Game Development Reference
calculating the images to be displayed. Using a head-mounted display thus requires
accurate positioning of the two screens in space as well as of these screens with respect
to the eyes. Several problems are to be identified:
Position of the sensor on the head-mounted display: First it is necessary to be able
to fix the sensor on the head-mounted display in its determined location;
Positioning the head-mounted display on the head: the display should be positioned
correctly on the user's head, i.e. we must be sure that we are displaying the images
of the correct point of view;
Location of the user's head: finally, the position and the orientation of the head-
mounted display in space must be known precisely so as to generate good images.
Is the accuracy of electromagnetic sensors sufficient? It is always possible to con-
sider using mechanical sensors, whose accuracy is much better and whose latency
time is much less. It is, however, necessary to check that the discomfort caused by
the use of mechanical sensors does not add to the discomfort created by the head-
mounted display, except when a systemmounted on a counterbalanced articulated
arm is used.
There are several sources of errors in the optical models. Some users of head-
mounted displays consider their optical systems as mere enlargers, not involving defects
of optics. These create aberrations (spherical aberration, astigmatism, distortion, etc.).
It would be preferable to take these defects into account in an optical model to be able
to correct the images using a software package. In stereoscopy, the two virtual images
must be displayed by giving a correct convergence of eyes in accordance with the depth
of the object observed. However, accommodation and convergence cannot be linked
as in reality, as the eyes have to accommodate to virtual images (read the explanations
in detail in chapter “Stereoscopic restitution of vision''). The interpupillary distance
varies from one person to the other. To avoid these errors, it is necessary to provide for
either a mechanical setting or optics with sufficiently large outlets so that anybody can
use it. The two screens can have slightly diverging optical axes and can be off-centred
to increase the visual field. In these conditions, it is difficult to determine the exact
characteristics of the visual field.
Once the problems of optics that every VR head-mounted display creates are
resolved, the problems of psychological discomfort caused by the isolation felt after
wearing a head-mounted display and the discomfort caused by its weight also need to
be tackled. Using a head-mounted display is intrusive, in the sense that it is heavy, its
movements can be more or less restricted due to the video cables, and the head's centre
of gravity is moved. Several questions thus need to be answered:
Does the weight of the head-mounted display have an effect on the head's position?
The shift in the centre of gravity of the head and display can make the user look
for a new, less tiring position;
Are the user's movements the same with or without the head-mounted display?
It is thus necessary to try to estimate the bias caused simply by wearing a head-
mounted display, even before thinking about the vision, if the posture of the user's
head is important for the application (for example, study of the visibility of a road