Game Development Reference
The number of objects present as well as the precision required for selecting them
are factors that increase the complexity of the task that we wish to accomplish.
characteristics of users: level of experience, learning ability, imagination ...
On this basis, the first aim regarding manipulation in virtual environments is to let
the user accomplish two basic tasks: selection and manipulation. Selection is the des-
ignation of one or more objects in the virtual universe for future use. Manipulation is
the modification of properties of one or more objects, most often of the position and
orientation, but may include other properties (texture, transparency, size ... ). We limit
this explanation to the presentation of techniques for positioning and orientation.
12.5.2 Interaction techniques
Naturally, manual interaction has become an important means of manipulation in vir-
tual environments. The methods of interaction in virtual environments can use different
modes of action of the user: voice, direction of eyes or the movement of other body
parts, but direct manual interaction remains the most natural and effective method.
3D manipulation techniques strongly affect the quality of an interface as a whole. We
will now present interaction techniques in 3D selection and manipulation.
The most natural technique is that of the virtual hand (Bolt, 1980) used for select-
ing and moving objects. It reproduces the user's hand movements linearly and at the
same scale. This method can be extended by using a “prop'' accessory making a more
intuitive manipulation possible. One of the most prominent examples is a neurosurgery
application (Hinckley et al., 1994) wherein a doctor uses a doll's head for manipulating
a digital brain model obtained by MRI. It can also control the position of a cutting
plane with the help of a rectangular tablet. “Props'' make it possible to benefit from the
advantages of a physical object (natural setting, coherence of sensory information, tac-
tile feedback), but the application context becomes very specific to the physical object
used. For these direct interaction methods, the volume of manipulation is reduced to
the span of movement of the user's hand.
To compensate for the limitation of movements of the virtual hand, the technique
of (“ray casting'') (Bolt, 1980) makes it possible for the user to point to objects with
the help of a virtual ray, controlled by his movements. The objects can be selected
at (theoretically) any distance. This is a natural method that requires less effort. It
is however difficult to use, for selecting distant or small objects, and has limitations
in positioning (using a ray helps to fix the distance of the user from the object) and
rotation of objects (limitation in degrees of freedom: only rotations around the axis of
the ray and radially around the user are possible).
To make the selection of distant objects easier, the ray casting technique was
extended. The tool of selection is no longer a ray but a cone whose peak is at the
level of the user's hand, reminding us of a torch light beam. This technique is called
“Flash Light'' (Liang & Green, 1994). Any object included in this selection volume is
taken into account. This method facilitates the selection of distant or small objects,
but leads to ambiguities in selection of nearby objects: several objects can be in the
selection volume. Similar to ray casting, this method has limitations in positioning
and rotations. An extension of the previous technique makes it possible for the user
to modify the opening of the selection cone, in the manipulation process, to avoid