Perceived Quality in Game Audio - Game Sound Technology and Player Interaction: Concepts and Developments - page 167

same lines, it was hypothesized that the subject's
ability to differentiate between different levels of
quality would decrease with an increase in dif-
ficulty of task/degree of interaction. The results
show that this is not generally the case. However,
when both task and main varying quality attribute
were located in the same modality, such an effect
could be observed.
More specifically, in the first experiment
(Jumisko-Pyykkö et al., 2007; Reiter and Jumisko-
Pyykkö, 2007) subjects were presented with a
scenario located in a virtual sports gym. In the
center of the gym, a loudspeaker was positioned
that played back music/speech signals with vary-
ing amounts of reverberation (time and strength).
Subjects were asked to rate the quality of reverbera-
tion under three different degrees of interaction:

possible but was not regarded as probable, given
the results of informal experiments with a simi-
lar variation in reverberation. The second, was
that the tasks (pressing a button, and navigating/
collecting objects) were not demanding enough
and that it was too easy for subjects to dedicate
part of their attention towards the quality-rating
task. This was contradicted by the claims of the
subjects themselves: a large majority claimed
to have been distracted by the navigation task.
The third possible explanation was subsequently
looked at in further experiments: The additional
cognitive load (pressing a button, navigating while
collecting objects) was located in the visual and
haptic domains, whereas the quality differences
to be rated were located in the auditory domain.
In a second round of experiments (compare
Reiter et al., 2007; Reiter, 2009), both the ad-
ditional cognitive load and the quality variations
were located in the auditory domain. A virtual room
(replica of the entrance hall of a large university
building) was equipped with a virtual loudspeaker
in the center, and subjects were asked to navigate
freely through the room using a computer mouse.
The loudspeaker played back a randomized se-
quence of numbers from 1 to 4 read out loud. The
reverberation time of the room acoustic simulation
could be adjusted between 1.0s and 3.0s in 0.5s
steps, with 2.0s considered the “reference” rever-
beration time. In the experiment, the reverberation
time was changed from reference to another value
at a single random point in time during a transition
time frame beginning 5 seconds after the start and
ending 5 seconds before the end of each 30 second
trial. A modified Degradation Category Rating
scale according to Recommendation ITU-T P.911
(1998) was used, consisting of 5 levels (much
shorter, shorter, equal, longer, much longer), to
have subjects compare the test reverberation time
with the reference reverberation time.
The additional cognitive load consisted of a
so-called n -back working memory task, similar
to what has been introduced by Kirchner (1958).
Here, subjects were asked to semantically compare

1. No interaction (watch task): subjects were
automatically moved on a pre-defined mo-
tion path through the virtual scenario
2. Limited interaction (watch and press button
task): subjects were moved on a pre-defined
motion path through the virtual scenario,
but were asked to press a button whenever
a certain object appeared within their field
of view
3. Full interaction (navigate and collect task):
subjects were asked to move freely through
the scenario by using the computer mouse
and to collect as many objects as possible
by approaching them.

Interestingly, the ability of subjects to rate
the quality of reverberation correctly did not
vary with the degree of interaction/difficulty of
the task (Friedman Χ 2 =3.3, df=2, p>0.05, ns).
Although subjects claimed to have experienced
more difficulties in the interactive tasks, this did
not show in the statistical analysis of the collected
data. Three possible explanations were looked at.
The first was that the quality differences were too
obvious, that is, the steps between the different
amounts of reverberation were too big. This is