Information Technology Reference
In-Depth Information
We have thus developed a new mixed approach for the sonification of a
hierarchical menu. This model links the ability of auditory icons to transmit meaning
directly with the efficiency of earcons for the representation of hierarchical relations
between items. Each of the six main level menus is therefore represented by an
auditory icon, the choice of which is based on different criteria: quality of sound,
relevance of the icon in relation to the represented notion and homogeneity of
auditory icons. The construction of earcons is then based on auditory synthesis by a
physical model, the parameters of which map the hierarchical position of each item.
Thus, each item in the menu is associated with a sound that comprises two types of
information: semantic (metaphorically with the auditory icon) and hierarchical
(symbolically with the earcon). Two instances of this new model were put forward
by the sound designer integrated in the project: a “hybrid” solution, the originality of
which resides in the manner in which the auditory icon and the earcon are
fused
;
and
a “museme” solution for which the semantic information of the item is ensured by
an orchestrated version of the auditory icon.
A model of the auditory HMI was then created to evaluate the ergonomics of this
new model and compare it with sonification based only on vocal synthesis. The
results of the test enable the following conclusions to be drawn:
- The model does not enable us to carry out navigation tasks more rapidly. This
result is in line with the results of Leplâtre and Brewster [LEP 00] regarding the
menu of a cell phone. However, it seems to us that in the case of automobile driving
the main stake consists of navigating menus without looking at the screen, even if
that means a slower navigation.
- A large majority of users deem that the model enables safer driving and better
visual attention to the driving scene. This result is obtained subjectively, and it will
now be relevant to implement objective methods in order to quantitatively evaluate
the efficiency of the model. Among the methods generally used to evaluate attention
in automobile driving, we can quote eye tracking, for example [SOD 02] or
measuring on a driving simulator (e.g. measure of the reaction time to braking
[SUM 00]).
- In the long term, users prefer the hybrid to the museme instance, as the
proposed sounds are judged to be more evocative.
- The model obtains better results than vocal synthesis for the acceptance of
sounds in the long term.
