Re: Dolphins & Acoustic Imaging (was "stereovision question")

[P3D John W Roberts <roberts@xxxxxxxxxxxxxxxxx>]

>Date: Tue, 9 Apr 1996 13:16:40 -0500
>From: P3D Josh Rubin  <jnr@xxxxxxxxx>
>Subject: Dolphins & Acoustic Imaging (was "stereovision question")

>I've always been fascinated by dolphins.  Their sonar is
>reportedly incredibly detailed and complex.  But more
>amazing, however, is their language.  Apparently, the
>information that they emit is very dense, and it
>resembles the sonar/echolocation information that they
>*receive* back from their environment, leading some to
>hypothesize that they communicate using moving pictures.
>They have several tongues or larynxes or something, and
>their rounded foreheads contain a lens-shaped dome made
>of fat, which functions as a focusing and directional
>lens for their sonic emissions.  They have more cerebral
>cortex than humans.

>Joshua N. Rubin (jnr@xxxxxxxxx) 

Some cetaceans put out very complex sounds - this could be useful from
a processing viewpoint, or it could be some sort of "ID" to allow them
to distinguish their signals from others. Individuals of some species can
apparently change their signals over the course of months, or learn calls
from others.

Some enthusiasts seem to assume that if some major component of the signal
is at x Hz, then the sea mammals are communicating speech information at
x bits per second. I've been skeptical of this extreme claim, but have often
wondered if there's a more reasonable estimate. A show on cable last week
discussed the communications activity of killer whales (which are thought to
be extremely intelligent, by many accounts more so than dolphins). The
patterns of learning and changing signals were described. A researcher
stated that he does not believe that the killer whales (orcas) have what a
human would consider true speech, but their handling and use of sounds to
convey information to one another appears to be a precursor of generalized
speech - something that could evolve over the course of time.

I would be interested in knowing whether dolphins can build up a full 3D
model of their surroundings without moving their heads. Humans seem to have
some ability to determine the height of a sound source as well as 
horizontal position. It is possible that the folds of the outer ear
change the spectral characteristics of the incoming sound in a position-
dependent way that facilitates this. If so, one would expect that some
sounds would be easier to track down than others. I seem to remember
noticing that pure tones are harder to locate than sounds that are richer
in harmonics. Discontinuities in the sound and motion of the head may be
useful too.

John R


------------------------------