This paper has introduced a general technique that allows to create
and to render mass scenes. It is impossible to make something like
a mathematical proof of completeness (that it can be used with all
mass scenes), but the technique is at least helpful in the process
of mass scene creation. It has been tested with many various scene
types, and it showed to be very easy to use and conductive.
It must be said that the proposed layout techniques do not consider
any interaction between the placed objects. An instance is a crowd
scene where a figure holds one hand of another figure. This small
interaction can be avoided by making not only one-character macro
objects for the layout, but also a couple should be a macro object.
There are still some scenes (like soap bubbles on the water) that
need a real interaction between the objects (bubbles) to modify the
object shape. To cover these scenes a small modification could be
made - the objects will not be really created in the layout process,
but only the reference points will be placed into some array and the
object creation macro will have the possibility to use this array, to
look around and see where and what is around. E.g. a soap bubble will
know its surroundings and according to them will modify its shape.
The future work includes the creation of additional layout parameters
according to the requirements of some special mass scenes, where the
set of current parameters would be insufficient. To achieve some
state of completeness maybe also the opportunity to create regular
layouts - multidimensional grids or rows should be added. Another
research path should be the possibility to create truly dynamic
animations.
This mass scenes rendering framework is not limited to the POV-Ray
platform. It can be easily ported to any other 3D environment, where
the adding of some functionality is possible (e.g. in 3DStudio the
possibility to make own plug-ins).