Overview

CC3SkinMeshNode is a CC3MeshNode specialized to use vertex skinning to draw the contents of its mesh. It is one of the key structural descendant nodes of a CC3SoftBodyNode instance.

The CC3SkinMeshNode contains a collection of skin sections, in the form of CC3SkinSection instances. Each CC3SkinSection instance relates a section of the mesh, in the form of a range of vertices, to a set of bones in the skeleton.

Each CC3SkinSection applies the transformations in the referenced bones to the the vertices in the section of the mesh that it controls, and draws that section of the mesh by drawing the vertices within its range in a single GL call.

After copying a CC3SkinMeshNode, the newly created copy will still be influenced by the original skeleton. The result is that both the original mesh and the copy will move and be deformed in tandem as the skeleton moves.

If you are creating a chorus line of dancing characters, this may be the effect you are after. However, if you are creating a squadron of similar, but independently moving, characters, each CC3SkinMeshNode copy should be controlled by a separate skeleton. To do this, after creating a copy of the skeleton bone node assembly as well, you can use the reattachBonesFrom: method to attach the skin mesh node to the new skeleton.

When copying a CC3SkinMeshNode as part of copying a CC3SoftBodyNode instance, a copy of the skeleton is also created, and the reattachBonesFrom: method is automatically invoked. When copying CC3SoftBodyNode, you do not need to invoke the reattachBonesFrom: method on the new CC3SkinMeshNode directly.

The use of bounding volumes with skinned meshes can be tricky, because the locations of the vertices are affected both by the location of the mesh node, as with any mesh, but also by the location of the bones. In addition, bone transformation is handled by the GPU, and unless the CPU is also tasked with transforming each vertex, it is difficult for the app to know the true range of the vertices.

Because of this, the createBoundingVolumes method will be ignored by instances of this class, and a bounding volume will not automatically be assigned to this node, to ensure that the mesh will not be culled when it shouldn’t if the automatic bounding volume is not the correct shape. This mesh will therefore be drawn for each frame, even if it is not in front of the camera (ie- inside the camera’s frustum).

It is left to the application to determine the best approach to managing the assignment of a bounding volume, possibly using one of the following approaches:

• You can choose to leave this node with no bounding volume, and allow it to be drawn on each frame. This may be the easiest approach if performance is not critical.

• For a CC3SkinMeshNode, the bounding volume that is created automatically when the createBoundingVolume method is invoked forms a sphere around the rest pose of the mesh. A spherical bounding volume is used because, for many bone-rigged characters, the bones remain within a sphere determmined by the rest pose. If you know that the vertices of the skinned mesh node will not move beyond the static bounding volume defined by the vertices in the rest pose, you can invoke the createBoundingVolume method to have a spherical bounding volume created automatically from the rest pose of this node. If this is a common requirement, you can also use the createSkinnedBoundingVolumes method on any ancestor node to have a spherical bounding volume automatically created for each descendant skinned mesh node.

• You can manually create a bounding volume of the right size and shape for the movement of the vertices from the perspective of a root bone of the skeleton. Assign the bounding volume to the root bone by using the boundingVolume property on the root bone and, once it has been assigned a root bone of the skeleton, use the setSkeletalBoundingVolume: method on an ancestor node of all of the CC3SkinMeshNodes that are to use that bounding volume, to assign that bounding volume to all of the appropriate CC3SkinMeshNodes. A good choice to target for the invocation of this method might be the CC3SoftBodyNode of the model, or even the CC3ResourceNode above it, if loaded from a file. During development, you can use the shouldDrawBoundingVolume property to make the bounding volume visible, to aid in determining and setting the right size and shape for it.