10. Acoustic Models of Sound Production
457
ter to dimensions within the sound production tract. Much of human speech
analysis is based on the presence of formants or resonant frequencies in the
sounds we make. The character of human speech provides indirect me surements of the speaker's vocal tract, as "each human voice attains its unique
tone quality (either speaking or singing) from the particular resonant frequencies of the three resonant cavities, the larynx, the mouth and the nasal
cavity" (Berg and Stork 1995, p. 108). Blue whale calls appear to lack such
formants, thus much of human voice analysis methodology (i.e., homologous deconvolution) may be inappropriate for analysis of blue whale calls.
Formants in a blue whale call would be seen as energy corresponding in'
time and duration with the fundamental frequency, but at nonharmonic frequencies. The type "A" pulsive call character is composed of temporally
offset sounds of noninteger frequencies repeated about 20 times in the California animals (Figure 1O.14b), thus resonance may playa role in the production of this sound, but because there is an apparent time offset between
pulses it is unclear how such a resonant system would work. The type "B"
call shows only harmonics, such as would appear in any system driven near
its physical limits, and lacks frequencies of noninteger multiples.
8. Predictions from the Blue Whale Sound
Production Model
8.1 Dive Profile and Feeding Incompatibility
If the proposed blue whale sound production model is valid, feeding activity would appear incompatible with the repetitive call behavior commonly
observed for tens of hours showing only breathing gaps. This incompatibility is predicted both by the necessary connection of the trachea with the
palatal sphincter during call production, which may block the esophagus,
and by the incompatibility of the required dive profile with searching for
and consuming prey. The type "B" call sound production model predicts the
whale to be either diving or ascending during call production and limits the
maximum type "B" call production depth to the lung collapse depth of
90m. Matched field acoustic methods (Baggeroer et al. 1993) allow the
depth of a blue whale during call production to be measured. Preliminary
results from analysis of a single call (D'Spain et al. 1995) suggest the call
depth during a type "B" call is 34m averaged over the entire call duration,
a depth consistent with the model, where the depth averaged in time occurs
in the upper half of the required 90-m dive or ascent. Further applications
of matched field methods to blue whale recordings, and the development
of array Doppler and localization methods for motion detection in calling
blue whales, may result in significant call depth and vertical motion data
that can be used to test the model in the near future.
The intensity of the blue whale type "B" call is nearly constant over the
duration of the call, suggesting the ideal dive profile for the sound produc-
457
ter to dimensions within the sound production tract. Much of human speech
analysis is based on the presence of formants or resonant frequencies in the
sounds we make. The character of human speech provides indirect me surements of the speaker's vocal tract, as "each human voice attains its unique
tone quality (either speaking or singing) from the particular resonant frequencies of the three resonant cavities, the larynx, the mouth and the nasal
cavity" (Berg and Stork 1995, p. 108). Blue whale calls appear to lack such
formants, thus much of human voice analysis methodology (i.e., homologous deconvolution) may be inappropriate for analysis of blue whale calls.
Formants in a blue whale call would be seen as energy corresponding in'
time and duration with the fundamental frequency, but at nonharmonic frequencies. The type "A" pulsive call character is composed of temporally
offset sounds of noninteger frequencies repeated about 20 times in the California animals (Figure 1O.14b), thus resonance may playa role in the production of this sound, but because there is an apparent time offset between
pulses it is unclear how such a resonant system would work. The type "B"
call shows only harmonics, such as would appear in any system driven near
its physical limits, and lacks frequencies of noninteger multiples.
8. Predictions from the Blue Whale Sound
Production Model
8.1 Dive Profile and Feeding Incompatibility
If the proposed blue whale sound production model is valid, feeding activity would appear incompatible with the repetitive call behavior commonly
observed for tens of hours showing only breathing gaps. This incompatibility is predicted both by the necessary connection of the trachea with the
palatal sphincter during call production, which may block the esophagus,
and by the incompatibility of the required dive profile with searching for
and consuming prey. The type "B" call sound production model predicts the
whale to be either diving or ascending during call production and limits the
maximum type "B" call production depth to the lung collapse depth of
90m. Matched field acoustic methods (Baggeroer et al. 1993) allow the
depth of a blue whale during call production to be measured. Preliminary
results from analysis of a single call (D'Spain et al. 1995) suggest the call
depth during a type "B" call is 34m averaged over the entire call duration,
a depth consistent with the model, where the depth averaged in time occurs
in the upper half of the required 90-m dive or ascent. Further applications
of matched field methods to blue whale recordings, and the development
of array Doppler and localization methods for motion detection in calling
blue whales, may result in significant call depth and vertical motion data
that can be used to test the model in the near future.
The intensity of the blue whale type "B" call is nearly constant over the
duration of the call, suggesting the ideal dive profile for the sound produc-