[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: auditory localization


Great question, and one that I have given considerable thought. I don't
know that there is any easy answer as to the "correct" number. However,
any time you allow participants to visually observe the total number of
possible sources, then the number of possible responses, and the
corresponding probability of being correct on any given trial is
restricted. Depending on the precision that you are looking for in the
measure, you would want to either include more sources or more potential
response locations. This could be done by either increasing the number
of sources, or by placing inactive loudspeakers between your present
loudspeakers. The use of inactive or "dummy" loudspeakers would increase
the number of total possible responses without increasing the number of
source locations. 

For instance, if you used only 4 loudspeakers as sources and response
possibilities, the chance of a participant getting any single
presentation correct would be 1/4 or 25%. If you increase the number of
response locations (to say, 10 -- 2 additional loudspeakers between each
pair of sources) without increasing the number of actual sources, you've
essentially tricked the participant into thinking that there are more
than 4 possible choices. I'm not 100% sure (my mind is not comprehending
the implications on probability for this), but the chances of a correct
response could reasonably be expected to decrease. Eventhough the number
of sources is the same, the number of possible responses has increased. 

One side note is that the use of additional loudspeakers to "stretch" a
listener's response range will work to a greater or lesser extent
depending on the signal characteristics, the distance between listener
and loudspeakers, and other related variables. For instance, if the
listener-to-loudspeaker distance is on the order of 1 m, the horizontal
separation between loudspeakers for a given azimuth is much smaller than
if the listener-to-loudspeaker distance is 4 m. So, the addition of
dummy loudspeakers in the smaller distance may be more successful than
in the larger distance. Our visual input is very strong and can override
our auditory input as evidenced by phenomena such as the ventriloquist
effect. For this reason, the participant has to believe that the sounds
can come from any of the loudspeakers, so things such as wires need to
be "connected" for all loudspeakers, including those that serve as

Good luck!

Paula P. Henry, Ph.D., CCC-A
Research Audiologist
Visual and Auditory Processes Branch
Human Research and Engineering Directorate
Army Research Laboratory
Aberdeen Proving Ground, MD

-----Original Message-----
From: AUDITORY Research in Auditory Perception
[mailto:AUDITORY@LISTS.MCGILL.CA] On Behalf Of Elyana Makowski
Sent: Friday, May 07, 2004 9:30 AM
Subject: auditory localization

I am a PhD student and my research project is on the use of localization
cues (ILD and ITD) in the  horizontal plane.  My set up consists of 11
speakers arranged at 18 degree intervals from -90 to +90 degrees in the
azimuthal plane.  Currently, I have tested normally hearing subjects,
and will soon be testing hearing aid users and cochlear implant users.
Recently, an issue has been brought up as to the number of speakers that
I am using, with the suggestion being made that better results could be
measured if I used more speakers.  My question is does anyone know how
many speakers are necessary to accurately measure the average
localization error?
  How many speakers are enough?
Thank you,

It's fast, it's easy and it's free. Get MSN Messenger today!