FPoseRootDataConstView

Navigation

API > API/Plugins > API/Plugins/AnimDatabase

Included here are various data structures used to store/represent/view raw pose data in a format which is appropriate for converting to and from the flat vectorized representation used by various AnimDatabase systems. Pose data itself is made up of three parts - root data, local bone transforms, and attribute data. The Pose Data structures here store pose data for multiple frames or characters, and allow operations in batch along this dimension. They also store velocities for locations, rotations, and scales. Functions and data structures are provided for performing conversion to global space using Forward Kinematics.

In most places these pose representations use single precision (one exception being the root location and global bone locations): this is because neural-network-based methods are far from accurate enough to make use of double-precision anyway.

Right now this pose representation does not have complete support for LODs and while it can be used to represent only a subset of joints, it often encodes the full skeleton. In the future proper support will be added once support for LODs is implemented in the various AnimDatabase systems that use this representation. Const View of Pose Root Data

Syntax

struct FPoseRootDataConstView  
Copy full snippet

struct FPoseRootDataConstView

Variables

Public

Functions

Public