bool raycast(game* g, vec2 from, vec2 to) {
ivec2 cur(fast_floor(from.x), fast_floor(from.y));
ivec2 end(fast_floor(to.x), fast_floor(to.y));
ivec2 sign(sign(to.x - from.x), sign(to.y - from.y));
vec2 abs_delta(abs(to.x - from.x), abs(to.y - from.y));
vec2 delta_t(vec2(1.0f) / abs_delta);
vec2 lb(to_vec2(cur));
vec2 ub(lb + vec2(1.0f));
vec2 t(vec2((from.x > to.x) ? (from.x - lb.x) : (ub.x - from.x), (from.y > to.y) ? (from.y - lb.y) : (ub.y - from.y)) / abs_delta);
for(;;) {
if (g->is_raycast_solid(cur.x, cur.y))
return true;
if (t.x <= t.y) {
if (cur.x == end.x) break;
t.x += delta_t.x;
cur.x += sign.x;
}
else {
if (cur.y == end.y) break;
t.y += delta_t.y;
cur.y += sign.y;
}
}
return false;
}
chunk_pos&
chunk_pos::operator= (const entity_pos &other)
{
this->x = fast_floor (other.x) / 16;
this->z = fast_floor (other.z) / 16;
return *this;
}
block_pos&
block_pos::operator= (const entity_pos &other)
{
this->x = fast_floor (other.x);
this->y = fast_floor (other.y);
this->z = fast_floor (other.z);
//if (this->y < 0) this->y = 0;
//if (this->y > 255) this->y = 255;
return *this;
}
float raw_noise_2d(float x, float y, const uint8_t* perm)
{
const float F2 = 0.5f * (std::sqrt(3.0f) - 1.0f);
float s = (x + y) * F2;
int32_t i = fast_floor(x + s);
int32_t j = fast_floor(y + s);
const float G2 = (3.0f - std::sqrt(3.0f)) / 6.0f;
float t = (i + j) * G2;
float x0 = x - (i - t);
float y0 = y - (j - t);
uint8_t i1 = 0;
uint8_t j1 = 1;
if(x0 > y0)
{
i1 = 1;
j1 = 0;
}
float x1 = x0 - i1 + G2;
float y1 = y0 - j1 + G2;
float x2 = x0 - 1.0f + 2.0f * G2;
float y2 = y0 - 1.0f + 2.0f * G2;
uint8_t ii = i & 255;
uint8_t jj = j & 255;
uint8_t gi0 = perm[ii + perm[jj ] ] % 12;
uint8_t gi1 = perm[ii + perm[jj + j1] + i1] % 12;
uint8_t gi2 = perm[ii + perm[jj + 1 ] + 1 ] % 12;
float t0 = 0.5f - x0 * x0 - y0 * y0;
float t1 = 0.5f - x1 * x1 - y1 * y1;
float t2 = 0.5f - x2 * x2 - y2 * y2;
float n0 = (t0 < 0) ? 0.0f : t0*t0*t0*t0 * dot(grad3[gi0], x0, y0);
float n1 = (t1 < 0) ? 0.0f : t1*t1*t1*t1 * dot(grad3[gi1], x1, y1);
float n2 = (t2 < 0) ? 0.0f : t2*t2*t2*t2 * dot(grad3[gi2], x2, y2);
return 70.0f * (n0 + n1 + n2);
}
// Classic Perlin noise, 3D version
float base(float x, float y)
{
x *= xscale; //replace with multiplication
y *= xscale;
//get grid point
int X = fast_floor(x);
int Y = fast_floor(y);
x = x - X;
y = y - Y;
int gi00 = get_gradient(X+0,Y+0);
int gi01 = get_gradient(X+0,Y+1);
int gi10 = get_gradient(X+1,Y+0);
int gi11 = get_gradient(X+1,Y+1);
// Calculate noise contributions from each of the eight corners
float n00= dot(grad+2*gi00, x, y);
float n10= dot(grad+2*gi10, x-1, y);
float n01= dot(grad+2*gi01, x, y-1);
float n11= dot(grad+2*gi11, x-1, y-1);
// Compute the fade curve value for each of x, y, z
#if 1
float u = fade(x);
float v = fade(y);
#else
float u = x;
float v = y;
#endif
// Interpolate along x the contributions from each of the corners
float nx00 = mix(n00, n10, u);
float nx10 = mix(n01, n11, u);
// Interpolate the four results along y
float nxy = mix(nx00, nx10, v);
if (nxy < -1 || nxy > 1 ) printf("Error: noise %f \n", nxy);
//return nxy / 0.707106781f; //-1 to 1
return nxy; //-1 to 1
}
aabb2 cast_aabb(game* g, aabb2 self, vec2 delta, int* clipped) {
aabb2 bounds(expand_toward(self, delta));
int x0 = fast_floor(bounds.min.x);
int y0 = fast_floor(bounds.min.y);
int x1 = fast_floor(bounds.max.x);
int y1 = fast_floor(bounds.max.y);
vec2 clipped_delta = delta;
for(int x = x0; x <= x1; x++) {
for(int y = y0; y <= y1; y++) {
if (g->get_tile(x, y) != TT_EMPTY) {
clipped_delta.x = clip_x(aabb2(vec2((float)x, (float)y), vec2((float)(x + 1), (float)(y + 1))), self, clipped_delta.x);
}
}
}
self.min.x += clipped_delta.x;
self.max.x += clipped_delta.x;
for(int x = x0; x <= x1; x++) {
for(int y = y0; y <= y1; y++) {
if (g->get_tile(x, y) != TT_EMPTY) {
clipped_delta.y = clip_y(aabb2(vec2((float)x, (float)y), vec2((float)(x + 1), (float)(y + 1))), self, clipped_delta.y);
}
}
}
self.min.y += clipped_delta.y;
self.max.y += clipped_delta.y;
if (clipped) {
*clipped = 0;
if (delta.x != clipped_delta.x) *clipped |= (delta.x > 0.0f) ? CLIPPED_XP : CLIPPED_XN;
if (delta.y != clipped_delta.y) *clipped |= (delta.y > 0.0f) ? CLIPPED_YP : CLIPPED_YN;
}
return self;
}
/** @brief MATLAB Driver.
**/
void
mexFunction(int nout, mxArray *out[],
int nin, const mxArray *in[])
{
int M,N,S=0,smin=0,K,num_levels=0 ;
const int* dimensions ;
const double* P_pt ;
const double* G_pt ;
float* descr_pt ;
float* buffer_pt ;
float sigma0 ;
float magnif = 3.0f ; /* Spatial bin extension factor. */
int NBP = 4 ; /* Number of bins for one spatial direction (even). */
int NBO = 8 ; /* Number of bins for the ortientation. */
int mode = NOSCALESPACE ;
int buffer_size=0;
enum {IN_G=0,IN_P,IN_SIGMA0,IN_S,IN_SMIN} ;
enum {OUT_L=0} ;
/* ------------------------------------------------------------------
** Check the arguments
** --------------------------------------------------------------- */
if (nin < 3) {
mexErrMsgTxt("At least three arguments are required") ;
} else if (nout > 1) {
mexErrMsgTxt("Too many output arguments.");
}
if( !uIsRealScalar(in[IN_SIGMA0]) ) {
mexErrMsgTxt("SIGMA0 should be a real scalar") ;
}
if(!mxIsDouble(in[IN_G]) ||
mxGetNumberOfDimensions(in[IN_G]) > 3) {
mexErrMsgTxt("G should be a real matrix or 3-D array") ;
}
sigma0 = (float) *mxGetPr(in[IN_SIGMA0]) ;
dimensions = mxGetDimensions(in[IN_G]) ;
M = dimensions[0] ;
N = dimensions[1] ;
G_pt = mxGetPr(in[IN_G]) ;
P_pt = mxGetPr(in[IN_P]) ;
K = mxGetN(in[IN_P]) ;
if( !uIsRealMatrix(in[IN_P],-1,-1)) {
mexErrMsgTxt("P should be a real matrix") ;
}
if ( mxGetM(in[IN_P]) == 4) {
/* Standard (scale space) mode */
mode = SCALESPACE ;
num_levels = dimensions[2] ;
if(nin < 5) {
mexErrMsgTxt("Five arguments are required in standard mode") ;
}
if( !uIsRealScalar(in[IN_S]) ) {
mexErrMsgTxt("S should be a real scalar") ;
}
if( !uIsRealScalar(in[IN_SMIN]) ) {
mexErrMsgTxt("SMIN should be a real scalar") ;
}
if( !uIsRealMatrix(in[IN_P],4,-1)) {
mexErrMsgTxt("When the e mode P should be a 4xK matrix.") ;
}
S = (int)(*mxGetPr(in[IN_S])) ;
smin = (int)(*mxGetPr(in[IN_SMIN])) ;
} else if ( mxGetM(in[IN_P]) == 3 ) {
mode = NOSCALESPACE ;
num_levels = 1 ;
S = 1 ;
smin = 0 ;
} else {
mexErrMsgTxt("P should be either a 3xK or a 4xK matrix.") ;
}
/* Parse the property-value pairs */
{
char str [80] ;
int arg = (mode == SCALESPACE) ? IN_SMIN + 1 : IN_SIGMA0 + 1 ;
while(arg < nin) {
int k ;
if( !uIsString(in[arg],-1) ) {
mexErrMsgTxt("The first argument in a property-value pair"
" should be a string") ;
}
mxGetString(in[arg], str, 80) ;
#ifdef WINDOWS
for(k = 0 ; properties[k] && strcmpi(str, properties[k]) ; ++k) ;
#else
for(k = 0 ; properties[k] && strcasecmp(str, properties[k]) ; ++k) ;
#endif
switch (k) {
case PROP_NBP:
if( !uIsRealScalar(in[arg+1]) ) {
mexErrMsgTxt("'NumSpatialBins' should be real scalar") ;
}
NBP = (int) *mxGetPr(in[arg+1]) ;
if( NBP <= 0 || (NBP & 0x1) ) {
mexErrMsgTxt("'NumSpatialBins' must be positive and even") ;
}
break ;
case PROP_NBO:
if( !uIsRealScalar(in[arg+1]) ) {
mexErrMsgTxt("'NumOrientBins' should be a real scalar") ;
}
NBO = (int) *mxGetPr(in[arg+1]) ;
if( NBO <= 0 ) {
mexErrMsgTxt("'NumOrientlBins' must be positive") ;
}
break ;
case PROP_MAGNIF:
if( !uIsRealScalar(in[arg+1]) ) {
mexErrMsgTxt("'Magnif' should be a real scalar") ;
}
magnif = (float) *mxGetPr(in[arg+1]) ;
if( magnif <= 0 ) {
mexErrMsgTxt("'Magnif' must be positive") ;
}
break ;
case PROP_UNKNOWN:
mexErrMsgTxt("Property unknown.") ;
break ;
}
arg += 2 ;
}
}
/* -----------------------------------------------------------------
* Pre-compute gradient and angles
* -------------------------------------------------------------- */
/* Alloc two buffers and make sure their size is multiple of 128 for
* better alignment (used also by the Altivec code below.)
*/
buffer_size = (M*N*num_levels + 0x7f) & (~ 0x7f) ;
buffer_pt = (float*) mxMalloc( sizeof(float) * 2 * buffer_size ) ;
descr_pt = (float*) mxCalloc( NBP*NBP*NBO*K, sizeof(float) ) ;
{
/* Offsets to move in the scale space. */
const int yo = 1 ;
const int xo = M ;
const int so = M*N ;
int x,y,s ;
#define at(x,y) (*(pt + (x)*xo + (y)*yo))
/* Compute the gradient */
for(s = 0 ; s < num_levels ; ++s) {
const double* pt = G_pt + s*so ;
for(x = 1 ; x < N-1 ; ++x) {
for(y = 1 ; y < M-1 ; ++y) {
float Dx = 0.5 * ( at(x+1,y) - at(x-1,y) ) ;
float Dy = 0.5 * ( at(x,y+1) - at(x,y-1) ) ;
buffer_pt[(x*xo+y*yo+s*so) + 0 ] = Dx ;
buffer_pt[(x*xo+y*yo+s*so) + buffer_size] = Dy ;
}
}
}
/* Compute angles and modules */
{
float* pt = buffer_pt ;
int j = 0 ;
while (j < N*M*num_levels) {
#if defined( MACOSX ) && defined( __ALTIVEC__ )
if( ((unsigned int)pt & 0x7) == 0 && j+3 < N*M*num_levels ) {
/* If aligned to 128 bit and there are at least 4 pixels left */
float4 a, b, c, d ;
a.vec = vec_ld(0,(vector float*)(pt )) ;
b.vec = vec_ld(0,(vector float*)(pt + buffer_size)) ;
c.vec = vatan2f(b.vec,a.vec) ;
a.x[0] = a.x[0]*a.x[0]+b.x[0]*b.x[0] ;
a.x[1] = a.x[1]*a.x[1]+b.x[1]*b.x[1] ;
a.x[2] = a.x[2]*a.x[2]+b.x[2]*b.x[2] ;
a.x[3] = a.x[3]*a.x[3]+b.x[3]*b.x[3] ;
d.vec = vsqrtf(a.vec) ;
vec_st(c.vec,0,(vector float*)(pt + buffer_size)) ;
vec_st(d.vec,0,(vector float*)(pt )) ;
j += 4 ;
pt += 4 ;
} else {
#endif
float Dx = *(pt ) ;
float Dy = *(pt + buffer_size) ;
*(pt ) = sqrtf(Dx*Dx + Dy*Dy) ;
if (*pt > 0)
*(pt + buffer_size) = atan2f(Dy, Dx) ;
else
*(pt + buffer_size) = 0 ;
j += 1 ;
pt += 1 ;
#if defined( MACOSX ) && defined( __ALTIVEC__ )
}
#endif
}
}
}
/* -----------------------------------------------------------------
* Do the job
* -------------------------------------------------------------- */
if(K > 0) {
int p ;
/* Offsets to move in the buffer */
const int yo = 1 ;
const int xo = M ;
const int so = M*N ;
/* Offsets to move in the descriptor. */
/* Use Lowe's convention. */
const int binto = 1 ;
const int binyo = NBO * NBP ;
const int binxo = NBO ;
const int bino = NBO * NBP * NBP ;
for(p = 0 ; p < K ; ++p, descr_pt += bino) {
/* The SIFT descriptor is a three dimensional histogram of the position
* and orientation of the gradient. There are NBP bins for each spatial
* dimesions and NBO bins for the orientation dimesion, for a total of
* NBP x NBP x NBO bins.
*
* The support of each spatial bin has an extension of SBP = 3sigma
* pixels, where sigma is the scale of the keypoint. Thus all the bins
* together have a support SBP x NBP pixels wide . Since weighting and
* interpolation of pixel is used, another half bin is needed at both
* ends of the extension. Therefore, we need a square window of SBP x
* (NBP + 1) pixels. Finally, since the patch can be arbitrarly rotated,
* we need to consider a window 2W += sqrt(2) x SBP x (NBP + 1) pixels
* wide.
*/
const float x = (float) *P_pt++ ;
const float y = (float) *P_pt++ ;
const float s = (float) (mode == SCALESPACE) ? (*P_pt++) : 0.0 ;
const float theta0 = (float) *P_pt++ ;
const float st0 = sinf(theta0) ;
const float ct0 = cosf(theta0) ;
const int xi = (int) floor(x+0.5) ; /* Round-off */
const int yi = (int) floor(y+0.5) ;
const int si = (int) floor(s+0.5) - smin ;
const float sigma = sigma0 * powf(2, s / S) ;
const float SBP = magnif * sigma ;
const int W = (int) floor( sqrt(2.0) * SBP * (NBP + 1) / 2.0 + 0.5) ;
int bin ;
int dxi ;
int dyi ;
const float* pt ;
float* dpt ;
/* Check that keypoints are within bounds . */
if(xi < 0 ||
xi > N-1 ||
yi < 0 ||
yi > M-1 ||
((mode==SCALESPACE) &&
(si < 0 ||
si > dimensions[2]-1) ) )
continue ;
/* Center the scale space and the descriptor on the current keypoint.
* Note that dpt is pointing to the bin of center (SBP/2,SBP/2,0).
*/
pt = buffer_pt + xi*xo + yi*yo + si*so ;
dpt = descr_pt + (NBP/2) * binyo + (NBP/2) * binxo ;
#define atd(dbinx,dbiny,dbint) (*(dpt + (dbint)*binto + (dbiny)*binyo + (dbinx)*binxo))
/*
* Process each pixel in the window and in the (1,1)-(M-1,N-1)
* rectangle.
*/
for(dxi = max(-W, 1-xi) ; dxi <= min(+W, N-2-xi) ; ++dxi) {
for(dyi = max(-W, 1-yi) ; dyi <= min(+W, M-2-yi) ; ++dyi) {
/* Compute the gradient. */
float mod = *(pt + dxi*xo + dyi*yo + 0 ) ;
float angle = *(pt + dxi*xo + dyi*yo + buffer_size ) ;
#ifdef LOWE_COMPATIBLE
float theta = fast_mod(-angle + theta0) ;
#else
float theta = fast_mod(angle - theta0) ;
#endif
/* Get the fractional displacement. */
float dx = ((float)(xi+dxi)) - x;
float dy = ((float)(yi+dyi)) - y;
/* Get the displacement normalized w.r.t. the keypoint orientation
* and extension. */
float nx = ( ct0 * dx + st0 * dy) / SBP ;
float ny = (-st0 * dx + ct0 * dy) / SBP ;
float nt = NBO * theta / (2*M_PI) ;
/* Get the gaussian weight of the sample. The gaussian window
* has a standard deviation of NBP/2. Note that dx and dy are in
* the normalized frame, so that -NBP/2 <= dx <= NBP/2. */
const float wsigma = NBP/2 ;
float win = expf(-(nx*nx + ny*ny)/(2.0 * wsigma * wsigma)) ;
/* The sample will be distributed in 8 adijacient bins.
* Now we get the ``lower-left'' bin. */
int binx = fast_floor( nx - 0.5 ) ;
int biny = fast_floor( ny - 0.5 ) ;
int bint = fast_floor( nt ) ;
float rbinx = nx - (binx+0.5) ;
float rbiny = ny - (biny+0.5) ;
float rbint = nt - bint ;
int dbinx ;
int dbiny ;
int dbint ;
/* Distribute the current sample into the 8 adijacient bins. */
for(dbinx = 0 ; dbinx < 2 ; ++dbinx) {
for(dbiny = 0 ; dbiny < 2 ; ++dbiny) {
for(dbint = 0 ; dbint < 2 ; ++dbint) {
if( binx+dbinx >= -(NBP/2) &&
binx+dbinx < (NBP/2) &&
biny+dbiny >= -(NBP/2) &&
biny+dbiny < (NBP/2) ) {
float weight = win
* mod
* fabsf(1 - dbinx - rbinx)
* fabsf(1 - dbiny - rbiny)
* fabsf(1 - dbint - rbint) ;
atd(binx+dbinx, biny+dbiny, (bint+dbint) % NBO) += weight ;
}
}
}
}
}
}
{
/* Normalize the histogram to L2 unit length. */
normalize_histogram(descr_pt, descr_pt + NBO*NBP*NBP) ;
/* Truncate at 0.2. */
for(bin = 0; bin < NBO*NBP*NBP ; ++bin) {
if (descr_pt[bin] > 0.2) descr_pt[bin] = 0.2;
}
/* Normalize again. */
normalize_histogram(descr_pt, descr_pt + NBO*NBP*NBP) ;
}
}
}
/* Restore pointer to the beginning of the descriptors. */
descr_pt -= NBO*NBP*NBP*K ;
{
int k ;
double* L_pt ;
out[OUT_L] = mxCreateDoubleMatrix(NBP*NBP*NBO, K, mxREAL) ;
L_pt = mxGetPr(out[OUT_L]) ;
for(k = 0 ; k < NBP*NBP*NBO*K ; ++k) {
L_pt[k] = descr_pt[k] ;
}
}
mxFree(descr_pt) ;
mxFree(buffer_pt) ;
}
static float generate(float x, float y)
{
const float F2 = 0.366025403f; // F2 = 0.5*(sqrt(3.0)-1.0)
const float G2 = 0.211324865f; // G2 = (3.0-Math.sqrt(3.0))/6.0
float n0, n1, n2; // Noise contributions from the three corners
// Skew the input space to determine which simplex cell we're in
float s = (x + y) * F2; // Hairy factor for 2D
float xs = x + s;
float ys = y + s;
sint32 i = fast_floor(xs);
sint32 j = fast_floor(ys);
float t = (float)(i + j) * G2;
float X0 = i - t; // Unskew the cell origin back to (x,y) space
float Y0 = j - t;
float x0 = x - X0; // The x,y distances from the cell origin
float y0 = y - Y0;
// For the 2D case, the simplex shape is an equilateral triangle.
// Determine which simplex we are in.
sint32 i1, j1; // Offsets for second (middle) corner of simplex in (i,j) coords
if (x0 > y0) { i1 = 1; j1 = 0; } // lower triangle, XY order: (0,0)->(1,0)->(1,1)
else { i1 = 0; j1 = 1; } // upper triangle, YX order: (0,0)->(0,1)->(1,1)
// A step of (1,0) in (i,j) means a step of (1-c,-c) in (x,y), and
// a step of (0,1) in (i,j) means a step of (-c,1-c) in (x,y), where
// c = (3-sqrt(3))/6
float x1 = x0 - i1 + G2; // Offsets for middle corner in (x,y) unskewed coords
float y1 = y0 - j1 + G2;
float x2 = x0 - 1.0f + 2.0f * G2; // Offsets for last corner in (x,y) unskewed coords
float y2 = y0 - 1.0f + 2.0f * G2;
// Wrap the integer indices at 256, to avoid indexing perm[] out of bounds
sint32 ii = i % 256;
sint32 jj = j % 256;
// Calculate the contribution from the three corners
float t0 = 0.5f - x0 * x0 - y0 * y0;
if (t0 < 0.0f) n0 = 0.0f;
else {
t0 *= t0;
n0 = t0 * t0 * grad(perm[ii + perm[jj]], x0, y0);
}
float t1 = 0.5f - x1 * x1 - y1 * y1;
if (t1 < 0.0f) n1 = 0.0f;
else {
t1 *= t1;
n1 = t1 * t1 * grad(perm[ii + i1 + perm[jj + j1]], x1, y1);
}
float t2 = 0.5f - x2 * x2 - y2 * y2;
if (t2 < 0.0f) n2 = 0.0f;
else {
t2 *= t2;
n2 = t2 * t2 * grad(perm[ii + 1 + perm[jj + 1]], x2, y2);
}
// Add contributions from each corner to get the final noise value.
// The result is scaled to return values in the interval [-1,1].
return 40.0f * (n0 + n1 + n2); // TODO: The scale factor is preliminary!
}
#include "catch/catch.hpp"
#include "cata_utility.h"
#include <cmath>
#include <random>
TEST_CASE( "fast_floor", "[math]" )
{
REQUIRE( fast_floor( -2.0 ) == -2 );
REQUIRE( fast_floor( -1.9 ) == -2 );
REQUIRE( fast_floor( -1.1 ) == -2 );
REQUIRE( fast_floor( -1.0 ) == -1 );
REQUIRE( fast_floor( -0.9 ) == -1 );
REQUIRE( fast_floor( -0.1 ) == -1 );
REQUIRE( fast_floor( 0.0 ) == 0 );
REQUIRE( fast_floor( 0.1 ) == 0 );
REQUIRE( fast_floor( 0.9 ) == 0 );
REQUIRE( fast_floor( 1.0 ) == 1 );
REQUIRE( fast_floor( 1.1 ) == 1 );
REQUIRE( fast_floor( 1.9 ) == 1 );
REQUIRE( fast_floor( 2.0 ) == 2 );
REQUIRE( fast_floor( 2.1 ) == 2 );
REQUIRE( fast_floor( 2.9 ) == 2 );
std::random_device rd;
std::mt19937 mt( rd() );
std::uniform_real_distribution<double> dist( -10.0, 10.0 );
for( size_t i = 0; i != 1000; ++i ) {
double val = dist( mt );
// Classic Perlin noise, 3D version
float base(float x, float y, float z)
{
/*
// Find unit grid cell containing point
int X = fast_floor(x);
int Y = fast_floor(y);
int Z = fast_floor(z);
// Get relative xyz coordinates of point within that cell
x = x - X;
y = y - Y;
z = z - Z;
// Wrap the integer cells at 255 (smaller integer period can be introduced here)
X = X & 255;
Y = Y & 255;
Z = Z & 255;
*/
x /= 8.0; //replace with multiplication
y /= 8.0;
z /= 4.0;
//get grid point
int X = fast_floor(x);
int Y = fast_floor(y);
int Z = fast_floor(z);
//get interpolation ratio
//if (z >= 32 || z < 0) printf("ERROR1 z = %f \n", z);
x = x - X;
y = y - Y;
z = z - Z;
//if (z >= 32 || z < 0) printf("ERROR2 z = %f \n", z);
//if (x<0 || y<0 || z<0) printf("x,y,z= %f %f %f \n", x,y,z);
//if (z >= 32) printf("x,y,z= %f %f %f X,Y,Z= %i %i %i \n", x+X,y+Y,z+Z,X,Y,Z);
// Calculate a set of eight hashed gradient indices
int gi000 = get_gradient(X+0,Y+0,Z+0);
int gi001 = get_gradient(X+0,Y+0,Z+1);
int gi010 = get_gradient(X+0,Y+1,Z+0);
int gi011 = get_gradient(X+0,Y+1,Z+1);
int gi100 = get_gradient(X+1,Y+0,Z+0);
int gi101 = get_gradient(X+1,Y+0,Z+1);
int gi110 = get_gradient(X+1,Y+1,Z+0);
int gi111 = get_gradient(X+1,Y+1,Z+1);
/*
int gi000 = perm[X+perm[Y+perm[Z]]] % 12;
int gi001 = perm[X+perm[Y+perm[Z+1]]] % 12;
int gi010 = perm[X+perm[Y+1+perm[Z]]] % 12;
int gi011 = perm[X+perm[Y+1+perm[Z+1]]] % 12;
int gi100 = perm[X+1+perm[Y+perm[Z]]] % 12;
int gi101 = perm[X+1+perm[Y+perm[Z+1]]] % 12;
int gi110 = perm[X+1+perm[Y+1+perm[Z]]] % 12;
int gi111 = perm[X+1+perm[Y+1+perm[Z+1]]] % 12;
*/
// The gradients of each corner are now:
// g000 = grad3[gi000];
// g001 = grad3[gi001];
// g010 = grad3[gi010];
// g011 = grad3[gi011];
// g100 = grad3[gi100];
// g101 = grad3[gi101];
// g110 = grad3[gi110];
// g111 = grad3[gi111];
// Calculate noise contributions from each of the eight corners
float n000= dot(_grad3[gi000], x, y, z);
float n100= dot(_grad3[gi100], x-1, y, z);
float n010= dot(_grad3[gi010], x, y-1, z);
float n110= dot(_grad3[gi110], x-1, y-1, z);
float n001= dot(_grad3[gi001], x, y, z-1);
float n101= dot(_grad3[gi101], x-1, y, z-1);
float n011= dot(_grad3[gi011], x, y-1, z-1);
float n111= dot(_grad3[gi111], x-1, y-1, z-1);
// Compute the fade curve value for each of x, y, z
#if 1
float u = fade(x);
float v = fade(y);
float w = fade(z);
#else
float u = x;
float v = y;
float w = z;
#endif
// Interpolate along x the contributions from each of the corners
float nx00 = mix(n000, n100, u);
float nx01 = mix(n001, n101, u);
float nx10 = mix(n010, n110, u);
float nx11 = mix(n011, n111, u);
// Interpolate the four results along y
float nxy0 = mix(nx00, nx10, v);
float nxy1 = mix(nx01, nx11, v);
// Interpolate the two last results along z
float nxyz = mix(nxy0, nxy1, w);
return nxyz * 0.707106781; //-1 to 1
}
float raw_noise_3d(float x, float y, float z, const uint8_t* perm)
{
constexpr float F3 = 1.0f / 3.0f;
float s = (x + y + z) * F3;
float i = fast_floor(x + s);
float j = fast_floor(y + s);
float k = fast_floor(z + s);
constexpr float G3 = 1.0f / 6.0f;
float t = (i + j + k) * G3;
float x0 = x - (i - t);
float y0 = y - (j - t);
float z0 = z - (k - t);
uint8_t i1 = 0, j1 = 1, k1 = 0;
uint8_t i2 = 1, j2 = 1, k2 = 0;
if(x0 >= y0)
{
if(y0 >= z0)
{
i1 = 1;
j1 = 0;
k1 = 0;
i2 = 1;
j2 = 1;
k2 = 0;
}
else if(x0 >= z0)
{
i1 = 1;
j1 = 0;
k1 = 0;
i2 = 1;
j2 = 0;
k2 = 1;
}
else
{
i1 = 0;
j1 = 0;
k1 = 1;
i2 = 1;
j2 = 0;
k2 = 1;
}
}
else
{
if(y0 < z0)
{
i1 = 0;
j1 = 0;
k1 = 1;
i2 = 0;
j2 = 1;
k2 = 1;
}
else if(x0 < z0)
{
i1 = 0;
j1 = 1;
k1 = 0;
i2 = 0;
j2 = 1;
k2 = 1;
}
}
float x1 = x0 - i1 + G3;
float y1 = y0 - j1 + G3;
float z1 = z0 - k1 + G3;
float x2 = x0 - i2 + G3 * 2.0f;
float y2 = y0 - j2 + G3 * 2.0f;
float z2 = z0 - k2 + G3 * 2.0f;
float x3 = x0 - 1.0f + G3 * 3.0f;
float y3 = y0 - 1.0f + G3 * 3.0f;
float z3 = z0 - 1.0f + G3 * 3.0f;
uint8_t ii = int32_t(i);
uint8_t jj = int32_t(j);
uint8_t kk = int32_t(k);
uint8_t gi0 = perm[ii + perm[jj + perm[kk ] ] ] % 12;
uint8_t gi1 = perm[ii + perm[jj + perm[kk + k1] + j1] + i1] % 12;
uint8_t gi2 = perm[ii + perm[jj + perm[kk + k2] + j2] + i2] % 12;
uint8_t gi3 = perm[ii + perm[jj + perm[kk + 1 ] + 1 ] + 1 ] % 12;
float t0 = 0.5f - x0 * x0 - y0 * y0 - z0 * z0;
float t1 = 0.5f - x1 * x1 - y1 * y1 - z1 * z1;
float t2 = 0.5f - x2 * x2 - y2 * y2 - z2 * z2;
float t3 = 0.5f - x3 * x3 - y3 * y3 - z3 * z3;
float n0 = (t0 < 0) ? 0.0f : t0*t0*t0*t0 * dot(grad3[gi0], x0, y0, z0);
float n1 = (t1 < 0) ? 0.0f : t1*t1*t1*t1 * dot(grad3[gi1], x1, y1, z1);
float n2 = (t2 < 0) ? 0.0f : t2*t2*t2*t2 * dot(grad3[gi2], x2, y2, z2);
float n3 = (t3 < 0) ? 0.0f : t3*t3*t3*t3 * dot(grad3[gi3], x3, y3, z3);
return 32.0f * (n0 + n1 + n2 + n3);
}