darwin.jopenctm.compression

Class CommonAlgorithms

java.lang.Object

darwin.jopenctm.compression.CommonAlgorithms

public class CommonAlgorithms
extends Object

Author:

daniel

Constructor Summary

Constructors

Constructor and Description
CommonAlgorithms()

Method Summary

Methods

Modifier and Type	Method and Description
static float[]	calcSmoothNormals(float[] vertices, int[] indices) Calculate the smooth normals for a given mesh.
static float[]	gridIdxToPoint(Grid grid, int idx) Convert a grid index to a point (the min x/y/z for the given grid box).
static float[]	makeNormalCoordSys(float[] normals, int offset) Create an ortho-normalized coordinate system where the Z-axis is aligned with the given normal.
static float[]	restoreVertices(int[] intVertices, int[] gridIndices, Grid grid, float vertexPrecision) Calculate inverse derivatives of the vertices.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

CommonAlgorithms

public CommonAlgorithms()

Method Detail

restoreVertices

public static float[] restoreVertices(int[] intVertices,
                      int[] gridIndices,
                      Grid grid,
                      float vertexPrecision)

Calculate inverse derivatives of the vertices.

gridIdxToPoint

public static float[] gridIdxToPoint(Grid grid,
                     int idx)

Convert a grid index to a point (the min x/y/z for the given grid box).

calcSmoothNormals

public static float[] calcSmoothNormals(float[] vertices,
                        int[] indices)

Calculate the smooth normals for a given mesh. These are used as the nominal normals for normal deltas & reconstruction.

makeNormalCoordSys

public static float[] makeNormalCoordSys(float[] normals,
                         int offset)

Create an ortho-normalized coordinate system where the Z-axis is aligned with the given normal. Note 1: This function is central to how the compressed normal data is interpreted, and it can not be changed (mathematically) without making the coder/decoder incompatible with other versions of the library! Note 2: Since we do this for every single normal, this routine needs to be fast. The current implementation uses: 12 MUL, 1 DIV, 1 SQRT, ~6 ADD.