47-channel burst-mode recording hydrophone system enabling
measurements of the dynamic echolocation behavior of
free-swimming dolphins (L)
Electrical Measurements, Faculty of Engineering LTH, Lund University, P.O. Box 118,
SE-221 00 Lund, Sweden
Kolmarden Wildlife Park, SE-618 92 Kolmarden, Sweden and Department for Physics, Chemistry and
Biology (IFM), Linkoping University, SE-581 83 Linkoping, Sweden
Johan Nilsson and Tomas Jansson
Electrical Measurements, Faculty of Engineering LTH, Lund University, P.O. Box 118, SE-221 00 Lund,
Stan A. Kuczaj
Department of Psychology, The University of Southern Mississippi, P.O. Box 5025, Hattiesburg, Mississippi
Monica Almqvist and Hans W. Persson
Electrical Measurements, Faculty of Engineering LTH, Lund University, P.O. Box 118,
SE-221 00 Lund, Sweden
共Received 9 February 2009; revised 30 June 2009; accepted 30 June 2009兲
Detailed echolocation behavior studies on free-swimming dolphins require a measurement system
that incorporates multiple hydrophones 共often ⬎16兲. However, the high data ﬂow rate of previous
systems has limited their usefulness since only minute long recordings have been manageable. To
address this problem, this report describes a 47-channel burst-mode recording hydrophone system
that enables highly resolved full beamwidth measurements on multiple free-swimming dolphins
during prolonged recording periods. The system facilitates a wide range of biosonar studies since it
eliminates the need to restrict the movement of animals in order to study the ﬁne details of their
sonar beams. © 2009 Acoustical Society of America. 关DOI: 10.1121/1.3184536兴
PACS number共s兲: 43.80.Ev, 43.80.Ka, 43.60.Qv 关WWA兴Pages: 959–962
The echolocation of dolphins and other odontocetes has
been extensively studied, a primary focus being the beam
axis 共Nachtigall and Moore, 1988;Thomas and Kastelein,
1990;Au, 1993;Villadsgaard et al., 2007;Kyhn et al.,
2009兲. Recording scenarios that focus on the beam axis typi-
cally provide accurate and effective measurements. However,
since these situations often require that the echolocating ani-
mal be kept stationary, it is likely that the full dynamics of
the sonar beam has not yet been described. In addition, these
static test conditions are by deﬁnition impossible to use in
other important contexts, such as in studies of the spontane-
ous use of echolocation by free-swimming dolphins, object
investigation behavior in groups of dolphins, and calf mim-
icry of their mother’s echolocation clicks. Although detailed
sonar studies have been conducted with free-swimming dol-
phins 共Sigurdson, 1996;Martin et al., 2005, among others兲,
these studies used relatively few hydrophones and conse-
quently had limited sonar beam coverage. As a result, much
is known about the beam axis, and little is known about the
rest of the beam.
Recording dolphin sonar in dynamic test conditions re-
quires a system able to deal with varying measurement pa-
rameters, including the animal’s relative distance to the re-
ceivers, the number of animals present, the orientation of the
beam relative to the receivers, and the required time for the
animal to respond to an echolocation task. Consequently, a
measurement system capable of long recording periods, large
beamwidth coverage, and high spatial and temporal resolu-
tions is needed in order to localize the beam axis and to
measure the rest of the beam with high accuracy from free-
Multi-channel sonar recording systems have previously
been reported by Miller and Tyack 共1998兲,Ball and Buck
共2005兲,Starkhammar et al. 共2007兲,Amundin et al. 共2008兲,
and Moore et al. 共2008兲, among others. The most extensive
system developed thus far was created by Moore et al.
共2008兲, who employed 24 hydrophones in an array. It was
designed for high spatial resolution across the array area
since it was used in a study of beamwidth control in a bottle-
nose dolphin 共Tursiops truncatus兲and so required measure-
a兲Author to whom correspondence should be addressed. Electronic mail:
J. Acoust. Soc. Am. 126 共3兲, September 2009 © 2009 Acoustical Society of America 9590001-4966/2009/126共3兲/959/4/$25.00
ments of small spatial alterations of the amplitude distribu-
tion across the whole beam cross-section. The system
recorded data during 5 s intervals with a sample rate of 312.5
kS/s, resulting in a data ﬂow rate of 16 Mbytes/s. Although
this system worked well for its purpose, the extremely high
data ﬂow rate makes it unsuitable for measuring echoloca-
tion behavior in free-swimming dolphins. Using this system,
a single minute of recordings would result in a 1 Gbyte large
A measurement system optimized for dynamic test con-
ditions requires large beamwidth coverage and higher spatial
and temporal resolutions. The system must also be able to
record for longer time periods in order to be useful in the
ﬁeld. This requires an increase in the number of array ele-
ments, the physical size of the array, and, preferably, also the
sample rate. In addition, the problems associated with ex-
tremely high data ﬂow rates must be solved.
This report describes a system with a measurement ap-
proach optimized for studies of dolphin sonar under dynamic
test conditions. The system uses a larger number of hydro-
phones 共47 channels兲, allows an increased sample rate 共1
MHz兲, and acquires data with lower data ﬂow rates than
previously reported multi-hydrophone systems. This facili-
tates full waveform recordings of the echolocation activity of
dolphins during prolonged time periods and comprehensive
beamwidth coverage, provided that the dolphins are within
reasonable distance from the screen. In addition, the ap-
proach enables real-time analysis and real-time visualization
of data during recordings. This measurement approach and
recording system enables researchers to investigate dolphin
sonar use in a wider range of contexts than has previously
been made possible. In the following sections, the authors
describe this system and provide examples of its potential
uses with free-swimming dolphins.
II. MATERIALS AND METHODS
A new 47-channel dolphin echolocation measurement
system was developed and tested with a group of 19 Atlantic
bottlenose dolphins, housed together in a large open sea pen
at Roatán Institute of Marine Science, Roatán, Honduras. All
dolphins were allowed to swim freely and to explore at will
objects suspended in front of or behind the recording hydro-
phone array. The size of array was 0.75⫻0.75 m2. Figure
1共A兲shows a schematic representation of the experimental
setup, and Fig. 1共B兲shows a photograph of the setup during
Typically, multi-hydrophone arrays used in biosonar ap-
plications produce a considerable amount of data due to the
relatively high sample rate required to reconstruct the full
waveform accurately in post-analysis 关approximately ten
times the maximum frequency for accurate visualization
共Buchla and McLachlan, 1992兲兴. The long recording times
required in test conditions with spontaneously echolocating
dolphins using a 47-element array system would result in
unmanageably large data ﬁles after only a few seconds, using
the data acquisition approaches in previous systems. There-
fore, an alternative approach to continuous sampling of all
parallel channels was needed for the 47-element hydrophone
In order to facilitate longer recordings and to keep the
data ﬂow rate manageable with a high sample rate, the sys-
tem was designed to be triggered by one echolocation click
at a time and to not sample data during the silent periods that
occur between clicks in click trains. Figure 2describes the
basic data acquisition method. Basically, the system only ac-
quires a small pre-set number of samples, containing only
the actual click, when a hydrophone output exceeds the cho-
sen trig level. This is referred to here as burst-mode sampling
共as opposed to continuous sampling兲. The time stamps 共T1
and T2 in Fig. 2兲correspond to the start time of each sample
burst and are stored in association with each click.
Successful burst-mode sampling required data acquisi-
tion hardware capable of extremely fast re-triggering of mea-
surements even after the particularly short inter-click inter-
vals 共⬍1ms兲that may occur in click train “buzzes” 共see
FIG. 1. 共Color online兲共A兲A schematic drawing of the system setup in the ﬁeld test. 共B兲An underwater photograph of the experimental setup.
FIG. 2. 共Color online兲Basic data acquisition method.
960 J. Acoust. Soc. Am., Vol. 126, No. 3, September 2009 Starkhammar et al.: Letters to the Editor
Herzing, 1996;Herzing and dos Santos, 2004兲or when mul-
tiple dolphins echolocate concurrently. The total time re-
quired to ﬁnish one burst-mode acquisition, re-trig, and start
a new one is referred to here as the system’s rearm time.
In order to capture the echolocation data across the en-
tire array, the system was designed to simultaneously trigger
all channels, regardless of which hydrophone was hit ﬁrst by
the sonar beam. Signals were acquired from 6⫻8 parallel
and synchronized channels using six PXI-digitizer cards 共NI
PXI-5105, National Instruments, USA兲, each with eight si-
multaneously sampling analog-to-digital-channels and 12 bit
voltage resolution. One of the digitizer channels was used as
a trig-channel. The 47 remaining channels were wired to the
47 individually pre-ampliﬁed hydrophones, all ampliﬁed
with either 35 or 50 dB, depending on the measurement situ-
ation. The sum of all the 47 pre-ampliﬁed hydrophone sig-
nals was wired to the pre-designated trig-channel using a
separate signal summarizing hardware circuitry.
The software was optimized to ensure fast rearm time
and real-time visualization and analysis of data, aspects that
were typically the most time-consuming as well as important
determinants of the overall system performance. This im-
provement was accomplished with software created in LAB-
VIEW 8.6™共National Instruments, USA兲enabling dual-core
operation of the CPU.
III. RESULTS AND DISCUSSION
The echolocation measurements of the free-swimming
dolphins in the group of 19 individuals demonstrated that the
system is capable of measuring the full waveforms of the
spontaneous sonar activity in the group. Measurements were
obtained with all 47 simultaneously sampling hydrophones at
a sample rate of 1 MHz during measurement sessions of
various lengths 共often ⬎15 min兲. Each acquisition in these
tests was set to record during a time window of 150
around each click event with a pre-trig time period of 40
共see Fig. 2兲. The total duration of the measurement sessions
was determined by the tourist activity at the facility. Sessions
were never aborted due to system failure.
Figure 3共A兲shows the corresponding waveforms of one
single click acquired simultaneously by all 47 hydrophones.
Each position of the numbered elements corresponds spa-
tially to the hydrophones shown in Fig. 1. The level of spa-
tial resolution and comprehensive beamwidth coverage pro-
vides new information concerning the entire cross-section of
the beam during one single click. As an example, Fig. 3共B兲
shows the relative energy distribution of the cross-section in
the beam. The energy in the click is coded as an indentation
of the interpolated three dimensional 共3D兲surface where
high energies “push” the surface downward, away from the
A suspended scuba tank provided a way to further dem-
onstrate the functionality of the system by shadowing the
hydrophones in the center of the screen when a dolphin
echolocated toward the tank from small bearing angles 共i.e.,
close to the perpendicular to the screen兲. This shadowing
effect is clearly seen as a ridge in the middle of the beam
energy plot in Fig. 3共B兲.
The presented level of spatial resolution and comprehen-
sive beamwidth coverage give an unprecedented detailed
measure of the spectral content within the cross-section of
the beam 关Fig. 3共C兲兴. The minimized data ﬂow rates make it
possible to view the spectral content even in real-time. The
system also allows researchers to study entire echolocation
scan sequences in detail by processing all successive clicks
and then re-playing them at variable frame rates. The reso-
lution of the measurements enables detailed re-plays of the
propagation of every single click across the array, further
facilitating quantiﬁed detailed studies of the dynamic varia-
tions in the echolocation behavior of dolphins during pro-
longed periods of time.
Benchmark tests of the system performance showed that
the low data ﬂow makes possible recordings during 20 min
of constantly echolocating animals with inter-click intervals
of 20 ms before the data ﬁle size reaches 1 Gbyte and be-
comes unreasonably large for commonly used post-
processing tools 共such as MATLAB®, The MathWorks™Inc.,
USA兲. This is a considerable improvement compared to pre-
viously published systems. The low data ﬂow rate of the
present system 共0.83 Mbyte/s under the conditions in the
benchmark tests兲is even more advantageous in more realis-
tic measurement scenarios, where free-swimming dolphins
echolocate spontaneously and when minute long silent peri-
ods in the recordings are likely to occur.
In conclusion, the presented 47-element hydrophone
system enables recordings with improved spatial and tempo-
ral resolutions of the cross-section of the echolocation beam
FIG. 3. 共Color online兲Visualization of one echolocation click acquired at the 47 hydrophone positions in the experimental setup. This click can be visualized
by plotting 共A兲the waveforms, 共B兲the relative energy distribution within the beam, or 共C兲the color coded frequency distribution within the beam. High click
energy in 共B兲is illustrated by letting it push the interpolated 3D surface downward, away from the echolocating dolphin.
J. Acoust. Soc. Am., Vol. 126, No. 3, September 2009 Starkhammar et al.: Letters to the Editor 961
of free-swimming dolphins. Moreover, the system makes
possible extended recording periods due to the minimized
data ﬂow rate. These features facilitate the reconstruction,
visualization, and re-play of signiﬁcant aspects of the clicks
during extended echolocation sequences. The system’s abil-
ity to process information from free-swimming dolphins in
groups opens the door to a completely new range of studies,
which will help us to better understand the functions of dol-
phin sonar since it eliminates the need to restrict the move-
ment of animals in order to study the ﬁne details in their
Amundin, M., Starkhammar, J., Evander, M., Almqvist, M., Lindstrom, K.,
and Persson, H. W. 共2008兲. “An echolocation visualization and interface
system for dolphin research,” J. Acoust. Soc. Am. 123, 1188–1194.
Au,W.L.共1993兲.The Sonar of Dolphins 共Springer, New York兲.
Ball, K. R., and Buck, J. R. 共2005兲. “A beamforming video recorder for
integrated observations of dolphin behavior and vocalizations,” J. Acoust.
Soc. Am. 11 7, 1005–1008.
Buchla, D., and McLachlan, W. 共1992兲.Applied Electronic Instrumentation
and Measurement 共Macmillan, New York兲, pp. 384–388.
Herzing, D. L. 共1996兲. “Vocalizations and associated underwater behavior of
free-ranging Atlantic spotted dolphins, Stenella frontalis and Bottlenose
dolphins, Tursiops truncatus,” Aquat. Mamm. 22, 61–79.
Herzing, D. L., and dos Santos, M. E. 共2004兲. “Functional aspects of echolo-
cation in dolphins,” in Echolocation in Bats and Dolphins, edited by J. A.
Thomas, C. F. Moss, and M. Vater 共University of Chicago Press, Chicago兲,
Kyhn, L. A., Tougaard, J., Jensen, J. F., Wahlberg, M., Stone, G., Yoshinaga,
A., Beedholm, K., and Madsen, P. T. 共2009兲. “Feeding at a high pitch:
Source parameters of narrow band, high-frequency clicks from echolocat-
ing off-shore hourglass dolphins and coastal Hector’s dolphins,” J. Acoust.
Soc. Am. 125, 1783–1791.
Martin, S. W., Phillips, M., Bauer, E. J., Moore, P. W., and Houser, D. S.
共2005兲. “Instrumenting free-swimming dolphins echolocating in open wa-
ter,” J. Acoust. Soc. Am. 11 7, 2301–2307.
Miller, P. J., and Tyack, P. L. 共1998兲. “A small towed beamforming array to
identify vocalizing resident killer whales 共Orcinus orca兲concurrent with
focal behavioral observations,” Deep-Sea Res., Part II 45, 1389–1405.
Moore, P. W., Dankiewicz, L. A., and Houser, D. S. 共2008兲. “Beamwidth
control and angular target detection in an echolocating bottlenose dolphin
共Tursiops truncatus兲,” J. Acoust. Soc. Am. 124, 3324–3332.
Nachtigall, P. E., and Moore, P. W. B. 共1988兲.Animal Sonar, Process and
Performance 共Plenum, New York兲.
Sigurdson, J. E. 共1996兲. “Open-water echolocation of bottom objects by
dolphins 共Tursiops Truncatus兲,” J. Acoust. Soc. Am. 100, 2610.
Starkhammar, J., Amundin, M., Olsén, H., Ahlmqvist, M., Lindstrom, K.,
and Persson, H. W. 共2007兲. “Acoustic touch screen for dolphins,” in Pro-
ceedings of Fourth International Conference of Bio-Acoustics, Institute of
Acoustics, Loughborough University, edited by S. Dimble, P. Dobbins, J.
Flint, E. Harland, and P. Lepper, pp. 55–60.
Thomas, J. A., and Kastelein, R. 共1990兲.Sensory Abilities of Cetaceans
共Plenum, New York兲.
Villadsgaard, A., Wahlberg, M., and Tougaard, J. 共2007兲. “Echolocation sig-
nals of wild harbour porpoises, Phocoena phocoena,” J. Exp. Biol. 210,
962 J. Acoust. Soc. Am., Vol. 126, No. 3, September 2009 Starkhammar et al.: Letters to the Editor