Mesh Rendering - Practical Rendering and Computation with Direct3D 11 - page 391

Figure 8.10. A sample mesh shown with and without its skeleton. Model courtesy of Radioactive Software,
LLC, www.radioactive-software.com . Created by Tomas Drinovsky, Danny Green.

To develop a model for rendering, we need to determine the number of bones that a
model will use, and how they will be connected. This is typically determined in the design
stage and can be chosen to suit the needs of a particular application. It is quite common for
humanoid models to define their bones in similar locations to where the human body has
joints between bones. The root of the bone hierarchy is typically chosen somewhere in the
lower back, where it can be used to position the entire object within world space. The cur-
rent bone transformation for each bone is the calculated by multiplying the transformation
matrix from the root bone with each successive bone's transformation matrix, until you

reach the current joint. An example mesh is shown
with and without its skeleton in Figure 8.10.
In this case, we see the skeleton oriented in
what is referred to as its bind pose. This is the base
orientation of all of the bones that a model will be
created in, and all of the vertices will have their
individual bones selected, and weights applied, in
this base orientation. After the model has been cre-
ated in this manner, the individual animations can
be created using the base information. Instead of
trying to manually place each vertex, the animator
can simply manipulate the position and orienta-
tion of the bones to create the desired pose of the
model. As an example, our sample model can be
seen in a crouching pose in Figure 8.11.

Figure 8.11. A sample model shown in
a crouching pose. Model courtesy of
Radioactive Software, LLC, www.radio-
active-software.com . Created by Tomas
Drinovsky, Danny Green.