the effect of signal onset/offset envelope on underwater detection thresholds of a harbor seal...

The effect of signal onset/offset envelope on underwater detection thresholds of a harbor seal (Phoca vitulina)

S. D. Turnbull and J. M. Terhune

Department of Biology, University of New Brunswick, P.O. Box 5050, Saint John, New Brunswick E2L 4L5, Canada

(Received 3 August 1994; accepted for publication 24 February 1995)

The effects of signal onset/offset envelope on the underwater hearing thresholds of a harbor seal (Phoca vitulina) were measured. Pure-tone, 540-ms pulses at 2, 4, 8, and 16 kHz were presented as test signals. An ANOVA revealed that there were no significant differences between repeated threshold measures for abrupt onset/offset signal envelopes versus slow onset/offset signal envelopes [F=4.380, d.f.=(1,18), p>0.05]. Seal vocalizations which have an abrupt onset/offset may be serving a short-range communicative function by helping the listener determine the direction of a nearby sender. ̧ 1995 Acoustical Society of America.

PACS numbers: 43.80.Lb, 43.80.Jz, 43.80.Ka

INTRODUCTION

Watkins and Schevill (1979) suggest that the characteristics of many harp seal (Phoca groenlandica) vocalizations are opposite to the equivalent features of random bursts of ambient noises found in the marine environment. These con-

trasting vocal features may enhance the signal detection abilities of conspecifics in the presence of ambient noise (Watkins and Schevill, 1979).

One such call feature is that of increasing amplitude throughout the sound sequence with an abrupt termination (Watkins and Schevill, 1979). Some ambient noises in the marine environment have their greatest sound-pressure levels toward the beginning or middle of the sound with diminish- ing intensity toward the end. The termination of such sounds is often lost due to transmission losses and reverberation

(Urick, 1983). In this study, the effect of abrupt versus gradual signal

onsets and offsets on the underwater pure-tone detection thresholds of a harbor seal (Phoca vitulina) were measured.

I. MATERIALS AND METHODS

The subject was an 11-yr-old male harbor seal whose pure-tone hearing thresholds are known (Terhune, 1988, 1991; Turnbull and Terhune, 1990, 1993, 1994). The seal was housed in an above ground indoor pool (see Turnbull and Terhune 1990, 1993).

The equipment was set up as shown in Fig. 1. An amplitude control box (ACB; Fig. 1), which contained four set- tings, controlled the output characteristics of the signal envelope. The four signal envelopes were abrupt on/abrupt off (A/A), slow on/slow off (S/S), abrupt on/slow off (A/S), and slow on/abrupt off (S/A) (Table I). There were no transients or switching noises generated by this equipment. A Wavetek 20 signal generator provided a continuous pure-tone signal into the ACB (Fig. 1). This signal generator was set at the frequency to be tested (either 2, 4, 8, or 16 kHz). The final signal output was broadcast by a Bruel & Kj•er 8100 hydrophone (Fig. 1).

Output signals were recorded for analysis. A Bruel & Kj•er 8100 hydrophone was placed where the seal's head

would be when he pressed the stimulus switch with his nose. The received signal was amplified using a Bruel & Kj•er 2635 amplifier. Sound-pressure levels were calibrated using a Bruel & Kj•er 2203 sound level meter, equipped with a 1613 octave filter. The received signals were recorded on a Sony TCD-D3 DAT recorder. The temporal, amplitude, and spectral characteristics of the test signals were analyzed using a Loughbourgh Sound Images Ltd. speech workstation (Fig. 2).

When set in the A/A mode, the onset (rise time) and offset (decay) durations were less than 2.0 ms. The signal had a constant amplitude for 540 ms. This was consistent for each of the four frequencies tested. In the S/S mode, the signal rise time was 105 ms before reaching peak amplitude and was constant for 300 ms. The signal decay time was 135 ms which was followed by an abrupt offset. These features were also constant for each of the four frequencies tested. The onset and offset characteristics of the signal types are given in Table I.

Thresholds (50% correct detection of the presence or absence of the signal) were calculated for each of the fre-

Stimulator c•

Signal Generator I

Resistance • Amplifier Box

Oscilloscope

•ntrol Soal's Switch

Signal Generator 2

Bandpass FlEer

Freq. Spec.

•o•' plitude :rol Box

Signal Generator 3

Hydrophone

FIG. 1. Block diagram of the equipment used in the abrupt onset/offset study.

78 J. Acoust. Sec. Am. 98 (1), July 1995 0001-4966/95/98(1)/78/3/$6.00 ¸ 1995 Acoustical Society of America 78

Redistribution subject to ASA license or copyright; see http://acousticalsociety.org/content/terms. Download to IP: 132.174.255.116 On: Fri, 28 Nov 2014 15:26:11

2 kHz

4 kHz

8 kHz

16 kHz

Time

FIG. 2. Envelope characteristics of two signal types; abrupt on/off on the left, slow on/off on the right. All pulses were 540 ms long.

quencies tested using a constant stimulus method (Guilford, 1954). The threshold measurement techniques are similar to those described in Turnbull and Terhune (1994).

Each testing session consisted of 40 trials, 30 of which consisted of 15 catch trials (no signal presented) and 15 signal trials (at the signal levels being tested), presented in an order which followed Gellerman's schedule (Gellerman, 1933). The first five served as a "warm-up" and the last five as a "retraining" exercise for the seal and were not used in threshold calculations (Turnbull and Terhune, 1994).

The 30 testing trials were further divided into three blocks of 10, each block consisting of 5 signal and 5 catch trials. The signal levels in each block were 4 dB apart. Three different signal levels (8-dB range) were presented to the seal in a random order during each testing session. Each group of three different signal levels were presented to the seal over the course of three separate, but consecutive, test-

TABLE I. Onset and decay characteristics of the pulses in the slow on and slow off mode.

Frequency Slow onset Slow offset Abrupt offset (kHz) (dB/105 ms) (dB/135 ms) (dB)

2 23 15 8

4 30 22 8

8 35 25 10

16 45 31 14

ing sessions. This gave 30 trials (15 signal and 15 catch trials) at each signal level (Turnbull and Terhune, 1994).

Once these three sessions were completed, the next three lower signal levels were presented to the seal in the same manner. This procedure continued until the seal's response rate fell to 60% correct or less (both signal and catch trials summed) for one of the signal levels. Following this session, the next signal level was 2 dB above that in which he scored 60% or below and the other two signal levels were 4 and 8 dB above this level. This gave three to six data points at 2-dB intervals from close to threshold to near certain detec-

tion levels (Turnbull and Terhune, 1994). Three threshold measurements were determined for each

frequency for the two main signal envelopes under investi- gation (A/A and S/S) and also at 2 kHz (S/A). The data from each of the repeated measures were combined to calculate one threshold at each frequency using a constant stimulus method (Guilford, 1954; see Table II).

Only one threshold measure was conducted at 16 kHz for each of the A/S and S/A modes. On two occasions, a fourth measure was taken (2 and 8 kHz, A/A). The threshold measures were determined in frequency blocks (i.e., all at 2 kHz) but the type of signal envelope was randomly presented.

An ANOVA was used to determine if there was a significant difference between the repeated threshold measures (dependent variable) of the two signal envelopes A/A and S/S. A Kruskal-Wallis test was performed on the combined data in Table II (A/A and S/S signal types only).

II. RESULTS

The greatest threshold shift between the two signal types (A/A and S/S) was at 2 kHz (12 dB). Three other threshold sets measured at this frequency overlapped or were within 1-2 dB of each other. An ANOVA revealed no significant difference between the repeated threshold measures for A/A

TABLE II. Underwater hearing thresholds of a harbor seal (Phoca vitulina) for pulses having various onset and offset characteristics (see text). Thresholds are given in dB re: 1 /xPa plus or minus one standard deviation. (n) =the number of combined threshold measures.

Thresholds

Frequency Signal characteristics (kHz) A/A (n) S/S (n) A/S (n) S/A (n)

2 64_+14 4 70_+9 3 ...... 63_+9 3

4 59_+ 8 3 59_+8 3 ............

8 62+11 4 61_+7 3 ............

16 52 +- 12 3 55_+9 3 51 _+ 12 1 52_+ 7 1

79 J. Acoust. Soc. Am., Vol. 98, No. 1, July 1995 S.D. Turnbull and J. M. Terhune: Underwater thresholds of a seal 79


and S/S [F=4.380, d.f.=(1,18), p>0.05]. There was no in- teraction between frequency and signal envelope [F = 1.877, d.f. = (3,18), p>0.10].

All data for the repeated thresholds (per frequency) were combined to produce one threshold estimate (Table TT). A Kruskal-Wallis test indicated no difference between the

thresholds of the A/A and S/S signal types given in Table TT [H-0.0211, d.f.=(1,8), p>0.10].

III. DISCUSSION

There are no differences between the threshold measures

of the two main signal types. The individual thresholds for A/A and S/S signal types overlapped at each of the frequencies tested. The variability within the animal itself was greater than the effect of the two main signal types. This was not the case for other studies using the same seal and methods (Turnbull and Terhune, 1993, 1994).

All of the seal's measured thresholds were above the

theoretical lower limit, the spectrum level of ambient noise, plus the highest critical ratios measured from this seal (Turn- bull and Terhune, 1990, 1993) and as such, were not likely masked by ambient noise levels. After finding no significant differences in the threshold measures, the decision was made to end the study without testing the seal with the other combinations of signal envelope (Table TT). If a signal having an A/A did not enhance the seal's signal detection abilities, it is unlikely that signals with combinations of S/A or A/S would produce any significant results. This is supported by the few measures that were taken at 16 kHz (Table II).

The S/S signal had a final abrupt offset of 8-14 dB. This abrupt offset would not likely provide any cue to the seal. The signal level was lowered until the seal reached threshold levels. The abrupt offset began far below the minimum signal level required to elicit a response from the seal, espe- cially at the higher frequencies.

Threshold determinations for this animal are similar

when presented with pulse durations of 50 ms or longer (Ter- hune, 1988). As such, increasing the duration of the signal would not likely produce any significant difference in the results. Signals longer than 200-300 ms do not result in lower signal detection thresholds in other species (Dooling, 1982; see Fay, 1988 for additional data; Gelfand, 1990).

Terhune (1974) suggests that the directional hearing abilities of harbor seals are typically mammalian. That is, seals are better at localizing a broadband noise than a pure tone because the broadband noise provides better interaural time cues. In some species of birds, broadband sounds with sudden onsets and offsets result in maximum locatability (Konishi, 1977; Marler, 1977). Such features may be impor- tant in short-range communication for birds (Wiley and Richards, 1982).

Harp seals produce short-duration clicks often in asso- ciation with other call types. These click vocalizations present a well-defined interaural time cue and may be serving to assist in localizing the direction of the nearby sender (M•hl et al., 1975). Vocalizations with abrupt onsets and offsets may serve the function of enabling the listener to determine the direction of the caller.

In traveling through the sea, sounds will be delayed, distorted, and weakened from multiple reflections and scat- tering (Urick, 1983). These multipath elements will obscure the abruptness of a signal's onset and offset (Wiley and Richards, 1982; Urick, 1983). Over great distances, up to 4 km for the harp seal (Terhune and Ronald, 1986) or up to 25 km for the Weddell (Leptonychotes weddelli) and bearded seals (Erignathus barbatus) (Thomas and Kuechle, 1982; Cleator et al., 1989), such effects would change the onset and offset of an A/A signal to more closely resemble that of an S/S signal type (Wiley and Richards, 1982). The findings of this study suggest that abrupt signal types will not likely increase the communication range relative to slow onset/ offset signal types.

Cleator, H. J., Stirling, I., and Smith, T. G. (1989). "Underwater vocalizations of the bearded seal (Erignathus barbatus)," Can. J. Zool. 67, 1900- 1910.

Dooling, R. J. (1982). "Auditory perception in birds," in Acoustic Commu- nication in Birds, edited by D. E. Kroodsma, E. H. Miller, and H. Ouellet (Academic, New York), Vol. I, pp. 95-130.

Fay, R. R. (1988). Hearing in Vertebrates: A Psychophysics Databook (Hill- Fay Associates, Winnetka, IL).

Gelfand, S. A. (1990). Hearing: An Introduction to Psychological and Physiological Methods (Marcel Dekker, New York), 2nd ed.

Gellerman, L. W. (1933). "Chance orders of alternating stimuli in visual discrimination experiments," J. Genet. Psychol. 42, 206-208.

Guilford, J.P. (1954). Psychometric Methods (McGraw-Hill, New York). Konishi, M. (1977). "Spatial localization of sound," in Dahlem Workshop

on Recognition of Complex Acoustic Signals, edited by T. Bullock (Dahlem, Berlin), pp. 127-143.

Marler, P. R. (1977). "The structure of animal communicative sounds," in Dahlem Workshop on Recognition of Complex Acoustic Signals, edited by T Bullock (Dahlem, Berlin), pp. 17-35.

Mfihl, B., Terhune, J. M., and Ronald, K. (1975). "Underwater calls of the harp seal (Pagophilus groenlandicus)," in Biology of the Seal, edited by K. Ronald and A. W. Mansfield (Rapp. P.-v. Reun. Cons. int. Explor. Mer) 169, 533-543.

Terhune, J. M. (1974). "Directional hearing of a harbour seal in air and water," J. Acoust. Soc. Am. 56, 1862-1865.

Terhune, J. M. (1988). "Detection thresholds of a harbor seal to repeated underwater high frequency, short duration, sinusoidal pulses," Can. J. Zool. 66, 1578-1582.

Terhune, J. M. (1991). "Masked and unmasked pure tone detection thresholds of a harbour seal listening in air," Can. J. Zool. 69, 2059-2066.

Terhune, J. M., and Ronald, K. (1986). "Distant and near-range functions of harp seal underwater calls," Can. J. Zool. 64, 1065-1070.

Thomas, J. A., and Kuechle, V. B. (1982). "Quantitative analysis of Weddell seal (Leptonychotes weddelli) underwater vocalizations at McMurdo Sound, Antarctica," J. Acoust. Soc. Am. 72, 1730-1738.

Turnbull, S. D., and Terhune, J. M. (1990). "White noise and pure tone masking of pure tone thresholds of a harbour seal listening in air and underwater," Can. J. Zool. 68, 2090-2097.

Turnbull, S. D., and Terhune, J. M. (1993). "Repetition enhances hearing detection thresholds in a harbour seal (Phoca vitulina)," Can. J. Zool. 71, 926-932.

Turnbull, S. D., and Terhune, J. M. (1994). "Descending frequency swept tones have lower thresholds than ascending frequency swept tones for a harbor seal (Phoca vitulina) and human listeners," J. Acoust. Soc. Am. 96, 2631-2636.

Urick, R. J. (1983). Principles of Underwater Sound, (McGraw-Hill, New York), 3rd ed.

Watkins, W. A., and Schevill, W. G. (1979). "Distinctive characteristics of underwater calls of the harp seal (Phoca groenlandica)," J. Acoust. Soc. Am. 66, 983-988.

Wiley, R. H., and Richards, D. G. (1982). "Adaptations for acoustic communication in birds: Sound transmission and signal detection," in Acoustic Communication in Birds, edited by D. E. Kroodsma, E. H. Miller, and H. Ouellet (Academic, New York), Vol. I, pp. 131-181.

80 J. Acoust. Soc. Am., Vol. 98, No. 1, July 1995 S.D. Turnbull and J. M. Terhune: Underwater thresholds of a seal 80


the effect of signal onset/offset envelope on underwater detection thresholds of a harbor seal...

Documents