DESCRIPTION
Most human displays seem to interact with the user via a visual screen. As more and more
information is placed on the screen, the screen becomes cluttered and the possibility
exists that important information will be overlooked. Using colour and increasing screen
size is not always the solution. Let's start taking advantage of the other four human
senses. Humans can naturally hear sounds in three dimensions. In other words, sounds can
be perceived from above, below, any direction, and distance. Without going into the
psychoacoustics details of 3D audio, suffice it to say that any monaural sound can be
transformed so that, when presented over headphones, the listener imagines that the sound
is outside of the head. Unlike dichotic presentation (i.e., a stereo signal), binaural
hearing---as natural three-dimensional hearing is known---can convey intimacy or urgency
when sounds approach the listener. When sounds are in the listener's ``sacred space'',
they can be difficult to ignore. Three-dimensional reproduction is being considered for a
wide range of applications (aside from the virtual reality community). Some preliminary
research has shown that binaural representation has several advantages over dichotic
reproduction. These areas of research include speech intelligibility, spatial orientation,
motor guidance, and teleconferencing. I shall talk about the motivation for 3D audio, and
discuss current and future work in 3D audio at DCIEM. The talk will conclude with a hands
on demo of Focal Point 3D Audio's binaural reproduction software and the opportunity to
try out different headphones.
SPEAKER
Robert Arrabito graduated with a M.Sc. in Computer Science from the University of Western
Ontario in 1990. His thesis investigated the feasibility of an automatic translation
system from publisher compositor tapes into Braille. Upon graduation, he joined the Human
Factors group at the Defence and Civil Institute of Environmental Medicine (DCIEM) as a
Defence Scientist where he is helping in the design and development of software to control
experiments. During the past year, he has been investigating the potential of binaural
reproduction as it applies to information display and decision making.
REVIEW
by Darrin Nowakowski It often requires someone of extraordinary circumstance to
illustrate the obvious. It is a matter of common knowledge that the vast majority of all
computer interfaces make pitiful use of sound as a form of information display. Most rely
on visual perception and motor skills for both input and output of information. Yet,
Robert Arrabito's blindness has not prevented him from pursuing a career as a software
developer and researcher at DCIEM. As an overview Arrabito explained how auditory events
can communicate information quite efficiently. Unlike visual events which are more often
continuous, -- they appear on the monitor continuously and therefore require more repeated
sampling -- auditory events are stop and go. A short burst of sound can be easily
identified with a specific event by varying one of many quantifiable facets of that sound
such as duration, frequency, pitch or timbre. Certain sounds also denote a meaning for the
user in the qualitative sense, thereby communicating information regarding the
significance of an event. For example a siren can signify a dangerous or an emergency
situation. Although Arrabito brought his monitorless computer with him he was quick to
acknowledge that this is not likely to become the norm. More attention to the aural
modality can significantly compliment existing visual modalities but is unlikely to
replace them outright. Employing this computer equipment along with several DAT playback
units he set up listening stations to demonstrate Focal Point 3D Audio's binaural
reproduction software. The 10 minute sound demonstration DAT included sound samples which
were first carefully recorded in a realistic 3 dimensional space these dichotic recordings
where then processed by the software and then reproduced for demonstration using
headphones which cover the entire ear. Dichotic sound -- ordinary stereo sound -- differs
from binaural sound in that a binaural sound can be placed spatially outside of the head.
Assuming the premise that humans can naturally hear sound in 3 dimensional space,
discerning both direction and distance, a wide range of possible research topics is
evident. As Robert Arrabito's research demonstrates, a sound can convey urgency when it
approaches the listener. Part of his demonstration included a speeding train approaching
from behind and to the left. The intimacy of a sound in a person's "sacred
space" -- in very close proximity to the body or head -- lends an immediacy to the
aural output which evokes a priority response. His demonstration also had people
whispering in your ears from the front and from behind. Binaural reproduction software has
obvious implications for Virtual Reality developers. Arrabito also indicated that his work
in developing binaural simulation software, has implications for other fields of research
which include speech intelligibility, teleconferencing and spatial orientation. To
summarize, Robert Arrabito provided a strong argument for exploring underutilized user
interface modalities such as the aural modality, to augment the strong bias that current
interfaces have toward the visual modality. Beyond the muted 2 dimensional sounds that
computer interfaces currently employ, only modestly, are clearer harmonies of 3
dimensional, binaural sounds; richer in meaning and broader in purpose.