DAGA 2016 Aachen

Comparison of standing wave ratio method and transfer function method for
measuring sound absorbing properties of 3d-printed samples
Foteini Setaki1, Martin Tenpierik2, Arjan van Timmeren3, Michela Turrin4

Faculty of Architecture and the Built Environment, TU Delft

Julianalaan 134, 2628 BL Delft, The Netherlands

1) E-mail: F.Setaki@tudelft.nl
2) E-mail: M.J.Tenpierik@tudelft.nl
3) E-mail: A.vanTimmeren@tudelft.nl
4) E-mail: M.Turrin@tudelft.nl
In the case of the standing wave ratio method, a movable
microphone is used to define the ratio of the maximum and
minimum pressure. From this ratio derives the reflection
Introduction coefficient and from there it is possible to get the impedance
Measurements in the impedance tube are very useful as they and absorption coefficient.
enable the measurement of sound absorption for normal The standing wave ratio method is very reliable and does not
incidence sound, under controlled conditions. Due to the need microphone calibration. However, it only measures
small size of the samples (a few centimetres in diameter), one frequency at a time and therefore the procedure becomes
they are frequently used in the development of new rather slow. It is a labour-intensive method and needs a lot of
materials. Alternative solutions usually demand the time to measure a large number of frequencies; as a result
construction of large samples, which is more difficult and typically only 1/3rd octave band centre frequencies are
expensive. Moreover, the impedance tube method can be measured.
carried out with relatively simple equipment in a normal
room and does not need specialised test chambers [1]. A lot In the case of the transfer function method, the sound source
has been written about the accuracy of the measurement is emitting a broadband stationary noise signal. The transfer
method, the effect of the mounting condition, etc. [2-4] function, which is defined as the ratio of pressures, between
However, a majority of these studies are concerned with two microphone positions is measured in the tube. This leads
open porous and homogenous materials. to the reflection coefficient of the sample for the same
frequency band as the broadband signal. The impedance and
This paper examines two 3d-printed samples with non- absorption coefficient can be derived as well.
homogenous and anisotropic geometry by measuring them
with two different methods in the impedance tube: the The second approach is accurate, reliable and obtains the
standing wave ratio method and the transfer function absorption coefficient and impedance of the surface for a
method. The tested samples contain open tube structures and large range of frequencies (within limits) considerably fast.
are produced with Selective Laser Sintering from polyamide
(PA12). Each sample (10cm in diameter) is measured in the Methodology
impedance tube, following the approach of each technique. This paper reports the results for two 3d-printed samples,
In the first case the measurements are taken at 1/3rd octave which contain interference absorbers, tested with the method
band centre frequencies and in the second from 50 to of the standing wave ratio and of the two microphone
1600Hz with a 2Hz interval. The measured results are transfer function methods. The measurements of the acoustic
juxtaposed to each other and to calculations based on absorption properties of each test sample were carried out in
acoustic theory. Besides indicating which method is more accordance with ISO 10534-1:1996 and ISO 10534-2:1998,
suitable for the specific type of non-porous absorbers, the correspondingly. The samples were measured 3-5 times,
advantages and disadvantages of each technique are each with different mounting conditions, minimising the
discussed. possibility of experimental error. For both methods,
cylindrical samples with diameter of 100 mm were used. For
Standing wave ratio and transfer function the standing wave ratio method (B&K, type 4002)
measurements were taken at the third-octave band centre
method frequencies: 100/ 125/ 160/ 200/ 250/ 315/ 400/ 630/ 800/
A loudspeaker is generating planar waves within the 1000/ 1250/ 1600 Hz. The transfer function method (B&K
impedance tube. The resulting standing wave relies on the type 4206) acquires results from 50 to 1600Hz, which were
acoustic properties of the inserted sample. By measuring it, measured with a resolution of 2Hz .
the normal incidence absorption coefficient can be calculated.
This is such a common technique in acoustics that it is The samples are fabricated with Selective Laser Sintering
described also by international standards[1, 5, 6], as follows: (SLS) from polyamide (PA12) (fig.1). SLS is a powder-
based additive manufacturing technique that works by

1042
 DAGA 2016 Aachen

selectively scanning and sintering the surface of the raw- calculations; as in the case of the standing wave ratio the
material with a laser [7]. SLS was chosen, among other AM percentage deviation is 5-12%. However, the α values are
techniques, because of its powder-based nature and its significantly higher in the case of the transfer matrix method,
economical and fast features in production. The self- when compared to the standing wave ratio. This is likely due
supporting capabilities of the technique allow the fabrication to the fact that the measurements with the standing wave
of highly complex internal air paths. Moreover, the post- ratio method are taken only at the third-octave band centre
processing only involves the removal of the unsintered frequencies. Hence, the maximum peaks might not have
powder using simple techniques such as using compressed been detected since they locate between the measured
air or brushes. Even though the sintered material presents frequencies. It is notable that the dashed curve represents the
interpolation of the α values between the measured
some micro-porosity, it is not expected to perform as an
frequencies, while the crosses show the measured values.
open porous material.
The tested samples contain tubes with two open ends. Their
resonant frequencies can be calculated based on [8]:

(2n  1)  c
f  , n = 1, 2, 3, … [Hz] (1)
2  ( Ltot  L)

As in Helmholtz resonators, the air mass co-vibrating at the

pipe inlets leads, according to Fuchs [8], to an inlet
correction to the length L dependent on the pipe radius r. In
equation 1, ΔL is calculated as 0.8r (= the radius of the
inlet). Figure 2: measured results for sample GP.3.4. in the impedance
tube, using the standing wave ratio(- + - + -) and the transfer
The designs of the two samples can be shown in Figures 2 matrix method (solid line)
and 3. Sample GP.3.4 contains one air-path with a total
length of 0.343 m and circular profile of 0.02m diameter. Larger differentiations are observed in the measurements of
The first design frequency is 476Hz (n=1). Sample GP.4.2 sample GP.4.2 (fig. 3). The transfer function method detects
contains three air-paths with diverse lengths: 0.343m (L1), 8 peak frequencies that are in satisfying agreement with
0.515m (L2) and 0.686m (L3). All of them have a diameter theory. The deviation percentage is 0-30% (see table 1). It is
of 0.015 m. The corresponding design frequencies for n=1 notable that the highest percentile deviation corresponds to
are: 482, 325 and 245 Hz. the longest air-path (L=0.686m). This fact might indicate
that the specific air-path might still contain un-sintered
powder, which is affecting the results, or that there is an
interference with neighbouring channels. Further
investigation needs to be conducted to clarify the reason.
The standing wave ratio method identifies only two major
peaks at 500 and 630Hz. This might relate to the fact that the
examined samples are performing sharp and narrow peaks
and the measurements are taken only at the third-octave
band centre frequencies. As a result, certain peaks are not
detected since they locate between the measured frequencies.
For this reason, it would be recommended to take
Figure 1: Samples GP.3.4 (right) and GP.4.2 (left) are measurements at smaller intervals, especially around the
tested in the impedance tube, using the transfer function expected peak frequencies.
method and the standing wave ratio method

Results
Figures 2 and 3 plot the measured normal sound incidence
absorption coefficient (α) as a function of frequency (dashed
and solid curves) together with calculated data (vertical
lines). The results show that within certain narrow frequency
bands high absorption peaks are found; the corresponding α
values are measured from 0.47 to 1 at peak frequencies
between 288-1600Hz.
The measurements of sample GP.3.4 are in satisfying
agreement with regard to the theoretical peak frequencies
and show only small deviations from these calculated peaks. Figure 3: measured results for sample GP.4.2. in the
As shown in table 1, in the case of the transfer function impedance tube, using the standing wave ratio (- + - + -)
method, the measured peaks deviate maximum 3% from and the transfer matrix method (solid line)

1043
DAGA 2016 Aachen

Literature
Table 1: calculated and measured peak frequencies and
corresponding percentage deviations [1] T.J. Cox, P. D’Antonio (2009), Acoustic Absorbers and
L n Calcula Meas. Percentage Meas. Percentage Diffusers: Theory, Design and Application. Taylor and
[m] ted Peaks - deviation Peak - deviation
Peak F Transfer [%] standing [%] Francis, Abingdon.
[Hz] matrix wave ratio

[2] K. V. Horoshenkov et al., “Reproducibility experiments

3.4 0,343 1 476 490 3 500 5 on measuring acoustical properties of rigid-frame
0,343 2 1429 1478 3 1600 12 porous media (round-robin tests)”, J. Acoust. Soc. Am.,
4.2 0,343 1 482 480 0 500 4 122(1), 345–53 (2007).
0,343 2 1446 1416 2 - - [3] A. Cummings, “Impedance tube measurements on
0,515 1 325 322 1 - - porous media: The effects of air-gaps around the
0,515 2 975 956 2 - - sample”, Journal of Sound and Vibration, 151 (1), 63-
0,515 3 1625 1600 2 1600 2 75 (1991)
0,686 1 245 288 17 - -
[4] N. Kino, T. Ueno, “Investigation of sample size effects
0,686 2 736 586 20 630 14 in impedance tube measurements”, Applied Acoustics,
0,686 3 1227 862 30 - - 68, (11–12), 1485-1493, (2007)

[5] Acoustics -- Determination of sound absorption

coefficient and impedance in impedance tubes -- Part 1:
Discussion Method using standing wave ratio, ISO 10534-1:1996
This paper examined two 3d-printed samples with two [6] Acoustics -- Determination of sound absorption
different measuring methods in the impedance tube: the coefficient and impedance in impedance tubes -- Part 2:
standing wave ratio and the transfer matrix method. The Transfer-function method, ISO 10534-2:1998
examined sound absorbers are non-homogenous and
anisotropic and perform sharp and narrow peaks.
[7] I. Gibson, D. Rosen, B. Stucker (2015), Additive
Even though the standing wave method is reliable, it Manufacturing Technologies: 3D Printing, Rapid
measures only one frequency at a time. In this experiment, Prototyping, and Direct Digital Manufacturing,
measurements were taken at the third-octave band centre Springer
frequencies. This frequency set should normally give a
sufficient impression for porous absorbers, which are [8] Fuchs, V.F. (2013), Applied Acoustics: Concepts,
expected to plot a more or less continuous curve as the Absorbers, and Silencers for Acoustical Comfort and
absorption spectrum. However, for the examined samples, Noise Control, Springer, Berlin
sound absorption depends on resonances. In this case, peak
frequencies and the corresponding α-values might not be
detected and give wrong impression on the performance of
the sample. For this reason, it would be recommended to
take measurements at smaller intervals, especially around the
expected peak frequencies.
The primary advantage of using the transfer matrix method
is that it obtains the absorption coefficient for a large range
of frequencies (within limits) with only a couple of quick
measurements. It seems that in the case of resonant
absorbers, which usually perform sharp and narrow peaks, it
is more beneficial to take measurements with the transfer
function method. However the standing wave ratio is a
robust method and is a useful second check on results but
requires much time in order to have high resolution results.

Acknowledgements
The project is supported by Technology Foundation STW.
More partners are involved: Materialise, Peutz, Merford and
ARUP. Special thanks to Evert de Ruiter, Diemer de Vries
and Lau Nijs.

1044