Skip to content

Core.Math.Geometry.Transform (Visera.Core.Math.Geometry.Transform)

3×4 affine transform type designed for GPU instancing. Memory layout matches Slang float3x4 for zero-cost CPU→GPU transfer via SSBO.

FTransform3x4F

Row-major 3×4 matrix stored as FVector4F Rows[3]. Each row is a float4; the implicit fourth row is (0, 0, 0, 1).

sizeof(FTransform3x4F) == 48, std::is_standard_layout_v == true.

Static factories

Method Parameters Description
Identity() — Returns the identity transform (rows = basis vectors + zero translation).
FromMatrix4x4(I_Matrix) const FMatrix4x4F& Extracts the top 3 rows from a column-major 4×4 matrix, transposing into row-major 3×4.
MakeTransform2D(I_Position, I_Scale, I_Rotation, I_Depth) FVector2F, FVector2F, FDegree, Float Builds a 2D affine transform: rotation around Z, non-uniform scale, translation in XY, depth in Z.
MakeTransform3D(I_Position, I_Rotation, I_Scale) const FVector3F&, const FQuaternion&, const FVector3F& Builds a 3D affine transform from position, quaternion rotation and scale.

Conversion

Method Return Description
ToMatrix4x4() FMatrix4x4F Reconstructs a column-major 4×4 matrix by appending row (0, 0, 0, 1).

Shader side

On the GPU, use import Visera.Shader.Core; to access:

  • InstanceData.transform — float3x4, bit-identical to FTransform3x4F.
  • ToMatrix4x4(float3x4 t) — helper in Visera.Shader.Core.Transform.

See also