package net.coderbot.iris.vertices; import org.joml.Vector3f; /** * Provides some utilities for working with packed normal vectors. Each normal component provides 8 bits of * precision in the range of [-1.0,1.0]. * Copied from Sodium, licensed under the LGPLv3. Modified to support a W component. * * | 32 | 24 | 16 | 8 | * | 0000 0000 | 0110 1100 | 0110 1100 | 0110 1100 | * | W | X | Y | Z | */ public class NormI8 { private static final int X_COMPONENT_OFFSET = 0; private static final int Y_COMPONENT_OFFSET = 8; private static final int Z_COMPONENT_OFFSET = 16; private static final int W_COMPONENT_OFFSET = 24; /** * The maximum value of a normal's vector component. */ private static final float COMPONENT_RANGE = 127.0f; /** * Constant value which can be multiplied with a floating-point vector component to get the normalized value. The * multiplication is slightly faster than a floating point division, and this code is a hot path which justifies it. */ private static final float NORM = 1.0f / COMPONENT_RANGE; public static int pack(Vector3f normal) { return pack(normal.x(), normal.y(), normal.z(), 0); } /** * Packs the specified vector components into a 32-bit integer in XYZ ordering with the 8 bits of padding at the * end. * @param x The x component of the normal's vector * @param y The y component of the normal's vector * @param z The z component of the normal's vector */ public static int pack(float x, float y, float z, float w) { return ((int) (x * 127) & 0xFF) | (((int) (y * 127) & 0xFF) << 8) | (((int) (z * 127) & 0xFF) << 16) | (((int) (w * 127) & 0xFF) << 24); } /** * Packs the specified vector components into a 32-bit integer in XYZ ordering with the 8 bits of padding at the * end. * @param x The x component of the normal's vector * @param y The y component of the normal's vector * @param z The z component of the normal's vector */ public static int packColor(float x, float y, float z, float w) { return ((int) (x * 127) & 0xFF) | (((int) (y * 127) & 0xFF) << 8) | (((int) (z * 127) & 0xFF) << 16) | (((int) w & 0xFF) << 24); } /** * Unpacks the x-component of the packed normal, denormalizing it to a float in the range of -1.0..1.0. * @param norm The packed normal */ public static float unpackX(int norm) { return ((byte) ((norm >> X_COMPONENT_OFFSET) & 0xFF)) * NORM; } /** * Unpacks the y-component of the packed normal, denormalizing it to a float in the range of -1.0..1.0. * @param norm The packed normal */ public static float unpackY(int norm) { return ((byte) ((norm >> Y_COMPONENT_OFFSET) & 0xFF)) * NORM; } /** * Unpacks the z-component of the packed normal, denormalizing it to a float in the range of -1.0..1.0. * @param norm The packed normal */ public static float unpackZ(int norm) { return ((byte) ((norm >> Z_COMPONENT_OFFSET) & 0xFF)) * NORM; } /** * Unpacks the w-component of the packed normal, denormalizing it to a float in the range of -1.0..1.0. * @param norm The packed normal */ public static float unpackW(int norm) { return ((byte) ((norm >> W_COMPONENT_OFFSET) & 0xFF)) * NORM; } }