Information Technology Reference
In-Depth Information
second case to the notion of
just-noticeable difference
(JND), which corresponds to
the smallest detectable variation in the considered unit,
-
confusion scaling method
, which consists of varying a physical unit and
measuring the variation of the corresponding differential threshold;
- the direct methods:
-
ratio scaling
,
which is based on the ability to carry out a numerical judgment
in relation to a difference in the level of perceived sensation. This method can be
implemented, either by
unit estimation
by associating a number proportional to the
perceived sensation or by
unit production
by directly adjusting the perceived
sensation level for a sound to a number,
-
cross modal matching
, which consists of linking a modality - e.g. auditory -
via the means of another modality - e.g. haptic (for instance, evaluation of the
auditory sensation of the sound level by the production of a sensation of muscular
force). This method can be used in the case of situations that evolve over time.
8.5.1.3.
Exploratory methods
When the auditory attributes are not easy to define or identify (conversely, for
example, to auditory intensity or duration) or are not known before, exploratory and
generally multidimensional methods are used.
In this domain, more current methods have also been listed by Susini in
[SUS 11]:
- The
semantic differential
method
, which consists of evaluating sounds studied
on a semantic scale marked by two opposed verbal descriptions (e.g.
“pleasant/unpleasant” or, in a slightly different way, “bright/non bright”). The
rigorous implementation of this method requires a few pre-requisites, in particular
the previous validation of the words used (also called semantic descriptors) in terms
of relevance in relation to the sensation studied but also in terms of common
understanding of their meaning. This method results in the description of a semantic
type profile that can also be put in relation to the acoustic sound descriptors.
- The method of
dissimilarity judgment
, which consists of evaluating the
similarity/dissimilarity that can be perceived between two sounds on a continuous
numerical scale, by repeating the process for combinations of pairs that come from a
given corpus. This method
in theory
enables us to reveal the perceptive dimensions
that underlie the perception of the sounds of the corpus and is therefore based on a
strong perceptive continuum hypothesis along these dimensions. Thus, this method
possesses a certain number of prerequisites including, among others, certain
homogeneity of stimuli that make up the corpus. Nonetheless, the associated
multi-
dimensional analysis
enables us to construct the perceptive (metric) space describing
