Game Development Reference
takes place. Thus, the role of stereoscopy certainly depends on the activity of the
subject. Is he trying to gauge the distance from an object in his peri-corporal space
(for example to grab it, which presents other problems in terms of sensory feedback
and motor interfaces) or gauge the distance from a vehicle in a simulated driving task?
In the first case, stereoscopy is certainly necessary (according to what is known about
the distance from which stereoscopic vision is considered useless). In the second case,
stereoscopy seems useless. However, the role of the “foreground/background'' relations
in the organisation of perceptual space (see chapter 13 on stereoscopic restitution) still
needs to be understood better.
As regards the spatial resolution required for a virtual reality system, we know
that we are wide of the mark with the present technologies, if we compare the graphic
rendering performances with the tiling of the photoreceptors of the human retina.
As for temporal resolution, the question is much more complex than it seems.
There is an implicit rule in virtual reality, which states that 30 images per second is an
adequate image rate. This is certainly false if we consider that the view of a continuous
movement is a basic functionality of our perceptual visual system, which is required
to control our spatial activities. Then, one of the problems presented is generation
(and perception) of quick movements. For example, if we refer to the classic data of
literature, we can assess that the view of a continuous movement (in the case of digital
stimulations) is lost when the angular deviation between two successive positions of
a moving object is more than 15-20 minutes of arc (Nakayama, 1985). This means
that, for an image rate of 30Hz, the maximum object speed that will be perceived as
a continuous movement is less than 10 degrees per second!
Possibly, it is necessary to put this into perspective while noting that this parameter
runs the risk of varying with the complexity of the visual scene.
Quite obviously, a lot of research is required before answering the question: What
is the correct parameter? Here, we can present a few questions from amongst the
numerous questions that thus seem related to the basic research.
It would certainly be futile to try to build a virtual reality system presenting “all''
the information in “absolute'' real time Firstly, it is technically impossible due to the
technical limitations of interfacing, and also as the computer systems have to process
and restore a considerable bulk of information in real time. What is more important is
the fact that we do not know what “all'' is, and that the chief concern of Behavioural
Sciences is to understand what information is processed, how it is processed and how
the sensorimotor control works, including in its temporal aspects. For example, at the
level of processing of visual information, it seems clear that we cannot do without
taking into account the original cognitive mechanisms of the regulation of perceptual
activity. In particular, we can cite the role of a past experience which tells us that,
while the initial perception of a complex or ambiguous scene can take a lot of time,
recognition of the same scene will be immediate during new presentations.
We can also cite numerous works that demonstrate the role of attention in percep-
tual processing, possibly in the plasticity of the aspects of the perceptual processing
mechanisms, which we could consider to be biologically stable (for example, the size
of the receptive fields in the cortical visual system).