Abstract
Sound families comprised of either musical
sounds (abstract musical signals),
communications sounds (based on traditional
electronic signals), or real world sounds (based
on naturally occurring, nonmusical events) were
designed to provide auditory feedback in a
graphical user interface. Typical business users
mapped the sounds to functions and rated their
confidence in the functional mapping. The
sounds which mapped well to functions were
then tested within the context of a graphical user
interface business communications simulation.
Users rated the pleasantness and appropriateness
of each sound and selected the best sound for
each function. In general, real world sounds
mapped most reliably to the functions. However,
users consistently preferred musical sounds to the
real world sounds. Of the eleven feedback
signals selected for specific functions four were
communications based and seven were musical.
None of the real world auditory signals were
acceptable.
Keywords:
Auditory Feedback, Auditory
Icons, Earcons, Sound.
Introduction
Real world signals, also known as auditory icons,
are based on naturally occurring, non-musical
events[3]. The use of real world sounds as
auditory feedback in graphical user interfaces has
been advocated due to the functional mapping
inherent in these sounds. This type of auditory
feedback allows users to use pre-existing
knowledge of the world to generalize to the
auditory metaphor of the user interface[1,2,3].
One issue regarding the implementation of real
world sounds concerns users' reactions to
listening to them repeatedly over time in an
actual business application. It has been
speculated that users may perceive the real world
sounds as annoying and trite. More abstract
sounds may be less intrusive; providing
appropriate feedback without being distracting to
the user or others.
Abstract sounds that may provide a successful
alternative include communications and musical
sounds. Communications oriented sounds are
musical tones traditionally associated with
electronic tasks. Musically oriented sounds are
abstract signals and there is generally no
association between musical sounds and a given
task or function.
A professional musician designed both
communications and musical feedback signals.
Real world signals were selected from sound
libraries. Experiment I investigated how well
these three families of feedback signals (i.e.,
musical, communications and real world)
mapped to eleven different communications
functions. Experiment II evaluated the sounds
based on subjective ratings.
EXPERIMENT I
The experiment was designed to find any natural
mapping of the sounds to the functions. The
sounds were also evaluated for their overall
pleasantness, and their appropriateness for a
business environment. Thirty-eight computer
literate office workers participated as paid
subjects. The subjects were tested in four groups
of approximately 10 per group. The 48 sounds
evaluated consisted of 22 musical sounds, 14
communications oriented sounds and 12 real
world sounds. The 12 real world sounds were
selected from a proposed set of 30 real world
sounds. Due to technical limitations of computer
audio output devices, the real world sounds were
limited to those that provided an adequate
auditory representation of the real world referent.
The subjects were familiarized with the 11
communications functions that require auditory
feedback. The functions included ringing,
confirmation and error signals, among others.
Each group was presented the 48 sounds in a
different random order. After each sound the
subjects indicated which of the 11 functions the
sound best mapped to, and then rated the sound
for its overall pleasantness, appropriateness
given the function chosen, and the confidence
they had in selecting that sound for that function.
After a intermediate task, the groups were
presented the sounds again in a different
randomized order and made the same mapping
and rating judgments.
The data were analyzed to determine which
sounds mapped most often to the functions. For
each function, we derived three sounds that
mapped best together with the highest
pleasantness and appropriateness ratings.
Overall, the real world sounds mapped most
predictably to the functions. The
communications sounds mapped to the functions
with reasonable consistency. The musical sounds
mapped least consistently to the functions, but
had the highest ratings for pleasantness and
appropriateness. Eleven real world sounds, 7
communications sounds and 15 musical sounds
were selected for further evaluation in the second
experiment.
EXPERIMENT II
The subset of sounds selected in Experiment I
was based primarily on functional mapping.
Experiment II was designed to test the sounds
within the context of a simulated business
communications application and collect
subjective ratings. Twenty-two computer literate
office workers participated as paid subjects. The
subjects were tested in groups of two to four.
The subjects were shown simulations that gave a
context and scenario for each of the functions.
Each group heard the three sounds within each
scenario in a different random order. After
hearing each sound, the subjects rated its
pleasantness and appropriateness. The subjects
then heard each sound in the scenario again
before selecting the sound that they would most
prefer for that function.
Eleven sounds, one for each function were
subsequently selected based on the rating data.
Seven of the sounds were musical sounds and
four were communications sounds. None of the
real world sounds were selected. The real world
sounds were consistently rated poorly on
pleasantness and appropriateness and were
seldom chosen as the best sound for any of the
functions. The musical sounds had reliably high
pleasantness and appropriateness ratings and
were most frequently selected as first choice.
Communications sounds varied in subjective
ratings and occasionally proved to be the best
selection for a given function.
DISCUSSION
In Experiment I the functional mapping of real
world sounds was very predictable. Subjects
generally selected the function for which the
sound had been intended. The communications
sounds, whose abstract musical tones were based
on learned pairings, had less consistent
functional mappings. The musical sounds, which
had no previous learned association, were poorly
mapped to the functions. In Experiment II, the
subjective ratings of the sounds were analyzed
within the context of a business application. The
more musically oriented sounds were most highly
preferred. Clearly, the type of auditory feedback
most preferred by subjects is not the same as
those sounds that provided the best functional
mapping. Rather, the users' reactions suggests
that real world signals are less appropriate for
actual applications relative to abstract musical
feedback.
ACKNOWLEDGEMENTS
The authors would like to thank Mark Barasch of
Mark Barasch Music Productions LTD for his
creative musical contributions and Dan Teibel
for his technical input and support.
References
1. Blattner, M., Sumikawa, D. & Greenberg, R.
(1989). Earcons and icons: Their structure and
common design principles. Human-Computer
Interaction, 4, 11-44.
2. Gaver, W. (1986). Auditory icons: Using
sound in computer interfaces. Human-Computer
Interaction, 2, 167-177.
3. Mynatt, E. Designing with Auditory Icons:
How Well do We Identify Auditory Cues?, in
Proc. CHI'94 Human Factors in Computing
Systems (Boston, April 24-28, 1994), ACM
Press, pp. 269-270.
