within- vs. across-stream organization

[Greg Sandell <sandell@xxxxxxxxxxxxxxxxxxx>]

Pierre, Bill and Punita,

Thanks for your suggestions in response to my posting.  However, all
the works cited (musical or experimental) seem to be concerned with
differences between voices that enhance streaming (e.g. one voice being
soft and the other loud, one voice being of timbre A and the other of
timbre B).  This fundamental aspect of stream segregation was not what I
was asking about (although I was glad to have been nudged towards Hartmann &
Johnson, and enjoyed looking over the pieces mentioned by Pierre).

The case where one voice has timbre A and the other timbre B enhances
streaming by highlighting differences *across* the voices.  This indeed
is one form of "organization." However, I am interested in forms of
organization *within* the individual voices that enhance their ability to
stream.

Perhaps the best way to clarify the question (which I sense it needs) is
by another example.

Say I have two alternating human voices singing a list of random numbers
on a single pitch, and the voices are the same person (your terminal
will need a non-proportionally-spaced font to read this):

voice 1:  5    1    8    2    3    6    7    9    1    4    0
voice 2:    6     2    1    8   2    2    6    3    5    3

          (time ------------------------------------------>  )

There will be nothing differentiating the voices, so only a single
stream will form.  We can cause streaming to occur, however, if we
differentiate the voices in some way.  We could put them at different
pitch levels; we could make them different persons; we could have one
voice reading odd numbers and the other voice reading even numbers
(although I'm not sure that would be enough).  That would be an across
stream organization.

But consider the following within-stream organization.  Suppose voice 1
sings short sequences of numbers that ascend, while voice 2 sings short
sequences of numbers that descend, as shown below.  This would be a
within-stream organization.

voice 1:  5    6    7    2    3    4    5    4    5    6    7
voice 2:    6    5    4    3    9    8    7    3    2    1

Will this form streams?  Maybe not on the basis of that difference
alone.  But will it *enhance* stream formation?  For example, if it had been
found that a 200 Hz separation between the voices in the first case was
the minimum distance at which streams would form, will this
within-stream organization reduce that required distance to, say, 150 Hz?

And of course, is there any research which looks into this kind of
thing?

I hope this example makes it more clear; thank you for your patience!

--
Gregory J. Sandell  (sandell@epunix.sussex.ac.uk)
Lab. of Experimental Psychology, University of Sussex
Brighton BN1 9QG England  +44-273-678058