See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/269577044

Feasibility of a length-limited parametric

source for active noise control applications
Conference Paper July 2014

CITATIONS

READS

55

3 authors:
Bhan Lam

Woon-Seng Gan

Nanyang Technological University

Nanyang Technological University

5 PUBLICATIONS 4 CITATIONS

241 PUBLICATIONS 1,609 CITATIONS

SEE PROFILE

SEE PROFILE

Chuang Shi
Nanyang Technological University
29 PUBLICATIONS 122 CITATIONS
SEE PROFILE

All in-text references underlined in blue are linked to publications on ResearchGate,

letting you access and read them immediately.

Available from: Chuang Shi

Retrieved on: 26 May 2016

The 21st International Congress on Sound and Vibration

13-17 July, 2014, Beijing/China

FEASIBILITY OF A LENGTH-LIMITED PARAMETRIC

SOURCE FOR ACTIVE NOISE CONTROL APPLICATIONS
Bhan Lam, Wong-Seng Gan, Chuang Shi
Digital Signal Processing Laboratory, School of Electrical & Electronic Engineering, Nanyang Technological University
e-mail: blam002@e.ntu.edu.sg
In a previous work presented by the authors in ICSV20, we proposed a theoretical framework
to generate a length-limited parametric source for active noise control (ANC). The lengthlimited parametric source consists of two parametric loudspeakers on the same axis. Two
separate ultrasonic carrier frequencies are applied to the different parametric loudspeakers,
resulting in different absorption lengths. By controlling the input phases and amplitudes of
the two parametric loudspeakers, their secondary sound waves can suppress each other in the
overlapped area but retain in the difference area. An advantage of the length-limited parametric source in active noise control is its ability to target a specific control region while minimizing the spillover effect of the conventional parametric source. However, the theoretical
framework from the previous study has not been experimentally validated, and this paper
aims to do the latter. A concentrically-nested parametric loudspeaker, which is driven by two
ultrasonic frequencies with user-controllable settings, was built to conduct this experiment.
The experimental setup consists of National Instruments data acquisition modules and Brel
& Kjrs microphones. The sound field along the length of the two absorption distances was
measured to evaluate the suppression achieved in the overlapped region and retention in the
difference region. Two types of signals were tested in the length-limited parametric source
setup: (1) single tone and (2) linear chirp. In a non-adaptive ANC setup, a length-limited parametric beam was employed as the anti-noise source to locally control the noise source produced by a conventional loudspeaker. From the experimental observations, further investigations can be conducted to improve the length-limited effect for narrowband signals.

1.

Length-limited parametric secondary sources in ANC

One limitation of active noise control (ANC) systems, which utilises dynamic speakers as anti-noise sources, is the physical difficulty in collocating the noise source and anti-noise source to
achieve an ideally wide quiet-zone.1 Thus, ANC systems usually produce localised quiet-zones
resulting in an undesired increase in sound pressure level (SPL) at areas outside the quiet-zones,
also known as the spill-over effect. Moreover, even if the aim of the ANC system is to produce a
localised quiet-zone, the spill-over effect cannot be avoided due to the Omni-directional nature
of dynamic speakers. On the contrary, a highly-directive anti-noise source generated by a parametric loudspeaker could potentially reduce the spill-over area.

ICSV21, Beijing, China, 13-17 July 2014

21st International Congress on Sound and Vibration (ICSV21), Beijing, China, 13-17 July 2014

Parametric
Sources

Second Beam

Length-limited Beam
First Beam

Figure 1: Illustration of proposed length-limited parametric source in ANC.2

Parametric speakers generate highly-directive audible sound beams based on the selfdemodulation effect in the nonlinearity of air.3-5 Since Brooks et al., numerous researchers have
tested ANC systems with parametric anti-noise sources.2, 5, 6 Notably, Kidner et al. have reported a
reduction in the spill-over area when a parametric anti-noise source was simulated based on the
numerical solution of the Khokhlov-Zabolotskaya-Kuznetsov (KZK) equation.7, 8 However, the
spill-over effect still exists in the area of the sound beam between the parametric source and the
quiet-zone location. To further reduce the spill-over effect in ANC systems using parametric
anti-noise sources, an ANC structure using a length-limited parametric anti-noise source was proposed.2
The length-limited parametric source, shown in Figure 1, is based on the principle of ANC,
where two beams on the same axis (with one at the same amplitude but out-of-phase with the other)
nullifying each other in the overlapped region.9, 10. Through simulations using the numerical solutions of the KZK equation, it has been shown that a parametric beam with a higher carrier frequency
(shorter absorption distance) was able to reduce the spill-over effect caused by the beam with a
lower carrier frequency.2 Hence, this paper aims to experimentally verify the existence of a lengthlimited parametric source and to deploy the length-limited beam (as the anti-noise source) in a simple non-adaptive ANC system.

2.

Physical implementation of the length-limited source

With reference to simulations conducted in the previous study by the authors, the generation
of the length-limited source is formulated based on the same source equations (Eq.(1) and Eq.(2))
employing the double-side band modulation (DSBAM) technique.2
p1 p0 [1 m1 sin(d t 1 )] sin(1t )

(1)

p2 p0 [1 m2 sin(d t 2 )]sin(2t )

(2)

However, the center frequencies 1 and 2 are set to 25 kHz and 40 kHz respectively, based
on the ultrasonic emitters (25 kHz and 40 kHz) that will be used in the experiment. The initial
sound pressure p0 of the source equations is set to 124 dB on the surface of both type of emitters.
Based on the environmental parameters of the test environment (temperature at 23C and humidity at 50%), the absorption distances are approximately, 13.7 m and 6.5 m for 25 kHz and 40
kHz parametric beams respectively. Since the absorption distance of the 40 kHz beam is shorter, its
parameters will be varied such that the optimum reduction is observed in the region where both
beams overlap. Thus, the phases, 1 is fixed at 1 0 and 2 varies from 2 0 to 330 at intervals of 30; and modulation indices, m1 is fixed at m1 1 and m2 varies from m2 0.5 to 0.9 at intervals of 0.2.
As both beams have to be located on the same axis and generated from the same source location, a concentrically-nested parametric loudspeaker was proposed and built. The parametric speaker is made up of an inner hexagonal array of 36 (40 kHz) emitters and an outer annular ring of 36
(25 kHz) emitters, as shown in Figure 2. The arrays are driven separately by two analog outputs.
ICSV21, Beijing, China, 13-17 July 2014

21st International Congress on Sound and Vibration (ICSV21), Beijing, China, 13-17 July 2014

Double-side Open-ended Duct

25 kHz Emitters
B&K 4961

40 kHz Emitters

36 X 40 kHz
Emitters (Silver)

Acoustic Domain
Analog Domain
Class D
Amp

36 X 25 kHz
Emitters (Yellow)

Class D
Amp

NI 9234
NI BNC
2110

NI PCI-6733

PC
(Labview)

Figure 2: Concentrically-nested parametric loudspeaker (Left). Block diagram of experimental setup.

(Right)

3.

Experimental validation of the length-limited source

The long absorption distances of the 40 kHz and 25 kHz emitters (~7 m & 14 m respectively)
coupled with the short Rayleigh distances (~1.5 m for both and beam loses directivity in the farfield), required a test environment that has sufficient length and is not affected by the reflections
caused by the lost in directivity. Thus, the experiment was confined in a 9.5 m long PVC pipe with
an internal diameter of 0.16 m.
The generation and real-time parameter control of the DSBAM signals and data acquisition of
acoustic measurements was developed in a system design software (NI Labview) and implemented
in a mix of hardware setup, as shown in Figure 2. In the setup, the parametric array is placed at one
end of the 9.5 m pipe and driven by two separately amplified DSBAM signals from the NI PCI6733 output module. Acoustic measurements are made by lowering the B&K 4961 multi-field microphone into the centre of the pipe at predefined distances from the parametric array (0.2 m to 4 m
at intervals of 0.2 m, 4 m to 9.2 m at intervals of 0.4 m). Acoustic data from the microphone is acquired via the NI 9234 analog input module for post-processing. At each distance point, acoustic
data is acquired for every combination of the varying m2 and 2 values (i.e., 3 12 36 data points
per distance point).
In line with the previous study, the experiment was conducted with the same modulating frequency, d 2 1.5 kHz.2 In addition to the 1.5 kHz single tone modulating frequency, the experiment was also conducted for a linear chirp from 0.5 kHz to 3 kHz with duration of 1 s.
3.1 Single Tone
Figure 3 shows the sound pressure level (SPL) plots of the 1.5 kHz single tone, for 2 from 0
to 330 along the entire length of the pipe at the defined locations, and for m2 = 0.5 (Figure 3(a)),
m2 = 0.7 (Figure 3(b)) and m2 = 0.9 (Figure 3(c)). From comparison of results plotted in Figure 3, a
trend of reduction can be observed in the first 4 m when the phase of the 40 kHz source, 2 = 150.
By observation, the most significant reduction is achieved when m2 = 0.9 at 2 = 150.
To determine the SPL reduction due to the shorter 40 kHz beam, a 1.5 kHz tone was generated with only the 25 kHz source. Figure 4 shows the comparison between the SPLs of the 1.5 kHz
tone without the 40 kHz source and the length-limited beams with m2 = 0.5, 0.7 and 0.9 at 2 =
150. The average SPL reduction is determined in the first 6.4 m of the pipe, which corresponds to

ICSV21, Beijing, China, 13-17 July 2014

21st International Congress on Sound and Vibration (ICSV21), Beijing, China, 13-17 July 2014
the absorption distance of 6.5 m for the 40 kHz source, and are: (1) 5.57 dB reduction for m2 = 0.5,
(2) 7.59 dB reduction for m2 = 0.7, and (3) 9.82 dB reduction for m2 = 0.9.
dB Plot for 1.5 kHz Sine at m = 0.5
2

dB/20Pa
75

330

70

300

65

270
60

Phase, (Deg)

240
210

55

180

50

150

45

120

40

90
35

60

30

30
0
0

3
4
5
6
7
Distance From Emitter in Pipe (m)

25

(a)
dB Plot for 1.5 kHz Sine at m = 0.7
2

dB/20Pa
75

330

70

300

65

270
60

Phase, (Deg)

240
210

55

180

50

150

45

120

40

90
35

60

30

30
0
0

3
4
5
6
7
Distance From Emitter in Pipe (m)

25

(b)
dB Plot for 1.5 kHz Sine at m = 0.9
2

dB/20Pa
75

330

70

300

65

Phase, (Deg)

270
60

240
210

55

180

50

150

45

120

40

90
35

60

30

30
0
0

3
4
5
6
7
Distance From Emitter in Pipe (m)

25

(c)
Figure 3: SPL values of the 1.5 kHz single tone along the length of the pipe for (a) m2 = 0.5, (b) m2 =
0.7 and (c) m2 = 0.9, for 2 from 0 to 330.

ICSV21, Beijing, China, 13-17 July 2014

21st International Congress on Sound and Vibration (ICSV21), Beijing, China, 13-17 July 2014

1.5 kHz dB Values of 25 kHz Beam and Length-limited Beams

80

Overlapped Area of 2 Beams

Length-limited Area

70

dB/20Pa

60

50

40
25kHz Beam Only
Length-limited Beam (=150, m=0.5)
Length-limited Beam (=150, m=0.7)
Length-limited Beam (=150, m=0.9)

30

20
0

3
4
5
6
Distance From Emitter in Pipe (cm)

Figure 4: Comparison of 1.5 kHz SPL for m = 0.5, 0.7 and 0.9 of the length-limited beam against the
25 kHz source without the 40 kHz source turned on.

3.2 Linear Chirp

The linear chirp test is signal is of 1 s duration from 0.5 kHz to 3 kHz. Figure 5 shows the averaged SPL plots of the linear chirp for 2 from 0 to 330 against the distance from the parametric
source in the pipe. The plots are generated for m2 = 0.5 (Figure 5(a)), m2 = 0.7 (Figure 5(b)) and
m2 = 0.9 (Figure 5(c)). From Figure 5, reduction can be observed at 2 = 180, but the lengthlimited effect is minimal at m2 = 0.5. To further investigate, additional SPL values were measured
for m2 values of 0.2, 0.3 and 0.4 at 2 = 180 to determine the optimal value to achieve a lengthlimited effect. Figure 6 shows the linear chirp averaged SPL plot of the 25 kHz source only and the
length-limited beams at 2 = 180 for m2 values of 0.2, 0.3, 0.4, 0.5, 0.7 and 0.9, along the length of
the pipe. By observation, the largest reduction in the first 6.4 m is at m2 = 0.4. Quantitatively, the
average SPL reductions in the first 6.4 m are: (1) 7.89 dB reduction at m2 = 0.2, (2) 8.94 dB reduction at m2 = 0.3, (3) 10.24 dB reduction at m2 =0.4, (4) 5.54 dB reduction at m2 = 0.5, (5) 4.48 dB
reduction at m2 = 0.7, and (6) 2.21 dB reduction at m2 = 0.9.
Due to the large reduction observed in the length-limited area, the SPL reductions in the
length-limited area have to be determined to verify the existence of the length-limited beam. The
reductions in the length-limited area are: (1) 7.18 dB reduction at m2 = 0.2, (2) 8.76 dB reduction at
m2 = 0.3, (3) 10.44 dB reduction at m2 = 0.4, (4) 7.04 dB reduction at m2 = 0.5, (5) 7.60 dB reduction at m2 = 0.7, and (6) 5.54 dB reduction at m2 = 0.9. Thus, a length-limited effect is not evident
for a linear chirp modulating frequency.

ICSV21, Beijing, China, 13-17 July 2014

21st International Congress on Sound and Vibration (ICSV21), Beijing, China, 13-17 July 2014
dB Plot for linear chirp at m = 0.5

dB/20Pa
55

330
300
270

50

Phase, (Deg)

240

210
180

45

150
120
90

40

60
30
0
0

3
4
5
6
7
Distance From Emitter in Pipe (m)

35

(a)
dB Plot for linear chirp at m = 0.7

dB/20Pa
55

330
300
50

240
210

Phase, (Deg)

270

180

45

150
120
90

40

60
30
0
0

3
4
5
6
7
Distance From Emitter in Pipe (m)

35

(b)
dB Plot for linear chirp at m = 0.9

dB/20Pa
55

330
300

Phase, (Deg)

270

50

240
210
180

45

150
120
90

40

60
30
0
0

3
4
5
6
7
Distance From Emitter in Pipe (m)

35

(c)
Figure 5: SPL values for the linear chirp signal along the length of the pipe for (a) m2 = 0.5, (b) m2 =
0.7 and (c) m2 = 0.9, for 2 from 0 to 330.

ICSV21, Beijing, China, 13-17 July 2014

21st International Congress on Sound and Vibration (ICSV21), Beijing, China, 13-17 July 2014
Comparision between 25 kHz Beam and Length-limited beams
75
25 kHz Beam Only
Length-limited Beam
Length-limited Beam
Length-limited Beam
Length-limited Beam
Length-limited Beam
Length-limited Beam

70
65

dB/20Pa

60
55

(=180, m=0.2)
(=180, m=0.3)
(=180, m=0.4)
(=180, m=0.5)
(=180, m=0.7)
(=180, m=0.9)

50
45
40
35

Overlapped Area of 2 Beams

30
25
0

Length-limited Area

3
4
5
6
Distance From Emitter in Pipe (m)

Figure 6: Comparison of linear chirp dB values along the length of the pipe between the 25 kHz source
length-limited beams.
ANC of 1.5kHz Tone at 8 m
70
ANC Off
ANC On

65

Noise Source at 8 m

dB/20Pa

60
55
50
45
40
35
0

Local Control at 8m
1

3
4
5
6
7
Distance From Emitter in Pipe(m)

10

Figure 7: SPL values of local ANC at 8 m with a length-limited parametric source as the anti-noise
source.

4.

ANC with a length-limited parametric anti-noise source

Since the existence of a length-limited parametric source has been validated experimentally
for the single tone case, the feasibility of deploying a length-limited parametric source for ANC
applications can be tested. Using the same experimental setup, a noise source (1.5 kHz single tone)
generated by a conventional speaker (GENELEC 1029A) is introduced at the 8 m position in the
pipe. The distance of 8 m was chosen as it is greater than the absorption length of a typical 40 kHz
parametric array (~6.5 m) and within the length-limited area (~6.5 m to 14 m). A length-limited
beam generated with 2 = 150 and m2 = 0.9 will be used as the anti-noise source to generate a
single tone (1.5 kHz) for local control of the conventional noise source at 8 m. In contrast to traditional ANC, the phase and amplitude of the noise source at 8 m is adjusted such that there is optimum cancellation by the length-limited anti-noise source.
ICSV21, Beijing, China, 13-17 July 2014

21st International Congress on Sound and Vibration (ICSV21), Beijing, China, 13-17 July 2014
SPL values along the length of the pipe for local control at 8 m are plotted in Figure 7. The
conventional noise source at 8 m was reduced by 18.40 dB from 58.56 dB to 40.16 dB. The discrepancy in the waveform for the first 7.6 m of the length-limited beam is due to perturbations
caused by the downstream reflections of the conventional noise source at 8 m.

5.

Conclusion

This paper investigated the feasibility of deploying a length-limited source in a local ANC
setup. Firstly, an experimental validation of the length-limited source proposed in a previous study
was conducted for three types of test frequencies.2 The experiment was confined in a 9.5 m long
PVC pipe and SPL measurements were taken at predefined intervals along the pipe. A reduction of
9.82 dB for single tone was achieved in the length of the pipe corresponding to the absorption distance of the 40 kHz beam. However, there is no conclusive evidence of the length-limited beam
formation for a linear chirp. Next, the local ANC of a noise source generated by a conventional
speaker at the 8 m position of the pipe was demonstrated. An 18.40 dB reduction of a 1.5 kHz single tone noise at 8 m was achieved with the length-limited parametric anti-noise source. Based on
the results, local ANC was achieved with reduced spill-over effect for a single tone. Although it
has been preliminarily shown that a length-limited source can be applied in ANC applications for
single tones, it is worth noting the perturbations caused by the beating phenomenon between the
carriers of the two parametric sources. Further investigation is required to prevent the introduction
of unwanted difference frequencies due to the beating effect.

REFERENCES
1

10

S. M. Kuo and D. R. Morgan, Active noise control systems: algorithms and DSP
implementations: Wiley, 1996.
C. Shi and W. S. Gan, "Using length-limited parametric source in active noise control
applications," presented at the 20th International Congress on Sound and Vibration
(ICSV20), Bangkok, Thailand, 2013.
M. Yoneyama, J. i. Fujimoto, Y. Kawamo, and S. Sasabe, "The audio spotlight: An
application of nonlinear interaction of sound waves to a new type of loudspeaker design,"
The Journal of the Acoustical Society of America, vol. 73, p. 1532, 1983.
C. Shi and W.-S. Gan, "Development of parametric loudspeaker," Potentials, IEEE, vol. 29,
pp. 20-24, 2010.
W.-S. Gan, J. Yang, and T. Kamakura, "A review of parametric acoustic array in air,"
Applied Acoustics, vol. 73, pp. 1211-1219, 2012.
L. A. Brooks, A. C. Zander, and C. H. Hansen, "Investigation into the feasibility of using a
parametric array control source in an active noise control system," in Proceedings of
ACOUSTICS, 2005, pp. 39-45.
Y.-S. Lee, "Numerical solution of the KZK equation for pulsed finite amplitude sound
beams in thermoviscous fluids," 1993.
M. R. F. Kidner, C. Petersen, A. C. Zander, and C. H. Hansen, "Feasibility study of
localized active noise control using an audio spotlight and virtual sensors," in Proc
acoustics, 2006.
C. Hedberg, K. Haller, and T. Kamakura, "A self-silenced sound beam," Acoustical Physics,
vol. 56, pp. 637-639, 2010.
H. Nomura, C. M. Hedberg, and T. Kamakura, "Numerical simulation of parametric sound
generation and its application to length-limited sound beam," Applied Acoustics, vol. 73, pp.
1231-1238, 2012.

ICSV21, Beijing, China, 13-17 July 2014