RBFSolver.h File Reference

#include "Animation/AnimTypes.h"
#include "Containers/Array.h"
#include "Containers/EnumAsByte.h"
#include "CoreMinimal.h"
#include "Curves/RichCurve.h"
#include "Math/Quat.h"
#include "Math/Rotator.h"
#include "Math/UnrealMathSSE.h"
#include "Math/Vector.h"
#include "Templates/SharedPointer.h"
#include "UObject/ObjectMacros.h"
#include "RBFSolver.generated.h"

Go to the source code of this file.

Classes
struct	FRBFEntry
struct	FRBFTarget
struct	FRBFOutputWeight
struct	FRBFParams
struct	FRBFSolver

Enumerations
enum class	ERBFSolverType : uint8 { Additive , Interpolative }
enum class	ERBFFunctionType : uint8 {   Gaussian , Exponential , Linear , Cubic ,   Quintic , DefaultFunction }
enum class	ERBFDistanceMethod : uint8 {   Euclidean , Quaternion , SwingAngle , TwistAngle ,   DefaultMethod }
enum class	ERBFNormalizeMethod : uint8 { OnlyNormalizeAboveOne , AlwaysNormalize , NormalizeWithinMedian , NoNormalization }

Variables
struct ANIMGRAPHRUNTIME_API	FRBFSolverData

Enumeration Type Documentation

ERBFDistanceMethod

enum class ERBFDistanceMethod : uint8

strong

Method for determining distance from input to targets

Enumerator
Euclidean	Standard n-dimensional distance measure
Quaternion	Treat inputs as quaternion
SwingAngle	Treat inputs as quaternion, and find distance between rotated TwistAxis direction
TwistAngle	Treat inputs as quaternion, and find distance between rotations around the TwistAxis direction
DefaultMethod	Uses the setting of the parent container

ERBFFunctionType

enum class ERBFFunctionType : uint8

strong

Function to use for each target falloff

Enumerator
Gaussian
Exponential
Linear
Cubic
Quintic
DefaultFunction	Uses the setting of the parent container

ERBFNormalizeMethod

enum class ERBFNormalizeMethod : uint8

strong

Method to normalize weights

Enumerator
OnlyNormalizeAboveOne	Only normalize above one
AlwaysNormalize	Always normalize. Zero distribution weights stay zero.
NormalizeWithinMedian	Normalize only within reference median. The median is a cone with a minimum and maximum angle within which the value will be interpolated between non-normalized and normalized. This helps to define the volume in which normalization is always required.
NoNormalization	Don't normalize at all. This should only be used with the interpolative method, if it is known that all input values will be within the area bounded by the targets.

ERBFSolverType

enum class ERBFSolverType : uint8

strong

The solver type to use. The two solvers have different requirements.

Enumerator
Additive	The additive solver sums up contributions from each target. It's faster but may require more targets for a good coverage, and requires the normalization step to be performed for smooth results.
Interpolative	The interpolative solver interpolates the values from each target based on distance. As long as the input values are within the area bounded by the targets, the interpolation is well-behaved and return weight values within the 0% - 100% limit with no normalization required. Interpolation also gives smoother results, with fewer targets, than additive but at a higher computational cost.

Variable Documentation

FRBFSolverData

struct ANIMGRAPHRUNTIME_API FRBFSolverData