
and static pressure equalization vent. There are different methods that can be used for high-
pressure calibration or testing. Traditionally such tests
have been performed with a pistonphone. This is a motor-
driven mechanical device that consists of a closed cav-
ity with a piston that can compress and decompress the
enclosed air. The generated pressure is a function of the
static pressure, the cavity volume and of the pistons vol-
ume displacement. The sound pressure is also influenced
by the so-called heat-conduction effect that occurs be-
tween the air and the walls of the cavity. The described
pistonphone is generally a simple and a rather cheap de-
vice. It usually suffers from lack of flexibility with respect
to selection of sound pressure and from its distortion that
may often be higher than that of the microphones that
are to be tested.
A dedicated and elaborated system for measurement
of dynamic linearity [25], but also a more complex sys-
Fig. 15. Sealed and vibration isolated measurement chamber tem, has been developed by Brüel & Kjaer. The sound is
for determining the inherent noise of condenser microphones produced by a loudspeaker. Via some acoustically cou-
and their preamplifiers. The volume of the chamber is about pled tubes the sound is lead to a position, where the
2000 cm3 . The shown chamber has been applied for testing diaphragms of the microphone being tested and of a
low-noise microphones with an inherent A-weighted noise level high-pressure reference microphone are placed. On its way
below 0 dB. through the tube system the pressure of the sound is dras-
tically magnified and “cleaned” for distortion components.
The system can, by the selected type of reference micro-
specs stated by the microphone manufacturer, but there
phone and by the optimized sound system, measure dy-
are cases, where testing of dynamic range is prescribed. An
namic RMS-linearity within 0.3% and distortion within
example is the standard for pattern evaluation of sound
0.5% all the way from 1 Pa up to 10 kPa RMS or 14 kPa
level meters that requires testing of both the inherent
peak (0.14 bar). The linearity of the system is ensured by
noise and the distortion at the upper end of the dynamic
two high-pressure reference microphones of different oper-
range. High pressure microphone testing may also be nec-
ation principles that over the dynamic range must trace
essary in connection with very costly, high-tech projects,
each other within narrow tolerances (see Fig. 16).
like launching of rockets and testing of jet engines and
air-crafts.
Low-level or inherent noise tests are performed by en- 14 Phase frequency response calibration
closing the microphone and its preamplifier in a sealed
cavity with very stiff sound-isolating walls. Depending on Measurement of sound intensity and mapping of sound
the levels to be measured, it may be necessary to make fields have created a need for making absolute and rel-
a set-up with soft springs that can insulate the chamber ative phase response measurements. This is necessary
from vibrations created by applied instruments or maybe for applying the intensity measurement principle that
occurring in the test facility (see Fig. 15). is standardized and described in IEC61043 [26]. Sound
 E. Frederiksen: Acoustic metrology – an overview of calibration methods and their uncertainties 105

Fig. 17. Wide band coupler for phase comparison calibration

of ø 12.7 mm microphones. used for matching of microphones
to sound intensity measurements. the coupler fully encloses the Fig. 18. Main unit of the B&K SLM calibration system
microphones in its cavity and works from 20 Hz to 16 kHz. that depending on the type of SLM works fully automated,
semi-automated or manually.

Table 5. Uncertainty of phase comparison calibration of

an intensity microphone pair obtained with the B&K phase
The pattern evaluation standard prescribes a very
calibration system and its dedicated couplers.
thorough checking of a number of general points, envi-
Uncertainty of 20–500 1 2 5 10 ronmental tests, electrostatic and radio frequency tests,
phase comparison Hz kHz kHz kHz kHz electrical performance tests and, of course, some acousti-
calibrations - deg 0.01 0.02 0.04 0.10 0.20 cal tests of:

– Directional responses.
pressure and air-particle velocity must be measured and – Frequency response (ref. direction).
multiplied to obtain the intensity, but because of the lack – “Free-field correction values”.
of suited particle velocity transducers both quantities are – “Adjustment value(s)”.
determined by two pressure-sensing microphones that, in – Self-generated noise.
order to ensure valid results, must have essentially iden-
tical phase frequency responses [27]. The requirements The long series of tests requires a large collection of
are so strong that it is not possible to produce the mi- specific instruments and several costly test facilities, in-
crophones with the required uniformity – matching pairs cluding a good and large an-echoic room. These tests are,
must be selected from large microphone batches. The se- therefore, only performed in relatively few laboratories.
lection is made with specifically designed couplers that In lack of local testing and results, it is very common to
expose the two microphones to the same magnitude and accept test results and approvals of sound level meters
phase of the pressure. Figure 17 shows such couplers that originating from other countries.
are designed for use with the phase response comparison
The verification of that a specific SLM meets the re-
system developed by B&K.
quirements of IEC61672-1 requires that its model has
Sound mapping is made with small and large micro-
passed the tests of IEC61672-2 and that the instrument
phone arrays. Some systems may consist of hundredths
itself has passed the periodical tests of IEC61672-3. Pe-
of microphones. The microphones of such arrays must
riodical tests are performed in many countries. Often
also fulfill certain requirements to the uniformity of their
these tests are made by calibration service centers, but
phase frequency responses, but fortunately a significantly
in smaller countries or countries that are just starting up
larger spread can be accepted. Maintenance and exten-
acoustic activities the tests are often performed by the
sion of such arrays requires reference microphones with
NMI. It was a goal for the IEC working group to de-
absolutely calibrated phase frequency responses. Modern
fine effective and relatively simple periodical tests that
reciprocity calibration systems that determine the com-
are not too costly to perform and that will, therefore,
plex microphone sensitivity meet the required tolerances
become more widely used. The testing of IEC61672-3 is,
and can cover this need (see Tab. 5).
thus, limited to a general inspection, to seven electrical
performance tests and to the following three acoustic tests:
15 Calibration of sound level meters – Indication at reference frequency.
– Inherent noise.
Specification and calibration of sound level meters (SLM)
– Frequency response. (measured by using actuator,
are standardized by IEC and covered by the following
multi-frequency calibrator or sound field).
three standards [28–30]:
– IEC61672-1 Specifications. SLM calibration systems that – without specific acoustic
– IEC61672-2 Pattern evaluation tests. facilities – can perform these ten tests are available from
– IEC61672-3 Periodical tests. different manufacturers (see Fig. 18).
106 International Journal of Metrology and Quality Engineering

uncertainty is based on reciprocity calibration of Labora-

tory Standard Microphones, LS1 or LS2.

17 Field calibration by sound level calibrators

Single-frequency sound level calibrators are widely used in
the field to ensure proper operation of sound measurement
systems, but often they are not used the optimal way. Con-
necting the calibrator to the microphone, starting this and
adjusting the system to display the level specified for the
Ø calibrator implies a risk for not discovering or for hiding
36 mm 72x72x40 mm a calibrator or a system error. Much higher confidence to
(a) (b) proper functioning of the system may be obtained by set-
ting this up in accordance with the stated sensitivity of
the microphone and then check with the calibrator that
the displayed level is within acceptable limits. With mod-
ern and well-maintained equipment the displayed value
should not differ by more than few tenths of a decibel from
the expected reading. After having observed the reading
one might either leave the system setting or re-adjust it,
depending on the degree of confidence one have in the
two calibration methods. If the deviation is too high, one
should find the reason and not take the risk to hide a
265x125x62 mm possible defect by correcting the microphone sensitivity
(c) parameter of the system.
Fig. 19. (a) Operation principle of pistonphone. (b) Typi-
cal feed-back calibrator 94 dB, 1000 Hz. (c) Multi-frequency 18 Summary
calibrator 94 dB–114 dB, 31.5 Hz–16 kHz.
Several different calibration methods and systems have
been briefly described. Some of them are very commonly
16 Calibration of sound level calibrators used by NMIs that operate acoustical departments, while
others are only available in case of specific needs in
Sound level calibrators are sound sources that produce the country. Pressure reciprocity calibration is the basic
known sound pressure levels and are, thus, well-suited for method that delivers reference standards for sensitivity
calibration or verification of sound level meters and other calibration of all types of measurement microphone and
sound measurement systems. They are now considered to for level calibration of sound calibrators. The method may
be necessary accessory parts to all measurement systems even be used for obtaining free-field responses of reference
and are used for system checking before and after per- standards LS1 and LS2, as the necessary free-field correc-
forming all important measurements. Nowadays there are tions are specified by IEC61094-7. In some cases it may,
basically two types. One type is a mechanical device with a thus, be avoided to establish a free-field calibration system
cavity and with, usually two, small pistons that are driven that is generally more costly and space requiring.
by a miniature electric motor (see Fig. 19a). They operate In lack of local calibration service centers some NMIs
at relatively low frequencies, typically 250 Hz, and are gen- find it necessary to establish secondary calibration of
erally very stable sound sources, which make them well- working standard microphones. The standardized, eas-
suited for reference standards for comparison calibration ily operated and widely used methods for this are cou-
of other types of calibrator. They are called pistonphones. pler comparison calibration for sensitivity at 250 Hz and
Today the most common calibrator type is the “feed- electrostatic actuator for frequency response calibration
back calibrator”. It contains some, not especially critical, combined with adding of corrections supplied by the
sound source, but a very stable microphone that measures microphone manufacturers or others.
the generated pressure and that via a feed-back circuit ad- SLM calibration is important in many countries. Many
justs this to the proper level. This principle has lead to NMIs do not perform pattern evaluations themselves.
some relatively small and highly stable calibrators whose They rely on those performed and reported by other in-
sound pressure is essentially independent of the ambient ternationally leading acoustic laboratories and establish
pressure. They typically work at the SLM reference fre- periodic testing only. This is far less costly and it does
quency 1000 Hz (see Fig. 19b). The feed-back principle is not require specific acoustic measurement facilities.
also used for multi-frequency calibrators that may cover More specific types of calibration and testing like infra-
nearly the full frequency range of common sound level me- sound calibration, phase comparison calibration and dy-
ters (see Fig. 19c). The standard IEC60942 [31] specifies namic linearity testing are today performed by relatively
sound level calibrators of different classes and describes few NMIs. If needs should occur for such applications, they
their way of calibration. The calibrator calibration and its may often be combined with common calibration systems
 E. Frederiksen: Acoustic metrology – an overview of calibration methods and their uncertainties 107

that nowadays are usually based on modern multi-function 16. E. Frederiksen, Turnkey Free-field Reciprocity System for
analyzers. Primary Microphone Calibration, Brüel & Kjaer Technical
Review, No. 1 (2009)
References 17. IEC 61094-7, Measurement microphones, Part 7: Values
for the difference between free-field and pressure sensitivity
1. IEC 61094-1/2/3/4/5/6/7, Measurement microphones, levels of laboratory standard microphones
series of microphone and calibration standards 18. H.G. Diestel, Reciprocity Calibration of Microphones in
2. IEC 61094-4, Measurement microphones, Part 4: a Diffuse Sound Field, J. Acoust. Soc. Am. 33, 514–518
Specifications for working standard microphones (1961)
3. IEC 61094-1, Measurement microphones, Part 1: 19. IEC 61183, Electroacoustics – Random-incidence and
Specifications for laboratory standard microphones diffuse-field calibration of sound level meters
4. IEC 50(801), Vocabulary: Acoustics and electroacoustics, 20. K. Brinkmann, H. Goydke, in Random-incidence and
25-8 – Reciprocal Transducer Diffuse-field Calibration, edited by G.S.K. Wong, T.F.W.
5. IEC 50(801), Vocabulary: Acoustics and electroacoustics, Embleton (AIP Handbook of Condenser Microphones,
23-28 – Free Sound Field American Institute of Physics, New York, 1994), pp. 120–
6. IEC 50(801), Vocabulary: Acoustics and electroacoustics, 135
23-31 – Diffuse Sound Field 21. IEC 61094-5, Measurement microphones, Part 5: Methods
7. R.K. Cook, Absolute pressure calibration of microphones, for pressure calibration of working standard microphones
J. Res. Natl. Bur. Stand. 25, 489–505 (1940) by comparison
8. R.K. Cook, Absolute pressure calibration of microphones, 22. IEC 61094-6, Measurement microphones, Part 6: Elec-
J. Acoust. Soc. Am. 12, 415–420 (1941) trostatic actuators for determination of frequency response
9. L.L. Beranek, Acoustic Measurements (John Wiley & 23. E. Frederiksen, in Electrostatic Actuator, edited by G.S.K.
Sons, Inc., New York, London, 1949), p. 113 Wong, T.F.W. Embleton (AIP Handbook of Condenser
10. IEC 61094-2, Measurement microphones, Part 2: Primary Microphones, American Institute of Physics, New York,
method for pressure calibration of laboratory standard 1994), p. 231
microphones by the reciprocity technique 24. E. Frederiksen, Infrasound Calibration of Measurement
11. IEC 61094-3, Measurement microphones, Part 3: Primary Microphones, Brüel & Kjaer Technical Review, No. 1
method for free-field calibration of laboratory standard (2008)
microphones by the reciprocity technique 25. E. Frederiksen, Verification of High-pressure Linearity and
12. G.S.K. Wong, in Primary Pressure Calibration by Distortion of Measurement Microphones, in Proceedings of
Reciprocity, edited by G.S.K. Wong, T.F.W. Embleton the International Conference on Acoustics, Kyoto, 2004
(AIP Handbook of Condenser Microphones, American 26. IEC 61043, Electroacoustics: Instruments for the measure-
Institute of Physics, New York, 1994), Chap. 4 ment of sound intensity Measurements with pairs of pres-
13. K. Rasmussen, The static pressure and temperature coeffi- sure sensing microphones
cients of laboratory standard microphones, Metrologia 36, 27. E. Frederiksen, Pressure Microphones for Intensity
265–273 (1999) Measurements with Significantly Improved Properties,
14. V. Nedzelnitsky, in Primary Method for Calibrating Free- Brüel & Kjaer Technical Review, No. 4 (1986)
field Response, edited by G.S.K. Wong, T.F.W. Embleton 28. IEC 61672-1, Electroacoustics - Sound Level Meters,
(AIP Handbook of Condenser Microphones, American Part 1: Specifications
Institute of Physics, New York, 1994), Chap. 4 29. IEC 61672-2, Electroacoustics - Sound Level Meters,
15. S. Barrera Figueroa, New Methods for Transducer Part 2: Pattern Evaluation Test
Calibration: ree-field reciprocity calibration of condenser 30. IEC 61672-3, Electroacoustics - Sound Level Meters,
microphones, Ph.D. thesis, Oersted-DTU, Acoustic Tech- Part 3: Periodic Test
nology, Technical University of Denmark, 2003 31. IEC 60942, Electroacoustics - Sound Calibrators