dll/directx/wine/d3dx9_36/math.c

   1 /*
   2  * Mathematical operations specific to D3DX9.
   3  *
   4  * Copyright (C) 2008 David Adam
   5  * Copyright (C) 2008 Luis Busquets
   6  * Copyright (C) 2008 Jérôme Gardou
   7  * Copyright (C) 2008 Philip Nilsson
   8  * Copyright (C) 2008 Henri Verbeet
   9  *
  10  * This library is free software; you can redistribute it and/or
  11  * modify it under the terms of the GNU Lesser General Public
  12  * License as published by the Free Software Foundation; either
  13  * version 2.1 of the License, or (at your option) any later version.
  14  *
  15  * This library is distributed in the hope that it will be useful,
  16  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  17  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  18  * Lesser General Public License for more details.
  19  *
  20  * You should have received a copy of the GNU Lesser General Public
  21  * License along with this library; if not, write to the Free Software
  22  * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA
  23  */
  24
  25 #include "d3dx9_36_private.h"
  26
  27 struct ID3DXMatrixStackImpl
  28 {
  29   ID3DXMatrixStack ID3DXMatrixStack_iface;
  30   LONG ref;
  31
  32   unsigned int current;
  33   unsigned int stack_size;
  34   D3DXMATRIX *stack;
  35 };
  36
  37 static const unsigned int INITIAL_STACK_SIZE = 32;
  38
  39 /*_________________D3DXColor____________________*/
  40
  41 D3DXCOLOR* WINAPI D3DXColorAdjustContrast(D3DXCOLOR *pout, const D3DXCOLOR *pc, FLOAT s)
  42 {
  43     TRACE("pout %p, pc %p, s %f\n", pout, pc, s);
  44
  45     pout->r = 0.5f + s * (pc->r - 0.5f);
  46     pout->g = 0.5f + s * (pc->g - 0.5f);
  47     pout->b = 0.5f + s * (pc->b - 0.5f);
  48     pout->a = pc->a;
  49     return pout;
  50 }
  51
  52 D3DXCOLOR* WINAPI D3DXColorAdjustSaturation(D3DXCOLOR *pout, const D3DXCOLOR *pc, FLOAT s)
  53 {
  54     FLOAT grey;
  55
  56     TRACE("pout %p, pc %p, s %f\n", pout, pc, s);
  57
  58     grey = pc->r * 0.2125f + pc->g * 0.7154f + pc->b * 0.0721f;
  59     pout->r = grey + s * (pc->r - grey);
  60     pout->g = grey + s * (pc->g - grey);
  61     pout->b = grey + s * (pc->b - grey);
  62     pout->a = pc->a;
  63     return pout;
  64 }
  65
  66 /*_________________Misc__________________________*/
  67
  68 FLOAT WINAPI D3DXFresnelTerm(FLOAT costheta, FLOAT refractionindex)
  69 {
  70     FLOAT a, d, g, result;
  71
  72     TRACE("costheta %f, refractionindex %f\n", costheta, refractionindex);
  73
  74     g = sqrtf(refractionindex * refractionindex + costheta * costheta - 1.0f);
  75     a = g + costheta;
  76     d = g - costheta;
  77     result = (costheta * a - 1.0f) * (costheta * a - 1.0f) / ((costheta * d + 1.0f) * (costheta * d + 1.0f)) + 1.0f;
  78     result *= 0.5f * d * d / (a * a);
  79
  80     return result;
  81 }
  82
  83 /*_________________D3DXMatrix____________________*/
  84
  85 D3DXMATRIX * WINAPI D3DXMatrixAffineTransformation(D3DXMATRIX *out, FLOAT scaling, const D3DXVECTOR3 *rotationcenter,
  86         const D3DXQUATERNION *rotation, const D3DXVECTOR3 *translation)
  87 {
  88     TRACE("out %p, scaling %f, rotationcenter %p, rotation %p, translation %p\n",
  89             out, scaling, rotationcenter, rotation, translation);
  90
  91     D3DXMatrixIdentity(out);
  92
  93     if (rotation)
  94     {
  95         FLOAT temp00, temp01, temp02, temp10, temp11, temp12, temp20, temp21, temp22;
  96
  97         temp00 = 1.0f - 2.0f * (rotation->y * rotation->y + rotation->z * rotation->z);
  98         temp01 = 2.0f * (rotation->x * rotation->y + rotation->z * rotation->w);
  99         temp02 = 2.0f * (rotation->x * rotation->z - rotation->y * rotation->w);
 100         temp10 = 2.0f * (rotation->x * rotation->y - rotation->z * rotation->w);
 101         temp11 = 1.0f - 2.0f * (rotation->x * rotation->x + rotation->z * rotation->z);
 102         temp12 = 2.0f * (rotation->y * rotation->z + rotation->x * rotation->w);
 103         temp20 = 2.0f * (rotation->x * rotation->z + rotation->y * rotation->w);
 104         temp21 = 2.0f * (rotation->y * rotation->z - rotation->x * rotation->w);
 105         temp22 = 1.0f - 2.0f * (rotation->x * rotation->x + rotation->y * rotation->y);
 106
 107         out->u.m[0][0] = scaling * temp00;
 108         out->u.m[0][1] = scaling * temp01;
 109         out->u.m[0][2] = scaling * temp02;
 110         out->u.m[1][0] = scaling * temp10;
 111         out->u.m[1][1] = scaling * temp11;
 112         out->u.m[1][2] = scaling * temp12;
 113         out->u.m[2][0] = scaling * temp20;
 114         out->u.m[2][1] = scaling * temp21;
 115         out->u.m[2][2] = scaling * temp22;
 116
 117         if (rotationcenter)
 118         {
 119             out->u.m[3][0] = rotationcenter->x * (1.0f - temp00) - rotationcenter->y * temp10
 120                     - rotationcenter->z * temp20;
 121             out->u.m[3][1] = rotationcenter->y * (1.0f - temp11) - rotationcenter->x * temp01
 122                     - rotationcenter->z * temp21;
 123             out->u.m[3][2] = rotationcenter->z * (1.0f - temp22) - rotationcenter->x * temp02
 124                     - rotationcenter->y * temp12;
 125         }
 126     }
 127     else
 128     {
 129         out->u.m[0][0] = scaling;
 130         out->u.m[1][1] = scaling;
 131         out->u.m[2][2] = scaling;
 132     }
 133
 134     if (translation)
 135     {
 136         out->u.m[3][0] += translation->x;
 137         out->u.m[3][1] += translation->y;
 138         out->u.m[3][2] += translation->z;
 139     }
 140
 141     return out;
 142 }
 143
 144 D3DXMATRIX * WINAPI D3DXMatrixAffineTransformation2D(D3DXMATRIX *out, FLOAT scaling,
 145         const D3DXVECTOR2 *rotationcenter, FLOAT rotation, const D3DXVECTOR2 *translation)
 146 {
 147     FLOAT tmp1, tmp2, s;
 148
 149     TRACE("out %p, scaling %f, rotationcenter %p, rotation %f, translation %p\n",
 150             out, scaling, rotationcenter, rotation, translation);
 151
 152     s = sinf(rotation / 2.0f);
 153     tmp1 = 1.0f - 2.0f * s * s;
 154     tmp2 = 2.0f * s * cosf(rotation / 2.0f);
 155
 156     D3DXMatrixIdentity(out);
 157     out->u.m[0][0] = scaling * tmp1;
 158     out->u.m[0][1] = scaling * tmp2;
 159     out->u.m[1][0] = -scaling * tmp2;
 160     out->u.m[1][1] = scaling * tmp1;
 161
 162     if (rotationcenter)
 163     {
 164         FLOAT x, y;
 165
 166         x = rotationcenter->x;
 167         y = rotationcenter->y;
 168
 169         out->u.m[3][0] = y * tmp2 - x * tmp1 + x;
 170         out->u.m[3][1] = -x * tmp2 - y * tmp1 + y;
 171     }
 172
 173     if (translation)
 174     {
 175         out->u.m[3][0] += translation->x;
 176         out->u.m[3][1] += translation->y;
 177     }
 178
 179     return out;
 180 }
 181
 182 HRESULT WINAPI D3DXMatrixDecompose(D3DXVECTOR3 *poutscale, D3DXQUATERNION *poutrotation, D3DXVECTOR3 *pouttranslation, const D3DXMATRIX *pm)
 183 {
 184     D3DXMATRIX normalized;
 185     D3DXVECTOR3 vec;
 186
 187     TRACE("poutscale %p, poutrotation %p, pouttranslation %p, pm %p\n", poutscale, poutrotation, pouttranslation, pm);
 188
 189     /*Compute the scaling part.*/
 190     vec.x=pm->u.m[0][0];
 191     vec.y=pm->u.m[0][1];
 192     vec.z=pm->u.m[0][2];
 193     poutscale->x=D3DXVec3Length(&vec);
 194
 195     vec.x=pm->u.m[1][0];
 196     vec.y=pm->u.m[1][1];
 197     vec.z=pm->u.m[1][2];
 198     poutscale->y=D3DXVec3Length(&vec);
 199
 200     vec.x=pm->u.m[2][0];
 201     vec.y=pm->u.m[2][1];
 202     vec.z=pm->u.m[2][2];
 203     poutscale->z=D3DXVec3Length(&vec);
 204
 205     /*Compute the translation part.*/
 206     pouttranslation->x=pm->u.m[3][0];
 207     pouttranslation->y=pm->u.m[3][1];
 208     pouttranslation->z=pm->u.m[3][2];
 209
 210     /*Let's calculate the rotation now*/
 211     if ( (poutscale->x == 0.0f) || (poutscale->y == 0.0f) || (poutscale->z == 0.0f) ) return D3DERR_INVALIDCALL;
 212
 213     normalized.u.m[0][0]=pm->u.m[0][0]/poutscale->x;
 214     normalized.u.m[0][1]=pm->u.m[0][1]/poutscale->x;
 215     normalized.u.m[0][2]=pm->u.m[0][2]/poutscale->x;
 216     normalized.u.m[1][0]=pm->u.m[1][0]/poutscale->y;
 217     normalized.u.m[1][1]=pm->u.m[1][1]/poutscale->y;
 218     normalized.u.m[1][2]=pm->u.m[1][2]/poutscale->y;
 219     normalized.u.m[2][0]=pm->u.m[2][0]/poutscale->z;
 220     normalized.u.m[2][1]=pm->u.m[2][1]/poutscale->z;
 221     normalized.u.m[2][2]=pm->u.m[2][2]/poutscale->z;
 222
 223     D3DXQuaternionRotationMatrix(poutrotation,&normalized);
 224     return S_OK;
 225 }
 226
 227 FLOAT WINAPI D3DXMatrixDeterminant(const D3DXMATRIX *pm)
 228 {
 229     FLOAT t[3], v[4];
 230
 231     TRACE("pm %p\n", pm);
 232
 233     t[0] = pm->u.m[2][2] * pm->u.m[3][3] - pm->u.m[2][3] * pm->u.m[3][2];
 234     t[1] = pm->u.m[1][2] * pm->u.m[3][3] - pm->u.m[1][3] * pm->u.m[3][2];
 235     t[2] = pm->u.m[1][2] * pm->u.m[2][3] - pm->u.m[1][3] * pm->u.m[2][2];
 236     v[0] = pm->u.m[1][1] * t[0] - pm->u.m[2][1] * t[1] + pm->u.m[3][1] * t[2];
 237     v[1] = -pm->u.m[1][0] * t[0] + pm->u.m[2][0] * t[1] - pm->u.m[3][0] * t[2];
 238
 239     t[0] = pm->u.m[1][0] * pm->u.m[2][1] - pm->u.m[2][0] * pm->u.m[1][1];
 240     t[1] = pm->u.m[1][0] * pm->u.m[3][1] - pm->u.m[3][0] * pm->u.m[1][1];
 241     t[2] = pm->u.m[2][0] * pm->u.m[3][1] - pm->u.m[3][0] * pm->u.m[2][1];
 242     v[2] = pm->u.m[3][3] * t[0] - pm->u.m[2][3] * t[1] + pm->u.m[1][3] * t[2];
 243     v[3] = -pm->u.m[3][2] * t[0] + pm->u.m[2][2] * t[1] - pm->u.m[1][2] * t[2];
 244
 245     return pm->u.m[0][0] * v[0] + pm->u.m[0][1] * v[1] +
 246         pm->u.m[0][2] * v[2] + pm->u.m[0][3] * v[3];
 247 }
 248
 249 D3DXMATRIX* WINAPI D3DXMatrixInverse(D3DXMATRIX *pout, FLOAT *pdeterminant, const D3DXMATRIX *pm)
 250 {
 251     FLOAT det, t[3], v[16];
 252     UINT i, j;
 253
 254     TRACE("pout %p, pdeterminant %p, pm %p\n", pout, pdeterminant, pm);
 255
 256     t[0] = pm->u.m[2][2] * pm->u.m[3][3] - pm->u.m[2][3] * pm->u.m[3][2];
 257     t[1] = pm->u.m[1][2] * pm->u.m[3][3] - pm->u.m[1][3] * pm->u.m[3][2];
 258     t[2] = pm->u.m[1][2] * pm->u.m[2][3] - pm->u.m[1][3] * pm->u.m[2][2];
 259     v[0] = pm->u.m[1][1] * t[0] - pm->u.m[2][1] * t[1] + pm->u.m[3][1] * t[2];
 260     v[4] = -pm->u.m[1][0] * t[0] + pm->u.m[2][0] * t[1] - pm->u.m[3][0] * t[2];
 261
 262     t[0] = pm->u.m[1][0] * pm->u.m[2][1] - pm->u.m[2][0] * pm->u.m[1][1];
 263     t[1] = pm->u.m[1][0] * pm->u.m[3][1] - pm->u.m[3][0] * pm->u.m[1][1];
 264     t[2] = pm->u.m[2][0] * pm->u.m[3][1] - pm->u.m[3][0] * pm->u.m[2][1];
 265     v[8] = pm->u.m[3][3] * t[0] - pm->u.m[2][3] * t[1] + pm->u.m[1][3] * t[2];
 266     v[12] = -pm->u.m[3][2] * t[0] + pm->u.m[2][2] * t[1] - pm->u.m[1][2] * t[2];
 267
 268     det = pm->u.m[0][0] * v[0] + pm->u.m[0][1] * v[4] +
 269         pm->u.m[0][2] * v[8] + pm->u.m[0][3] * v[12];
 270     if (det == 0.0f)
 271         return NULL;
 272     if (pdeterminant)
 273         *pdeterminant = det;
 274
 275     t[0] = pm->u.m[2][2] * pm->u.m[3][3] - pm->u.m[2][3] * pm->u.m[3][2];
 276     t[1] = pm->u.m[0][2] * pm->u.m[3][3] - pm->u.m[0][3] * pm->u.m[3][2];
 277     t[2] = pm->u.m[0][2] * pm->u.m[2][3] - pm->u.m[0][3] * pm->u.m[2][2];
 278     v[1] = -pm->u.m[0][1] * t[0] + pm->u.m[2][1] * t[1] - pm->u.m[3][1] * t[2];
 279     v[5] = pm->u.m[0][0] * t[0] - pm->u.m[2][0] * t[1] + pm->u.m[3][0] * t[2];
 280
 281     t[0] = pm->u.m[0][0] * pm->u.m[2][1] - pm->u.m[2][0] * pm->u.m[0][1];
 282     t[1] = pm->u.m[3][0] * pm->u.m[0][1] - pm->u.m[0][0] * pm->u.m[3][1];
 283     t[2] = pm->u.m[2][0] * pm->u.m[3][1] - pm->u.m[3][0] * pm->u.m[2][1];
 284     v[9] = -pm->u.m[3][3] * t[0] - pm->u.m[2][3] * t[1]- pm->u.m[0][3] * t[2];
 285     v[13] = pm->u.m[3][2] * t[0] + pm->u.m[2][2] * t[1] + pm->u.m[0][2] * t[2];
 286
 287     t[0] = pm->u.m[1][2] * pm->u.m[3][3] - pm->u.m[1][3] * pm->u.m[3][2];
 288     t[1] = pm->u.m[0][2] * pm->u.m[3][3] - pm->u.m[0][3] * pm->u.m[3][2];
 289     t[2] = pm->u.m[0][2] * pm->u.m[1][3] - pm->u.m[0][3] * pm->u.m[1][2];
 290     v[2] = pm->u.m[0][1] * t[0] - pm->u.m[1][1] * t[1] + pm->u.m[3][1] * t[2];
 291     v[6] = -pm->u.m[0][0] * t[0] + pm->u.m[1][0] * t[1] - pm->u.m[3][0] * t[2];
 292
 293     t[0] = pm->u.m[0][0] * pm->u.m[1][1] - pm->u.m[1][0] * pm->u.m[0][1];
 294     t[1] = pm->u.m[3][0] * pm->u.m[0][1] - pm->u.m[0][0] * pm->u.m[3][1];
 295     t[2] = pm->u.m[1][0] * pm->u.m[3][1] - pm->u.m[3][0] * pm->u.m[1][1];
 296     v[10] = pm->u.m[3][3] * t[0] + pm->u.m[1][3] * t[1] + pm->u.m[0][3] * t[2];
 297     v[14] = -pm->u.m[3][2] * t[0] - pm->u.m[1][2] * t[1] - pm->u.m[0][2] * t[2];
 298
 299     t[0] = pm->u.m[1][2] * pm->u.m[2][3] - pm->u.m[1][3] * pm->u.m[2][2];
 300     t[1] = pm->u.m[0][2] * pm->u.m[2][3] - pm->u.m[0][3] * pm->u.m[2][2];
 301     t[2] = pm->u.m[0][2] * pm->u.m[1][3] - pm->u.m[0][3] * pm->u.m[1][2];
 302     v[3] = -pm->u.m[0][1] * t[0] + pm->u.m[1][1] * t[1] - pm->u.m[2][1] * t[2];
 303     v[7] = pm->u.m[0][0] * t[0] - pm->u.m[1][0] * t[1] + pm->u.m[2][0] * t[2];
 304
 305     v[11] = -pm->u.m[0][0] * (pm->u.m[1][1] * pm->u.m[2][3] - pm->u.m[1][3] * pm->u.m[2][1]) +
 306         pm->u.m[1][0] * (pm->u.m[0][1] * pm->u.m[2][3] - pm->u.m[0][3] * pm->u.m[2][1]) -
 307         pm->u.m[2][0] * (pm->u.m[0][1] * pm->u.m[1][3] - pm->u.m[0][3] * pm->u.m[1][1]);
 308
 309     v[15] = pm->u.m[0][0] * (pm->u.m[1][1] * pm->u.m[2][2] - pm->u.m[1][2] * pm->u.m[2][1]) -
 310         pm->u.m[1][0] * (pm->u.m[0][1] * pm->u.m[2][2] - pm->u.m[0][2] * pm->u.m[2][1]) +
 311         pm->u.m[2][0] * (pm->u.m[0][1] * pm->u.m[1][2] - pm->u.m[0][2] * pm->u.m[1][1]);
 312
 313     det = 1.0f / det;
 314
 315     for (i = 0; i < 4; i++)
 316         for (j = 0; j < 4; j++)
 317             pout->u.m[i][j] = v[4 * i + j] * det;
 318
 319     return pout;
 320 }
 321
 322 D3DXMATRIX * WINAPI D3DXMatrixLookAtLH(D3DXMATRIX *out, const D3DXVECTOR3 *eye, const D3DXVECTOR3 *at,
 323         const D3DXVECTOR3 *up)
 324 {
 325     D3DXVECTOR3 right, upn, vec;
 326
 327     TRACE("out %p, eye %p, at %p, up %p\n", out, eye, at, up);
 328
 329     D3DXVec3Subtract(&vec, at, eye);
 330     D3DXVec3Normalize(&vec, &vec);
 331     D3DXVec3Cross(&right, up, &vec);
 332     D3DXVec3Cross(&upn, &vec, &right);
 333     D3DXVec3Normalize(&right, &right);
 334     D3DXVec3Normalize(&upn, &upn);
 335     out->u.m[0][0] = right.x;
 336     out->u.m[1][0] = right.y;
 337     out->u.m[2][0] = right.z;
 338     out->u.m[3][0] = -D3DXVec3Dot(&right, eye);
 339     out->u.m[0][1] = upn.x;
 340     out->u.m[1][1] = upn.y;
 341     out->u.m[2][1] = upn.z;
 342     out->u.m[3][1] = -D3DXVec3Dot(&upn, eye);
 343     out->u.m[0][2] = vec.x;
 344     out->u.m[1][2] = vec.y;
 345     out->u.m[2][2] = vec.z;
 346     out->u.m[3][2] = -D3DXVec3Dot(&vec, eye);
 347     out->u.m[0][3] = 0.0f;
 348     out->u.m[1][3] = 0.0f;
 349     out->u.m[2][3] = 0.0f;
 350     out->u.m[3][3] = 1.0f;
 351
 352     return out;
 353 }
 354
 355 D3DXMATRIX * WINAPI D3DXMatrixLookAtRH(D3DXMATRIX *out, const D3DXVECTOR3 *eye, const D3DXVECTOR3 *at,
 356         const D3DXVECTOR3 *up)
 357 {
 358     D3DXVECTOR3 right, upn, vec;
 359
 360     TRACE("out %p, eye %p, at %p, up %p\n", out, eye, at, up);
 361
 362     D3DXVec3Subtract(&vec, at, eye);
 363     D3DXVec3Normalize(&vec, &vec);
 364     D3DXVec3Cross(&right, up, &vec);
 365     D3DXVec3Cross(&upn, &vec, &right);
 366     D3DXVec3Normalize(&right, &right);
 367     D3DXVec3Normalize(&upn, &upn);
 368     out->u.m[0][0] = -right.x;
 369     out->u.m[1][0] = -right.y;
 370     out->u.m[2][0] = -right.z;
 371     out->u.m[3][0] = D3DXVec3Dot(&right, eye);
 372     out->u.m[0][1] = upn.x;
 373     out->u.m[1][1] = upn.y;
 374     out->u.m[2][1] = upn.z;
 375     out->u.m[3][1] = -D3DXVec3Dot(&upn, eye);
 376     out->u.m[0][2] = -vec.x;
 377     out->u.m[1][2] = -vec.y;
 378     out->u.m[2][2] = -vec.z;
 379     out->u.m[3][2] = D3DXVec3Dot(&vec, eye);
 380     out->u.m[0][3] = 0.0f;
 381     out->u.m[1][3] = 0.0f;
 382     out->u.m[2][3] = 0.0f;
 383     out->u.m[3][3] = 1.0f;
 384
 385     return out;
 386 }
 387
 388 D3DXMATRIX* WINAPI D3DXMatrixMultiply(D3DXMATRIX *pout, const D3DXMATRIX *pm1, const D3DXMATRIX *pm2)
 389 {
 390     D3DXMATRIX out;
 391     int i,j;
 392
 393     TRACE("pout %p, pm1 %p, pm2 %p\n", pout, pm1, pm2);
 394
 395     for (i=0; i<4; i++)
 396     {
 397         for (j=0; j<4; j++)
 398         {
 399             out.u.m[i][j] = pm1->u.m[i][0] * pm2->u.m[0][j] + pm1->u.m[i][1] * pm2->u.m[1][j] + pm1->u.m[i][2] * pm2->u.m[2][j] + pm1->u.m[i][3] * pm2->u.m[3][j];
 400         }
 401     }
 402
 403     *pout = out;
 404     return pout;
 405 }
 406
 407 D3DXMATRIX* WINAPI D3DXMatrixMultiplyTranspose(D3DXMATRIX *pout, const D3DXMATRIX *pm1, const D3DXMATRIX *pm2)
 408 {
 409     D3DXMATRIX temp;
 410     int i, j;
 411
 412     TRACE("pout %p, pm1 %p, pm2 %p\n", pout, pm1, pm2);
 413
 414     for (i = 0; i < 4; i++)
 415         for (j = 0; j < 4; j++)
 416             temp.u.m[j][i] = pm1->u.m[i][0] * pm2->u.m[0][j] + pm1->u.m[i][1] * pm2->u.m[1][j] + pm1->u.m[i][2] * pm2->u.m[2][j] + pm1->u.m[i][3] * pm2->u.m[3][j];
 417
 418     *pout = temp;
 419     return pout;
 420 }
 421
 422 D3DXMATRIX* WINAPI D3DXMatrixOrthoLH(D3DXMATRIX *pout, FLOAT w, FLOAT h, FLOAT zn, FLOAT zf)
 423 {
 424     TRACE("pout %p, w %f, h %f, zn %f, zf %f\n", pout, w, h, zn, zf);
 425
 426     D3DXMatrixIdentity(pout);
 427     pout->u.m[0][0] = 2.0f / w;
 428     pout->u.m[1][1] = 2.0f / h;
 429     pout->u.m[2][2] = 1.0f / (zf - zn);
 430     pout->u.m[3][2] = zn / (zn - zf);
 431     return pout;
 432 }
 433
 434 D3DXMATRIX* WINAPI D3DXMatrixOrthoOffCenterLH(D3DXMATRIX *pout, FLOAT l, FLOAT r, FLOAT b, FLOAT t, FLOAT zn, FLOAT zf)
 435 {
 436     TRACE("pout %p, l %f, r %f, b %f, t %f, zn %f, zf %f\n", pout, l, r, b, t, zn, zf);
 437
 438     D3DXMatrixIdentity(pout);
 439     pout->u.m[0][0] = 2.0f / (r - l);
 440     pout->u.m[1][1] = 2.0f / (t - b);
 441     pout->u.m[2][2] = 1.0f / (zf -zn);
 442     pout->u.m[3][0] = -1.0f -2.0f *l / (r - l);
 443     pout->u.m[3][1] = 1.0f + 2.0f * t / (b - t);
 444     pout->u.m[3][2] = zn / (zn -zf);
 445     return pout;
 446 }
 447
 448 D3DXMATRIX* WINAPI D3DXMatrixOrthoOffCenterRH(D3DXMATRIX *pout, FLOAT l, FLOAT r, FLOAT b, FLOAT t, FLOAT zn, FLOAT zf)
 449 {
 450     TRACE("pout %p, l %f, r %f, b %f, t %f, zn %f, zf %f\n", pout, l, r, b, t, zn, zf);
 451
 452     D3DXMatrixIdentity(pout);
 453     pout->u.m[0][0] = 2.0f / (r - l);
 454     pout->u.m[1][1] = 2.0f / (t - b);
 455     pout->u.m[2][2] = 1.0f / (zn -zf);
 456     pout->u.m[3][0] = -1.0f -2.0f *l / (r - l);
 457     pout->u.m[3][1] = 1.0f + 2.0f * t / (b - t);
 458     pout->u.m[3][2] = zn / (zn -zf);
 459     return pout;
 460 }
 461
 462 D3DXMATRIX* WINAPI D3DXMatrixOrthoRH(D3DXMATRIX *pout, FLOAT w, FLOAT h, FLOAT zn, FLOAT zf)
 463 {
 464     TRACE("pout %p, w %f, h %f, zn %f, zf %f\n", pout, w, h, zn, zf);
 465
 466     D3DXMatrixIdentity(pout);
 467     pout->u.m[0][0] = 2.0f / w;
 468     pout->u.m[1][1] = 2.0f / h;
 469     pout->u.m[2][2] = 1.0f / (zn - zf);
 470     pout->u.m[3][2] = zn / (zn - zf);
 471     return pout;
 472 }
 473
 474 D3DXMATRIX* WINAPI D3DXMatrixPerspectiveFovLH(D3DXMATRIX *pout, FLOAT fovy, FLOAT aspect, FLOAT zn, FLOAT zf)
 475 {
 476     TRACE("pout %p, fovy %f, aspect %f, zn %f, zf %f\n", pout, fovy, aspect, zn, zf);
 477
 478     D3DXMatrixIdentity(pout);
 479     pout->u.m[0][0] = 1.0f / (aspect * tanf(fovy/2.0f));
 480     pout->u.m[1][1] = 1.0f / tanf(fovy/2.0f);
 481     pout->u.m[2][2] = zf / (zf - zn);
 482     pout->u.m[2][3] = 1.0f;
 483     pout->u.m[3][2] = (zf * zn) / (zn - zf);
 484     pout->u.m[3][3] = 0.0f;
 485     return pout;
 486 }
 487
 488 D3DXMATRIX* WINAPI D3DXMatrixPerspectiveFovRH(D3DXMATRIX *pout, FLOAT fovy, FLOAT aspect, FLOAT zn, FLOAT zf)
 489 {
 490     TRACE("pout %p, fovy %f, aspect %f, zn %f, zf %f\n", pout, fovy, aspect, zn, zf);
 491
 492     D3DXMatrixIdentity(pout);
 493     pout->u.m[0][0] = 1.0f / (aspect * tanf(fovy/2.0f));
 494     pout->u.m[1][1] = 1.0f / tanf(fovy/2.0f);
 495     pout->u.m[2][2] = zf / (zn - zf);
 496     pout->u.m[2][3] = -1.0f;
 497     pout->u.m[3][2] = (zf * zn) / (zn - zf);
 498     pout->u.m[3][3] = 0.0f;
 499     return pout;
 500 }
 501
 502 D3DXMATRIX* WINAPI D3DXMatrixPerspectiveLH(D3DXMATRIX *pout, FLOAT w, FLOAT h, FLOAT zn, FLOAT zf)
 503 {
 504     TRACE("pout %p, w %f, h %f, zn %f, zf %f\n", pout, w, h, zn, zf);
 505
 506     D3DXMatrixIdentity(pout);
 507     pout->u.m[0][0] = 2.0f * zn / w;
 508     pout->u.m[1][1] = 2.0f * zn / h;
 509     pout->u.m[2][2] = zf / (zf - zn);
 510     pout->u.m[3][2] = (zn * zf) / (zn - zf);
 511     pout->u.m[2][3] = 1.0f;
 512     pout->u.m[3][3] = 0.0f;
 513     return pout;
 514 }
 515
 516 D3DXMATRIX* WINAPI D3DXMatrixPerspectiveOffCenterLH(D3DXMATRIX *pout, FLOAT l, FLOAT r, FLOAT b, FLOAT t, FLOAT zn, FLOAT zf)
 517 {
 518     TRACE("pout %p, l %f, r %f, b %f, t %f, zn %f, zf %f\n", pout, l, r, b, t, zn, zf);
 519
 520     D3DXMatrixIdentity(pout);
 521     pout->u.m[0][0] = 2.0f * zn / (r - l);
 522     pout->u.m[1][1] = -2.0f * zn / (b - t);
 523     pout->u.m[2][0] = -1.0f - 2.0f * l / (r - l);
 524     pout->u.m[2][1] = 1.0f + 2.0f * t / (b - t);
 525     pout->u.m[2][2] = - zf / (zn - zf);
 526     pout->u.m[3][2] = (zn * zf) / (zn -zf);
 527     pout->u.m[2][3] = 1.0f;
 528     pout->u.m[3][3] = 0.0f;
 529     return pout;
 530 }
 531
 532 D3DXMATRIX* WINAPI D3DXMatrixPerspectiveOffCenterRH(D3DXMATRIX *pout, FLOAT l, FLOAT r, FLOAT b, FLOAT t, FLOAT zn, FLOAT zf)
 533 {
 534     TRACE("pout %p, l %f, r %f, b %f, t %f, zn %f, zf %f\n", pout, l, r, b, t, zn, zf);
 535
 536     D3DXMatrixIdentity(pout);
 537     pout->u.m[0][0] = 2.0f * zn / (r - l);
 538     pout->u.m[1][1] = -2.0f * zn / (b - t);
 539     pout->u.m[2][0] = 1.0f + 2.0f * l / (r - l);
 540     pout->u.m[2][1] = -1.0f -2.0f * t / (b - t);
 541     pout->u.m[2][2] = zf / (zn - zf);
 542     pout->u.m[3][2] = (zn * zf) / (zn -zf);
 543     pout->u.m[2][3] = -1.0f;
 544     pout->u.m[3][3] = 0.0f;
 545     return pout;
 546 }
 547
 548 D3DXMATRIX* WINAPI D3DXMatrixPerspectiveRH(D3DXMATRIX *pout, FLOAT w, FLOAT h, FLOAT zn, FLOAT zf)
 549 {
 550     TRACE("pout %p, w %f, h %f, zn %f, zf %f\n", pout, w, h, zn, zf);
 551
 552     D3DXMatrixIdentity(pout);
 553     pout->u.m[0][0] = 2.0f * zn / w;
 554     pout->u.m[1][1] = 2.0f * zn / h;
 555     pout->u.m[2][2] = zf / (zn - zf);
 556     pout->u.m[3][2] = (zn * zf) / (zn - zf);
 557     pout->u.m[2][3] = -1.0f;
 558     pout->u.m[3][3] = 0.0f;
 559     return pout;
 560 }
 561
 562 D3DXMATRIX* WINAPI D3DXMatrixReflect(D3DXMATRIX *pout, const D3DXPLANE *pplane)
 563 {
 564     D3DXPLANE Nplane;
 565
 566     TRACE("pout %p, pplane %p\n", pout, pplane);
 567
 568     D3DXPlaneNormalize(&Nplane, pplane);
 569     D3DXMatrixIdentity(pout);
 570     pout->u.m[0][0] = 1.0f - 2.0f * Nplane.a * Nplane.a;
 571     pout->u.m[0][1] = -2.0f * Nplane.a * Nplane.b;
 572     pout->u.m[0][2] = -2.0f * Nplane.a * Nplane.c;
 573     pout->u.m[1][0] = -2.0f * Nplane.a * Nplane.b;
 574     pout->u.m[1][1] = 1.0f - 2.0f * Nplane.b * Nplane.b;
 575     pout->u.m[1][2] = -2.0f * Nplane.b * Nplane.c;
 576     pout->u.m[2][0] = -2.0f * Nplane.c * Nplane.a;
 577     pout->u.m[2][1] = -2.0f * Nplane.c * Nplane.b;
 578     pout->u.m[2][2] = 1.0f - 2.0f * Nplane.c * Nplane.c;
 579     pout->u.m[3][0] = -2.0f * Nplane.d * Nplane.a;
 580     pout->u.m[3][1] = -2.0f * Nplane.d * Nplane.b;
 581     pout->u.m[3][2] = -2.0f * Nplane.d * Nplane.c;
 582     return pout;
 583 }
 584
 585 D3DXMATRIX * WINAPI D3DXMatrixRotationAxis(D3DXMATRIX *out, const D3DXVECTOR3 *v, FLOAT angle)
 586 {
 587     D3DXVECTOR3 nv;
 588     FLOAT sangle, cangle, cdiff;
 589
 590     TRACE("out %p, v %p, angle %f\n", out, v, angle);
 591
 592     D3DXVec3Normalize(&nv, v);
 593     sangle = sinf(angle);
 594     cangle = cosf(angle);
 595     cdiff = 1.0f - cangle;
 596
 597     out->u.m[0][0] = cdiff * nv.x * nv.x + cangle;
 598     out->u.m[1][0] = cdiff * nv.x * nv.y - sangle * nv.z;
 599     out->u.m[2][0] = cdiff * nv.x * nv.z + sangle * nv.y;
 600     out->u.m[3][0] = 0.0f;
 601     out->u.m[0][1] = cdiff * nv.y * nv.x + sangle * nv.z;
 602     out->u.m[1][1] = cdiff * nv.y * nv.y + cangle;
 603     out->u.m[2][1] = cdiff * nv.y * nv.z - sangle * nv.x;
 604     out->u.m[3][1] = 0.0f;
 605     out->u.m[0][2] = cdiff * nv.z * nv.x - sangle * nv.y;
 606     out->u.m[1][2] = cdiff * nv.z * nv.y + sangle * nv.x;
 607     out->u.m[2][2] = cdiff * nv.z * nv.z + cangle;
 608     out->u.m[3][2] = 0.0f;
 609     out->u.m[0][3] = 0.0f;
 610     out->u.m[1][3] = 0.0f;
 611     out->u.m[2][3] = 0.0f;
 612     out->u.m[3][3] = 1.0f;
 613
 614     return out;
 615 }
 616
 617 D3DXMATRIX* WINAPI D3DXMatrixRotationQuaternion(D3DXMATRIX *pout, const D3DXQUATERNION *pq)
 618 {
 619     TRACE("pout %p, pq %p\n", pout, pq);
 620
 621     D3DXMatrixIdentity(pout);
 622     pout->u.m[0][0] = 1.0f - 2.0f * (pq->y * pq->y + pq->z * pq->z);
 623     pout->u.m[0][1] = 2.0f * (pq->x *pq->y + pq->z * pq->w);
 624     pout->u.m[0][2] = 2.0f * (pq->x * pq->z - pq->y * pq->w);
 625     pout->u.m[1][0] = 2.0f * (pq->x * pq->y - pq->z * pq->w);
 626     pout->u.m[1][1] = 1.0f - 2.0f * (pq->x * pq->x + pq->z * pq->z);
 627     pout->u.m[1][2] = 2.0f * (pq->y *pq->z + pq->x *pq->w);
 628     pout->u.m[2][0] = 2.0f * (pq->x * pq->z + pq->y * pq->w);
 629     pout->u.m[2][1] = 2.0f * (pq->y *pq->z - pq->x *pq->w);
 630     pout->u.m[2][2] = 1.0f - 2.0f * (pq->x * pq->x + pq->y * pq->y);
 631     return pout;
 632 }
 633
 634 D3DXMATRIX* WINAPI D3DXMatrixRotationX(D3DXMATRIX *pout, FLOAT angle)
 635 {
 636     TRACE("pout %p, angle %f\n", pout, angle);
 637
 638     D3DXMatrixIdentity(pout);
 639     pout->u.m[1][1] = cosf(angle);
 640     pout->u.m[2][2] = cosf(angle);
 641     pout->u.m[1][2] = sinf(angle);
 642     pout->u.m[2][1] = -sinf(angle);
 643     return pout;
 644 }
 645
 646 D3DXMATRIX* WINAPI D3DXMatrixRotationY(D3DXMATRIX *pout, FLOAT angle)
 647 {
 648     TRACE("pout %p, angle %f\n", pout, angle);
 649
 650     D3DXMatrixIdentity(pout);
 651     pout->u.m[0][0] = cosf(angle);
 652     pout->u.m[2][2] = cosf(angle);
 653     pout->u.m[0][2] = -sinf(angle);
 654     pout->u.m[2][0] = sinf(angle);
 655     return pout;
 656 }
 657
 658 D3DXMATRIX * WINAPI D3DXMatrixRotationYawPitchRoll(D3DXMATRIX *out, FLOAT yaw, FLOAT pitch, FLOAT roll)
 659 {
 660     FLOAT sroll, croll, spitch, cpitch, syaw, cyaw;
 661
 662     TRACE("out %p, yaw %f, pitch %f, roll %f\n", out, yaw, pitch, roll);
 663
 664     sroll = sinf(roll);
 665     croll = cosf(roll);
 666     spitch = sinf(pitch);
 667     cpitch = cosf(pitch);
 668     syaw = sinf(yaw);
 669     cyaw = cosf(yaw);
 670
 671     out->u.m[0][0] = sroll * spitch * syaw + croll * cyaw;
 672     out->u.m[0][1] = sroll * cpitch;
 673     out->u.m[0][2] = sroll * spitch * cyaw - croll * syaw;
 674     out->u.m[0][3] = 0.0f;
 675     out->u.m[1][0] = croll * spitch * syaw - sroll * cyaw;
 676     out->u.m[1][1] = croll * cpitch;
 677     out->u.m[1][2] = croll * spitch * cyaw + sroll * syaw;
 678     out->u.m[1][3] = 0.0f;
 679     out->u.m[2][0] = cpitch * syaw;
 680     out->u.m[2][1] = -spitch;
 681     out->u.m[2][2] = cpitch * cyaw;
 682     out->u.m[2][3] = 0.0f;
 683     out->u.m[3][0] = 0.0f;
 684     out->u.m[3][1] = 0.0f;
 685     out->u.m[3][2] = 0.0f;
 686     out->u.m[3][3] = 1.0f;
 687
 688     return out;
 689 }
 690
 691 D3DXMATRIX* WINAPI D3DXMatrixRotationZ(D3DXMATRIX *pout, FLOAT angle)
 692 {
 693     TRACE("pout %p, angle %f\n", pout, angle);
 694
 695     D3DXMatrixIdentity(pout);
 696     pout->u.m[0][0] = cosf(angle);
 697     pout->u.m[1][1] = cosf(angle);
 698     pout->u.m[0][1] = sinf(angle);
 699     pout->u.m[1][0] = -sinf(angle);
 700     return pout;
 701 }
 702
 703 D3DXMATRIX* WINAPI D3DXMatrixScaling(D3DXMATRIX *pout, FLOAT sx, FLOAT sy, FLOAT sz)
 704 {
 705     TRACE("pout %p, sx %f, sy %f, sz %f\n", pout, sx, sy, sz);
 706
 707     D3DXMatrixIdentity(pout);
 708     pout->u.m[0][0] = sx;
 709     pout->u.m[1][1] = sy;
 710     pout->u.m[2][2] = sz;
 711     return pout;
 712 }
 713
 714 D3DXMATRIX* WINAPI D3DXMatrixShadow(D3DXMATRIX *pout, const D3DXVECTOR4 *plight, const D3DXPLANE *pplane)
 715 {
 716     D3DXPLANE Nplane;
 717     FLOAT dot;
 718
 719     TRACE("pout %p, plight %p, pplane %p\n", pout, plight, pplane);
 720
 721     D3DXPlaneNormalize(&Nplane, pplane);
 722     dot = D3DXPlaneDot(&Nplane, plight);
 723     pout->u.m[0][0] = dot - Nplane.a * plight->x;
 724     pout->u.m[0][1] = -Nplane.a * plight->y;
 725     pout->u.m[0][2] = -Nplane.a * plight->z;
 726     pout->u.m[0][3] = -Nplane.a * plight->w;
 727     pout->u.m[1][0] = -Nplane.b * plight->x;
 728     pout->u.m[1][1] = dot - Nplane.b * plight->y;
 729     pout->u.m[1][2] = -Nplane.b * plight->z;
 730     pout->u.m[1][3] = -Nplane.b * plight->w;
 731     pout->u.m[2][0] = -Nplane.c * plight->x;
 732     pout->u.m[2][1] = -Nplane.c * plight->y;
 733     pout->u.m[2][2] = dot - Nplane.c * plight->z;
 734     pout->u.m[2][3] = -Nplane.c * plight->w;
 735     pout->u.m[3][0] = -Nplane.d * plight->x;
 736     pout->u.m[3][1] = -Nplane.d * plight->y;
 737     pout->u.m[3][2] = -Nplane.d * plight->z;
 738     pout->u.m[3][3] = dot - Nplane.d * plight->w;
 739     return pout;
 740 }
 741
 742 D3DXMATRIX* WINAPI D3DXMatrixTransformation(D3DXMATRIX *pout, const D3DXVECTOR3 *pscalingcenter, const D3DXQUATERNION *pscalingrotation, const D3DXVECTOR3 *pscaling, const D3DXVECTOR3 *protationcenter, const D3DXQUATERNION *protation, const D3DXVECTOR3 *ptranslation)
 743 {
 744     D3DXMATRIX m1, m2, m3, m4, m5, m6, m7;
 745     D3DXQUATERNION prc;
 746     D3DXVECTOR3 psc, pt;
 747
 748     TRACE("pout %p, pscalingcenter %p, pscalingrotation %p, pscaling %p, protationcentr %p, protation %p, ptranslation %p\n",
 749         pout, pscalingcenter, pscalingrotation, pscaling, protationcenter, protation, ptranslation);
 750
 751     if ( !pscalingcenter )
 752     {
 753         psc.x = 0.0f;
 754         psc.y = 0.0f;
 755         psc.z = 0.0f;
 756     }
 757     else
 758     {
 759         psc.x = pscalingcenter->x;
 760         psc.y = pscalingcenter->y;
 761         psc.z = pscalingcenter->z;
 762     }
 763
 764     if ( !protationcenter )
 765     {
 766         prc.x = 0.0f;
 767         prc.y = 0.0f;
 768         prc.z = 0.0f;
 769     }
 770     else
 771     {
 772         prc.x = protationcenter->x;
 773         prc.y = protationcenter->y;
 774         prc.z = protationcenter->z;
 775     }
 776
 777     if ( !ptranslation )
 778     {
 779         pt.x = 0.0f;
 780         pt.y = 0.0f;
 781         pt.z = 0.0f;
 782     }
 783     else
 784     {
 785         pt.x = ptranslation->x;
 786         pt.y = ptranslation->y;
 787         pt.z = ptranslation->z;
 788     }
 789
 790     D3DXMatrixTranslation(&m1, -psc.x, -psc.y, -psc.z);
 791
 792     if ( !pscalingrotation )
 793     {
 794         D3DXMatrixIdentity(&m2);
 795         D3DXMatrixIdentity(&m4);
 796     }
 797     else
 798     {
 799         D3DXMatrixRotationQuaternion(&m4, pscalingrotation);
 800         D3DXMatrixInverse(&m2, NULL, &m4);
 801     }
 802
 803     if ( !pscaling ) D3DXMatrixIdentity(&m3);
 804     else D3DXMatrixScaling(&m3, pscaling->x, pscaling->y, pscaling->z);
 805
 806     if ( !protation ) D3DXMatrixIdentity(&m6);
 807     else D3DXMatrixRotationQuaternion(&m6, protation);
 808
 809     D3DXMatrixTranslation(&m5, psc.x - prc.x,  psc.y - prc.y,  psc.z - prc.z);
 810     D3DXMatrixTranslation(&m7, prc.x + pt.x, prc.y + pt.y, prc.z + pt.z);
 811     D3DXMatrixMultiply(&m1, &m1, &m2);
 812     D3DXMatrixMultiply(&m1, &m1, &m3);
 813     D3DXMatrixMultiply(&m1, &m1, &m4);
 814     D3DXMatrixMultiply(&m1, &m1, &m5);
 815     D3DXMatrixMultiply(&m1, &m1, &m6);
 816     D3DXMatrixMultiply(pout, &m1, &m7);
 817     return pout;
 818 }
 819
 820 D3DXMATRIX* WINAPI D3DXMatrixTransformation2D(D3DXMATRIX *pout, const D3DXVECTOR2 *pscalingcenter, FLOAT scalingrotation, const D3DXVECTOR2 *pscaling, const D3DXVECTOR2 *protationcenter, FLOAT rotation, const D3DXVECTOR2 *ptranslation)
 821 {
 822     D3DXQUATERNION rot, sca_rot;
 823     D3DXVECTOR3 rot_center, sca, sca_center, trans;
 824
 825     TRACE("pout %p, pscalingcenter %p, scalingrotation %f, pscaling %p, protztioncenter %p, rotation %f, ptranslation %p\n",
 826         pout, pscalingcenter, scalingrotation, pscaling, protationcenter, rotation, ptranslation);
 827
 828     if ( pscalingcenter )
 829     {
 830         sca_center.x=pscalingcenter->x;
 831         sca_center.y=pscalingcenter->y;
 832         sca_center.z=0.0f;
 833     }
 834     else
 835     {
 836         sca_center.x=0.0f;
 837         sca_center.y=0.0f;
 838         sca_center.z=0.0f;
 839     }
 840
 841     if ( pscaling )
 842     {
 843         sca.x=pscaling->x;
 844         sca.y=pscaling->y;
 845         sca.z=1.0f;
 846     }
 847     else
 848     {
 849         sca.x=1.0f;
 850         sca.y=1.0f;
 851         sca.z=1.0f;
 852     }
 853
 854     if ( protationcenter )
 855     {
 856         rot_center.x=protationcenter->x;
 857         rot_center.y=protationcenter->y;
 858         rot_center.z=0.0f;
 859     }
 860     else
 861     {
 862         rot_center.x=0.0f;
 863         rot_center.y=0.0f;
 864         rot_center.z=0.0f;
 865     }
 866
 867     if ( ptranslation )
 868     {
 869         trans.x=ptranslation->x;
 870         trans.y=ptranslation->y;
 871         trans.z=0.0f;
 872     }
 873     else
 874     {
 875         trans.x=0.0f;
 876         trans.y=0.0f;
 877         trans.z=0.0f;
 878     }
 879
 880     rot.w=cosf(rotation/2.0f);
 881     rot.x=0.0f;
 882     rot.y=0.0f;
 883     rot.z=sinf(rotation/2.0f);
 884
 885     sca_rot.w=cosf(scalingrotation/2.0f);
 886     sca_rot.x=0.0f;
 887     sca_rot.y=0.0f;
 888     sca_rot.z=sinf(scalingrotation/2.0f);
 889
 890     D3DXMatrixTransformation(pout, &sca_center, &sca_rot, &sca, &rot_center, &rot, &trans);
 891
 892     return pout;
 893 }
 894
 895 D3DXMATRIX* WINAPI D3DXMatrixTranslation(D3DXMATRIX *pout, FLOAT x, FLOAT y, FLOAT z)
 896 {
 897     TRACE("pout %p, x %f, y %f, z %f\n", pout, x, y, z);
 898
 899     D3DXMatrixIdentity(pout);
 900     pout->u.m[3][0] = x;
 901     pout->u.m[3][1] = y;
 902     pout->u.m[3][2] = z;
 903     return pout;
 904 }
 905
 906 D3DXMATRIX* WINAPI D3DXMatrixTranspose(D3DXMATRIX *pout, const D3DXMATRIX *pm)
 907 {
 908     const D3DXMATRIX m = *pm;
 909     int i,j;
 910
 911     TRACE("pout %p, pm %p\n", pout, pm);
 912
 913     for (i=0; i<4; i++)
 914         for (j=0; j<4; j++) pout->u.m[i][j] = m.u.m[j][i];
 915
 916     return pout;
 917 }
 918
 919 /*_________________D3DXMatrixStack____________________*/
 920
 921
 922 static inline struct ID3DXMatrixStackImpl *impl_from_ID3DXMatrixStack(ID3DXMatrixStack *iface)
 923 {
 924   return CONTAINING_RECORD(iface, struct ID3DXMatrixStackImpl, ID3DXMatrixStack_iface);
 925 }
 926
 927 static HRESULT WINAPI ID3DXMatrixStackImpl_QueryInterface(ID3DXMatrixStack *iface, REFIID riid, void **out)
 928 {
 929     TRACE("iface %p, riid %s, out %p.\n", iface, debugstr_guid(riid), out);
 930
 931     if (IsEqualGUID(riid, &IID_ID3DXMatrixStack)
 932             || IsEqualGUID(riid, &IID_IUnknown))
 933     {
 934         ID3DXMatrixStack_AddRef(iface);
 935         *out = iface;
 936         return S_OK;
 937     }
 938
 939     WARN("%s not implemented, returning E_NOINTERFACE.\n", debugstr_guid(riid));
 940
 941     *out = NULL;
 942     return E_NOINTERFACE;
 943 }
 944
 945 static ULONG WINAPI ID3DXMatrixStackImpl_AddRef(ID3DXMatrixStack *iface)
 946 {
 947     struct ID3DXMatrixStackImpl *This = impl_from_ID3DXMatrixStack(iface);
 948     ULONG ref = InterlockedIncrement(&This->ref);
 949     TRACE("(%p) : AddRef from %d\n", This, ref - 1);
 950     return ref;
 951 }
 952
 953 static ULONG WINAPI ID3DXMatrixStackImpl_Release(ID3DXMatrixStack *iface)
 954 {
 955     struct ID3DXMatrixStackImpl *This = impl_from_ID3DXMatrixStack(iface);
 956     ULONG ref = InterlockedDecrement(&This->ref);
 957     if (!ref)
 958     {
 959         HeapFree(GetProcessHeap(), 0, This->stack);
 960         HeapFree(GetProcessHeap(), 0, This);
 961     }
 962     TRACE("(%p) : ReleaseRef to %d\n", This, ref);
 963     return ref;
 964 }
 965
 966 static D3DXMATRIX* WINAPI ID3DXMatrixStackImpl_GetTop(ID3DXMatrixStack *iface)
 967 {
 968     struct ID3DXMatrixStackImpl *This = impl_from_ID3DXMatrixStack(iface);
 969
 970     TRACE("iface %p\n", iface);
 971
 972     return &This->stack[This->current];
 973 }
 974
 975 static HRESULT WINAPI ID3DXMatrixStackImpl_LoadIdentity(ID3DXMatrixStack *iface)
 976 {
 977     struct ID3DXMatrixStackImpl *This = impl_from_ID3DXMatrixStack(iface);
 978
 979     TRACE("iface %p\n", iface);
 980
 981     D3DXMatrixIdentity(&This->stack[This->current]);
 982
 983     return D3D_OK;
 984 }
 985
 986 static HRESULT WINAPI ID3DXMatrixStackImpl_LoadMatrix(ID3DXMatrixStack *iface, const D3DXMATRIX *pm)
 987 {
 988     struct ID3DXMatrixStackImpl *This = impl_from_ID3DXMatrixStack(iface);
 989
 990     TRACE("iface %p, pm %p\n", iface, pm);
 991
 992     This->stack[This->current] = *pm;
 993
 994     return D3D_OK;
 995 }
 996
 997 static HRESULT WINAPI ID3DXMatrixStackImpl_MultMatrix(ID3DXMatrixStack *iface, const D3DXMATRIX *pm)
 998 {
 999     struct ID3DXMatrixStackImpl *This = impl_from_ID3DXMatrixStack(iface);
1000
1001     TRACE("iface %p, pm %p\n", iface, pm);
1002
1003     D3DXMatrixMultiply(&This->stack[This->current], &This->stack[This->current], pm);
1004
1005     return D3D_OK;
1006 }
1007
1008 static HRESULT WINAPI ID3DXMatrixStackImpl_MultMatrixLocal(ID3DXMatrixStack *iface, const D3DXMATRIX *pm)
1009 {
1010     struct ID3DXMatrixStackImpl *This = impl_from_ID3DXMatrixStack(iface);
1011
1012     TRACE("iface %p, pm %p\n", iface, pm);
1013
1014     D3DXMatrixMultiply(&This->stack[This->current], pm, &This->stack[This->current]);
1015
1016     return D3D_OK;
1017 }
1018
1019 static HRESULT WINAPI ID3DXMatrixStackImpl_Pop(ID3DXMatrixStack *iface)
1020 {
1021     struct ID3DXMatrixStackImpl *This = impl_from_ID3DXMatrixStack(iface);
1022
1023     TRACE("iface %p\n", iface);
1024
1025     /* Popping the last element on the stack returns D3D_OK, but does nothing. */
1026     if (!This->current) return D3D_OK;
1027
1028     if (This->current <= This->stack_size / 4 && This->stack_size >= INITIAL_STACK_SIZE * 2)
1029     {
1030         unsigned int new_size;
1031         D3DXMATRIX *new_stack;
1032
1033         new_size = This->stack_size / 2;
1034         new_stack = HeapReAlloc(GetProcessHeap(), 0, This->stack, new_size * sizeof(*new_stack));
1035         if (new_stack)
1036         {
1037             This->stack_size = new_size;
1038             This->stack = new_stack;
1039         }
1040     }
1041
1042     --This->current;
1043
1044     return D3D_OK;
1045 }
1046
1047 static HRESULT WINAPI ID3DXMatrixStackImpl_Push(ID3DXMatrixStack *iface)
1048 {
1049     struct ID3DXMatrixStackImpl *This = impl_from_ID3DXMatrixStack(iface);
1050
1051     TRACE("iface %p\n", iface);
1052
1053     if (This->current == This->stack_size - 1)
1054     {
1055         unsigned int new_size;
1056         D3DXMATRIX *new_stack;
1057
1058         if (This->stack_size > UINT_MAX / 2) return E_OUTOFMEMORY;
1059
1060         new_size = This->stack_size * 2;
1061         new_stack = HeapReAlloc(GetProcessHeap(), 0, This->stack, new_size * sizeof(*new_stack));
1062         if (!new_stack) return E_OUTOFMEMORY;
1063
1064         This->stack_size = new_size;
1065         This->stack = new_stack;
1066     }
1067
1068     ++This->current;
1069     This->stack[This->current] = This->stack[This->current - 1];
1070
1071     return D3D_OK;
1072 }
1073
1074 static HRESULT WINAPI ID3DXMatrixStackImpl_RotateAxis(ID3DXMatrixStack *iface, const D3DXVECTOR3 *pv, FLOAT angle)
1075 {
1076     D3DXMATRIX temp;
1077     struct ID3DXMatrixStackImpl *This = impl_from_ID3DXMatrixStack(iface);
1078
1079     TRACE("iface %p, pv %p, angle %f\n", iface, pv, angle);
1080
1081     D3DXMatrixRotationAxis(&temp, pv, angle);
1082     D3DXMatrixMultiply(&This->stack[This->current], &This->stack[This->current], &temp);
1083
1084     return D3D_OK;
1085 }
1086
1087 static HRESULT WINAPI ID3DXMatrixStackImpl_RotateAxisLocal(ID3DXMatrixStack *iface, const D3DXVECTOR3 *pv, FLOAT angle)
1088 {
1089     D3DXMATRIX temp;
1090     struct ID3DXMatrixStackImpl *This = impl_from_ID3DXMatrixStack(iface);
1091
1092     TRACE("iface %p, pv %p, angle %f\n", iface, pv, angle);
1093
1094     D3DXMatrixRotationAxis(&temp, pv, angle);
1095     D3DXMatrixMultiply(&This->stack[This->current], &temp, &This->stack[This->current]);
1096
1097     return D3D_OK;
1098 }
1099
1100 static HRESULT WINAPI ID3DXMatrixStackImpl_RotateYawPitchRoll(ID3DXMatrixStack *iface, FLOAT x, FLOAT y, FLOAT z)
1101 {
1102     D3DXMATRIX temp;
1103     struct ID3DXMatrixStackImpl *This = impl_from_ID3DXMatrixStack(iface);
1104
1105     TRACE("iface %p, x %f, y %f, z %f\n", iface, x, y, z);
1106
1107     D3DXMatrixRotationYawPitchRoll(&temp, x, y, z);
1108     D3DXMatrixMultiply(&This->stack[This->current], &This->stack[This->current], &temp);
1109
1110     return D3D_OK;
1111 }
1112
1113 static HRESULT WINAPI ID3DXMatrixStackImpl_RotateYawPitchRollLocal(ID3DXMatrixStack *iface, FLOAT x, FLOAT y, FLOAT z)
1114 {
1115     D3DXMATRIX temp;
1116     struct ID3DXMatrixStackImpl *This = impl_from_ID3DXMatrixStack(iface);
1117
1118     TRACE("iface %p, x %f, y %f, z %f\n", iface, x, y, z);
1119
1120     D3DXMatrixRotationYawPitchRoll(&temp, x, y, z);
1121     D3DXMatrixMultiply(&This->stack[This->current], &temp, &This->stack[This->current]);
1122
1123     return D3D_OK;
1124 }
1125
1126 static HRESULT WINAPI ID3DXMatrixStackImpl_Scale(ID3DXMatrixStack *iface, FLOAT x, FLOAT y, FLOAT z)
1127 {
1128     D3DXMATRIX temp;
1129     struct ID3DXMatrixStackImpl *This = impl_from_ID3DXMatrixStack(iface);
1130
1131     TRACE("iface %p,x %f, y %f, z %f\n", iface, x, y, z);
1132
1133     D3DXMatrixScaling(&temp, x, y, z);
1134     D3DXMatrixMultiply(&This->stack[This->current], &This->stack[This->current], &temp);
1135
1136     return D3D_OK;
1137 }
1138
1139 static HRESULT WINAPI ID3DXMatrixStackImpl_ScaleLocal(ID3DXMatrixStack *iface, FLOAT x, FLOAT y, FLOAT z)
1140 {
1141     D3DXMATRIX temp;
1142     struct ID3DXMatrixStackImpl *This = impl_from_ID3DXMatrixStack(iface);
1143
1144     TRACE("iface %p, x %f, y %f, z %f\n", iface, x, y, z);
1145
1146     D3DXMatrixScaling(&temp, x, y, z);
1147     D3DXMatrixMultiply(&This->stack[This->current], &temp, &This->stack[This->current]);
1148
1149     return D3D_OK;
1150 }
1151
1152 static HRESULT WINAPI ID3DXMatrixStackImpl_Translate(ID3DXMatrixStack *iface, FLOAT x, FLOAT y, FLOAT z)
1153 {
1154     D3DXMATRIX temp;
1155     struct ID3DXMatrixStackImpl *This = impl_from_ID3DXMatrixStack(iface);
1156
1157     TRACE("iface %p, x %f, y %f, z %f\n", iface, x, y, z);
1158
1159     D3DXMatrixTranslation(&temp, x, y, z);
1160     D3DXMatrixMultiply(&This->stack[This->current], &This->stack[This->current], &temp);
1161
1162     return D3D_OK;
1163 }
1164
1165 static HRESULT WINAPI ID3DXMatrixStackImpl_TranslateLocal(ID3DXMatrixStack *iface, FLOAT x, FLOAT y, FLOAT z)
1166 {
1167     D3DXMATRIX temp;
1168     struct ID3DXMatrixStackImpl *This = impl_from_ID3DXMatrixStack(iface);
1169
1170     TRACE("iface %p, x %f, y %f, z %f\n", iface, x, y, z);
1171
1172     D3DXMatrixTranslation(&temp, x, y, z);
1173     D3DXMatrixMultiply(&This->stack[This->current], &temp,&This->stack[This->current]);
1174
1175     return D3D_OK;
1176 }
1177
1178 static const ID3DXMatrixStackVtbl ID3DXMatrixStack_Vtbl =
1179 {
1180     ID3DXMatrixStackImpl_QueryInterface,
1181     ID3DXMatrixStackImpl_AddRef,
1182     ID3DXMatrixStackImpl_Release,
1183     ID3DXMatrixStackImpl_Pop,
1184     ID3DXMatrixStackImpl_Push,
1185     ID3DXMatrixStackImpl_LoadIdentity,
1186     ID3DXMatrixStackImpl_LoadMatrix,
1187     ID3DXMatrixStackImpl_MultMatrix,
1188     ID3DXMatrixStackImpl_MultMatrixLocal,
1189     ID3DXMatrixStackImpl_RotateAxis,
1190     ID3DXMatrixStackImpl_RotateAxisLocal,
1191     ID3DXMatrixStackImpl_RotateYawPitchRoll,
1192     ID3DXMatrixStackImpl_RotateYawPitchRollLocal,
1193     ID3DXMatrixStackImpl_Scale,
1194     ID3DXMatrixStackImpl_ScaleLocal,
1195     ID3DXMatrixStackImpl_Translate,
1196     ID3DXMatrixStackImpl_TranslateLocal,
1197     ID3DXMatrixStackImpl_GetTop
1198 };
1199
1200 HRESULT WINAPI D3DXCreateMatrixStack(DWORD flags, ID3DXMatrixStack **stack)
1201 {
1202     struct ID3DXMatrixStackImpl *object;
1203
1204     TRACE("flags %#x, stack %p.\n", flags, stack);
1205
1206     if (!(object = HeapAlloc(GetProcessHeap(), HEAP_ZERO_MEMORY, sizeof(*object))))
1207     {
1208         *stack = NULL;
1209         return E_OUTOFMEMORY;
1210     }
1211     object->ID3DXMatrixStack_iface.lpVtbl = &ID3DXMatrixStack_Vtbl;
1212     object->ref = 1;
1213
1214     if (!(object->stack = HeapAlloc(GetProcessHeap(), 0, INITIAL_STACK_SIZE * sizeof(*object->stack))))
1215     {
1216         HeapFree(GetProcessHeap(), 0, object);
1217         *stack = NULL;
1218         return E_OUTOFMEMORY;
1219     }
1220
1221     object->current = 0;
1222     object->stack_size = INITIAL_STACK_SIZE;
1223     D3DXMatrixIdentity(&object->stack[0]);
1224
1225     TRACE("Created matrix stack %p.\n", object);
1226
1227     *stack = &object->ID3DXMatrixStack_iface;
1228     return D3D_OK;
1229 }
1230
1231 /*_________________D3DXPLANE________________*/
1232
1233 D3DXPLANE* WINAPI D3DXPlaneFromPointNormal(D3DXPLANE *pout, const D3DXVECTOR3 *pvpoint, const D3DXVECTOR3 *pvnormal)
1234 {
1235     TRACE("pout %p, pvpoint %p, pvnormal %p\n", pout, pvpoint, pvnormal);
1236
1237     pout->a = pvnormal->x;
1238     pout->b = pvnormal->y;
1239     pout->c = pvnormal->z;
1240     pout->d = -D3DXVec3Dot(pvpoint, pvnormal);
1241     return pout;
1242 }
1243
1244 D3DXPLANE* WINAPI D3DXPlaneFromPoints(D3DXPLANE *pout, const D3DXVECTOR3 *pv1, const D3DXVECTOR3 *pv2, const D3DXVECTOR3 *pv3)
1245 {
1246     D3DXVECTOR3 edge1, edge2, normal, Nnormal;
1247
1248     TRACE("pout %p, pv1 %p, pv2 %p, pv3 %p\n", pout, pv1, pv2, pv3);
1249
1250     edge1.x = 0.0f; edge1.y = 0.0f; edge1.z = 0.0f;
1251     edge2.x = 0.0f; edge2.y = 0.0f; edge2.z = 0.0f;
1252     D3DXVec3Subtract(&edge1, pv2, pv1);
1253     D3DXVec3Subtract(&edge2, pv3, pv1);
1254     D3DXVec3Cross(&normal, &edge1, &edge2);
1255     D3DXVec3Normalize(&Nnormal, &normal);
1256     D3DXPlaneFromPointNormal(pout, pv1, &Nnormal);
1257     return pout;
1258 }
1259
1260 D3DXVECTOR3* WINAPI D3DXPlaneIntersectLine(D3DXVECTOR3 *pout, const D3DXPLANE *pp, const D3DXVECTOR3 *pv1, const D3DXVECTOR3 *pv2)
1261 {
1262     D3DXVECTOR3 direction, normal;
1263     FLOAT dot, temp;
1264
1265     TRACE("pout %p, pp %p, pv1 %p, pv2 %p\n", pout, pp, pv1, pv2);
1266
1267     normal.x = pp->a;
1268     normal.y = pp->b;
1269     normal.z = pp->c;
1270     direction.x = pv2->x - pv1->x;
1271     direction.y = pv2->y - pv1->y;
1272     direction.z = pv2->z - pv1->z;
1273     dot = D3DXVec3Dot(&normal, &direction);
1274     if ( !dot ) return NULL;
1275     temp = ( pp->d + D3DXVec3Dot(&normal, pv1) ) / dot;
1276     pout->x = pv1->x - temp * direction.x;
1277     pout->y = pv1->y - temp * direction.y;
1278     pout->z = pv1->z - temp * direction.z;
1279     return pout;
1280 }
1281
1282 D3DXPLANE * WINAPI D3DXPlaneNormalize(D3DXPLANE *out, const D3DXPLANE *p)
1283 {
1284     FLOAT norm;
1285
1286     TRACE("out %p, p %p\n", out, p);
1287
1288     norm = sqrtf(p->a * p->a + p->b * p->b + p->c * p->c);
1289     if (norm)
1290     {
1291         out->a = p->a / norm;
1292         out->b = p->b / norm;
1293         out->c = p->c / norm;
1294         out->d = p->d / norm;
1295     }
1296     else
1297     {
1298         out->a = 0.0f;
1299         out->b = 0.0f;
1300         out->c = 0.0f;
1301         out->d = 0.0f;
1302     }
1303
1304     return out;
1305 }
1306
1307 D3DXPLANE* WINAPI D3DXPlaneTransform(D3DXPLANE *pout, const D3DXPLANE *pplane, const D3DXMATRIX *pm)
1308 {
1309     const D3DXPLANE plane = *pplane;
1310
1311     TRACE("pout %p, pplane %p, pm %p\n", pout, pplane, pm);
1312
1313     pout->a = pm->u.m[0][0] * plane.a + pm->u.m[1][0] * plane.b + pm->u.m[2][0] * plane.c + pm->u.m[3][0] * plane.d;
1314     pout->b = pm->u.m[0][1] * plane.a + pm->u.m[1][1] * plane.b + pm->u.m[2][1] * plane.c + pm->u.m[3][1] * plane.d;
1315     pout->c = pm->u.m[0][2] * plane.a + pm->u.m[1][2] * plane.b + pm->u.m[2][2] * plane.c + pm->u.m[3][2] * plane.d;
1316     pout->d = pm->u.m[0][3] * plane.a + pm->u.m[1][3] * plane.b + pm->u.m[2][3] * plane.c + pm->u.m[3][3] * plane.d;
1317     return pout;
1318 }
1319
1320 D3DXPLANE* WINAPI D3DXPlaneTransformArray(D3DXPLANE* out, UINT outstride, const D3DXPLANE* in, UINT instride, const D3DXMATRIX* matrix, UINT elements)
1321 {
1322     UINT i;
1323
1324     TRACE("out %p, outstride %u, in %p, instride %u, matrix %p, elements %u\n", out, outstride, in, instride, matrix, elements);
1325
1326     for (i = 0; i < elements; ++i) {
1327         D3DXPlaneTransform(
1328             (D3DXPLANE*)((char*)out + outstride * i),
1329             (const D3DXPLANE*)((const char*)in + instride * i),
1330             matrix);
1331     }
1332     return out;
1333 }
1334
1335 /*_________________D3DXQUATERNION________________*/
1336
1337 D3DXQUATERNION* WINAPI D3DXQuaternionBaryCentric(D3DXQUATERNION *pout, const D3DXQUATERNION *pq1, const D3DXQUATERNION *pq2, const D3DXQUATERNION *pq3, FLOAT f, FLOAT g)
1338 {
1339     D3DXQUATERNION temp1, temp2;
1340
1341      TRACE("pout %p, pq1 %p, pq2 %p, pq3 %p, f %f, g %f\n", pout, pq1, pq2, pq3, f, g);
1342
1343     D3DXQuaternionSlerp(pout, D3DXQuaternionSlerp(&temp1, pq1, pq2, f + g), D3DXQuaternionSlerp(&temp2, pq1, pq3, f+g), g / (f + g));
1344     return pout;
1345 }
1346
1347 D3DXQUATERNION * WINAPI D3DXQuaternionExp(D3DXQUATERNION *out, const D3DXQUATERNION *q)
1348 {
1349     FLOAT norm;
1350
1351     TRACE("out %p, q %p\n", out, q);
1352
1353     norm = sqrtf(q->x * q->x + q->y * q->y + q->z * q->z);
1354     if (norm)
1355     {
1356         out->x = sinf(norm) * q->x / norm;
1357         out->y = sinf(norm) * q->y / norm;
1358         out->z = sinf(norm) * q->z / norm;
1359         out->w = cosf(norm);
1360     }
1361     else
1362     {
1363         out->x = 0.0f;
1364         out->y = 0.0f;
1365         out->z = 0.0f;
1366         out->w = 1.0f;
1367     }
1368
1369     return out;
1370 }
1371
1372 D3DXQUATERNION* WINAPI D3DXQuaternionInverse(D3DXQUATERNION *pout, const D3DXQUATERNION *pq)
1373 {
1374     D3DXQUATERNION out;
1375     FLOAT norm;
1376
1377     TRACE("pout %p, pq %p\n", pout, pq);
1378
1379     norm = D3DXQuaternionLengthSq(pq);
1380
1381     out.x = -pq->x / norm;
1382     out.y = -pq->y / norm;
1383     out.z = -pq->z / norm;
1384     out.w = pq->w / norm;
1385
1386     *pout =out;
1387     return pout;
1388 }
1389
1390 D3DXQUATERNION * WINAPI D3DXQuaternionLn(D3DXQUATERNION *out, const D3DXQUATERNION *q)
1391 {
1392     FLOAT t;
1393
1394     TRACE("out %p, q %p\n", out, q);
1395
1396     if ((q->w >= 1.0f) || (q->w == -1.0f))
1397         t = 1.0f;
1398     else
1399         t = acosf(q->w) / sqrtf(1.0f - q->w * q->w);
1400
1401     out->x = t * q->x;
1402     out->y = t * q->y;
1403     out->z = t * q->z;
1404     out->w = 0.0f;
1405
1406     return out;
1407 }
1408
1409 D3DXQUATERNION* WINAPI D3DXQuaternionMultiply(D3DXQUATERNION *pout, const D3DXQUATERNION *pq1, const D3DXQUATERNION *pq2)
1410 {
1411     D3DXQUATERNION out;
1412
1413     TRACE("pout %p, pq1 %p, pq2 %p\n", pout, pq1, pq2);
1414
1415     out.x = pq2->w * pq1->x + pq2->x * pq1->w + pq2->y * pq1->z - pq2->z * pq1->y;
1416     out.y = pq2->w * pq1->y - pq2->x * pq1->z + pq2->y * pq1->w + pq2->z * pq1->x;
1417     out.z = pq2->w * pq1->z + pq2->x * pq1->y - pq2->y * pq1->x + pq2->z * pq1->w;
1418     out.w = pq2->w * pq1->w - pq2->x * pq1->x - pq2->y * pq1->y - pq2->z * pq1->z;
1419     *pout = out;
1420     return pout;
1421 }
1422
1423 D3DXQUATERNION * WINAPI D3DXQuaternionNormalize(D3DXQUATERNION *out, const D3DXQUATERNION *q)
1424 {
1425     FLOAT norm;
1426
1427     TRACE("out %p, q %p\n", out, q);
1428
1429     norm = D3DXQuaternionLength(q);
1430
1431     out->x = q->x / norm;
1432     out->y = q->y / norm;
1433     out->z = q->z / norm;
1434     out->w = q->w / norm;
1435
1436     return out;
1437 }
1438
1439 D3DXQUATERNION * WINAPI D3DXQuaternionRotationAxis(D3DXQUATERNION *out, const D3DXVECTOR3 *v, FLOAT angle)
1440 {
1441     D3DXVECTOR3 temp;
1442
1443     TRACE("out %p, v %p, angle %f\n", out, v, angle);
1444
1445     D3DXVec3Normalize(&temp, v);
1446
1447     out->x = sinf(angle / 2.0f) * temp.x;
1448     out->y = sinf(angle / 2.0f) * temp.y;
1449     out->z = sinf(angle / 2.0f) * temp.z;
1450     out->w = cosf(angle / 2.0f);
1451
1452     return out;
1453 }
1454
1455 D3DXQUATERNION * WINAPI D3DXQuaternionRotationMatrix(D3DXQUATERNION *out, const D3DXMATRIX *m)
1456 {
1457     FLOAT s, trace;
1458
1459     TRACE("out %p, m %p\n", out, m);
1460
1461     trace = m->u.m[0][0] + m->u.m[1][1] + m->u.m[2][2] + 1.0f;
1462     if (trace > 1.0f)
1463     {
1464         s = 2.0f * sqrtf(trace);
1465         out->x = (m->u.m[1][2] - m->u.m[2][1]) / s;
1466         out->y = (m->u.m[2][0] - m->u.m[0][2]) / s;
1467         out->z = (m->u.m[0][1] - m->u.m[1][0]) / s;
1468         out->w = 0.25f * s;
1469     }
1470     else
1471     {
1472         int i, maxi = 0;
1473
1474         for (i = 1; i < 3; i++)
1475         {
1476             if (m->u.m[i][i] > m->u.m[maxi][maxi])
1477                 maxi = i;
1478         }
1479
1480         switch (maxi)
1481         {
1482             case 0:
1483                 s = 2.0f * sqrtf(1.0f + m->u.m[0][0] - m->u.m[1][1] - m->u.m[2][2]);
1484                 out->x = 0.25f * s;
1485                 out->y = (m->u.m[0][1] + m->u.m[1][0]) / s;
1486                 out->z = (m->u.m[0][2] + m->u.m[2][0]) / s;
1487                 out->w = (m->u.m[1][2] - m->u.m[2][1]) / s;
1488                 break;
1489
1490             case 1:
1491                 s = 2.0f * sqrtf(1.0f + m->u.m[1][1] - m->u.m[0][0] - m->u.m[2][2]);
1492                 out->x = (m->u.m[0][1] + m->u.m[1][0]) / s;
1493                 out->y = 0.25f * s;
1494                 out->z = (m->u.m[1][2] + m->u.m[2][1]) / s;
1495                 out->w = (m->u.m[2][0] - m->u.m[0][2]) / s;
1496                 break;
1497
1498             case 2:
1499                 s = 2.0f * sqrtf(1.0f + m->u.m[2][2] - m->u.m[0][0] - m->u.m[1][1]);
1500                 out->x = (m->u.m[0][2] + m->u.m[2][0]) / s;
1501                 out->y = (m->u.m[1][2] + m->u.m[2][1]) / s;
1502                 out->z = 0.25f * s;
1503                 out->w = (m->u.m[0][1] - m->u.m[1][0]) / s;
1504                 break;
1505         }
1506     }
1507
1508     return out;
1509 }
1510
1511 D3DXQUATERNION * WINAPI D3DXQuaternionRotationYawPitchRoll(D3DXQUATERNION *out, FLOAT yaw, FLOAT pitch, FLOAT roll)
1512 {
1513     FLOAT syaw, cyaw, spitch, cpitch, sroll, croll;
1514
1515     TRACE("out %p, yaw %f, pitch %f, roll %f\n", out, yaw, pitch, roll);
1516
1517     syaw = sinf(yaw / 2.0f);
1518     cyaw = cosf(yaw / 2.0f);
1519     spitch = sinf(pitch / 2.0f);
1520     cpitch = cosf(pitch / 2.0f);
1521     sroll = sinf(roll / 2.0f);
1522     croll = cosf(roll / 2.0f);
1523
1524     out->x = syaw * cpitch * sroll + cyaw * spitch * croll;
1525     out->y = syaw * cpitch * croll - cyaw * spitch * sroll;
1526     out->z = cyaw * cpitch * sroll - syaw * spitch * croll;
1527     out->w = cyaw * cpitch * croll + syaw * spitch * sroll;
1528
1529     return out;
1530 }
1531
1532 D3DXQUATERNION * WINAPI D3DXQuaternionSlerp(D3DXQUATERNION *out, const D3DXQUATERNION *q1,
1533         const D3DXQUATERNION *q2, FLOAT t)
1534 {
1535     FLOAT dot, temp;
1536
1537     TRACE("out %p, q1 %p, q2 %p, t %f\n", out, q1, q2, t);
1538
1539     temp = 1.0f - t;
1540     dot = D3DXQuaternionDot(q1, q2);
1541     if (dot < 0.0f)
1542     {
1543         t = -t;
1544         dot = -dot;
1545     }
1546
1547     if (1.0f - dot > 0.001f)
1548     {
1549         FLOAT theta = acosf(dot);
1550
1551         temp = sinf(theta * temp) / sinf(theta);
1552         t = sinf(theta * t) / sinf(theta);
1553     }
1554
1555     out->x = temp * q1->x + t * q2->x;
1556     out->y = temp * q1->y + t * q2->y;
1557     out->z = temp * q1->z + t * q2->z;
1558     out->w = temp * q1->w + t * q2->w;
1559
1560     return out;
1561 }
1562
1563 D3DXQUATERNION* WINAPI D3DXQuaternionSquad(D3DXQUATERNION *pout, const D3DXQUATERNION *pq1, const D3DXQUATERNION *pq2, const D3DXQUATERNION *pq3, const D3DXQUATERNION *pq4, FLOAT t)
1564 {
1565     D3DXQUATERNION temp1, temp2;
1566
1567     TRACE("pout %p, pq1 %p, pq2 %p, pq3 %p, pq4 %p, t %f\n", pout, pq1, pq2, pq3, pq4, t);
1568
1569     D3DXQuaternionSlerp(pout, D3DXQuaternionSlerp(&temp1, pq1, pq4, t), D3DXQuaternionSlerp(&temp2, pq2, pq3, t), 2.0f * t * (1.0f - t));
1570     return pout;
1571 }
1572
1573 static D3DXQUATERNION add_diff(const D3DXQUATERNION *q1, const D3DXQUATERNION *q2, const FLOAT add)
1574 {
1575     D3DXQUATERNION temp;
1576
1577     temp.x = q1->x + add * q2->x;
1578     temp.y = q1->y + add * q2->y;
1579     temp.z = q1->z + add * q2->z;
1580     temp.w = q1->w + add * q2->w;
1581
1582     return temp;
1583 }
1584
1585 void WINAPI D3DXQuaternionSquadSetup(D3DXQUATERNION *paout, D3DXQUATERNION *pbout, D3DXQUATERNION *pcout, const D3DXQUATERNION *pq0, const D3DXQUATERNION *pq1, const D3DXQUATERNION *pq2, const D3DXQUATERNION *pq3)
1586 {
1587     D3DXQUATERNION q, temp1, temp2, temp3, zero;
1588
1589     TRACE("paout %p, pbout %p, pcout %p, pq0 %p, pq1 %p, pq2 %p, pq3 %p\n", paout, pbout, pcout, pq0, pq1, pq2, pq3);
1590
1591     zero.x = 0.0f;
1592     zero.y = 0.0f;
1593     zero.z = 0.0f;
1594     zero.w = 0.0f;
1595
1596     if ( D3DXQuaternionDot(pq0, pq1) <  0.0f )
1597         temp2 = add_diff(&zero, pq0, -1.0f);
1598     else
1599         temp2 = *pq0;
1600
1601     if ( D3DXQuaternionDot(pq1, pq2) < 0.0f )
1602         *pcout = add_diff(&zero, pq2, -1.0f);
1603     else
1604         *pcout = *pq2;
1605
1606     if ( D3DXQuaternionDot(pcout, pq3) < 0.0f )
1607         temp3 = add_diff(&zero, pq3, -1.0f);
1608     else
1609         temp3 = *pq3;
1610
1611     D3DXQuaternionInverse(&temp1, pq1);
1612     D3DXQuaternionMultiply(&temp2, &temp1, &temp2);
1613     D3DXQuaternionLn(&temp2, &temp2);
1614     D3DXQuaternionMultiply(&q, &temp1, pcout);
1615     D3DXQuaternionLn(&q, &q);
1616     temp1 = add_diff(&temp2, &q, 1.0f);
1617     temp1.x *= -0.25f;
1618     temp1.y *= -0.25f;
1619     temp1.z *= -0.25f;
1620     temp1.w *= -0.25f;
1621     D3DXQuaternionExp(&temp1, &temp1);
1622     D3DXQuaternionMultiply(paout, pq1, &temp1);
1623
1624     D3DXQuaternionInverse(&temp1, pcout);
1625     D3DXQuaternionMultiply(&temp2, &temp1, pq1);
1626     D3DXQuaternionLn(&temp2, &temp2);
1627     D3DXQuaternionMultiply(&q, &temp1, &temp3);
1628     D3DXQuaternionLn(&q, &q);
1629     temp1 = add_diff(&temp2, &q, 1.0f);
1630     temp1.x *= -0.25f;
1631     temp1.y *= -0.25f;
1632     temp1.z *= -0.25f;
1633     temp1.w *= -0.25f;
1634     D3DXQuaternionExp(&temp1, &temp1);
1635     D3DXQuaternionMultiply(pbout, pcout, &temp1);
1636
1637     return;
1638 }
1639
1640 void WINAPI D3DXQuaternionToAxisAngle(const D3DXQUATERNION *pq, D3DXVECTOR3 *paxis, FLOAT *pangle)
1641 {
1642     TRACE("pq %p, paxis %p, pangle %p\n", pq, paxis, pangle);
1643
1644     paxis->x = pq->x;
1645     paxis->y = pq->y;
1646     paxis->z = pq->z;
1647     *pangle = 2.0f * acosf(pq->w);
1648 }
1649
1650 /*_________________D3DXVec2_____________________*/
1651
1652 D3DXVECTOR2* WINAPI D3DXVec2BaryCentric(D3DXVECTOR2 *pout, const D3DXVECTOR2 *pv1, const D3DXVECTOR2 *pv2, const D3DXVECTOR2 *pv3, FLOAT f, FLOAT g)
1653 {
1654     TRACE("pout %p, pv1 %p, pv2 %p, pv3 %p, f %f, g %f\n", pout, pv1, pv2, pv3, f, g);
1655
1656     pout->x = (1.0f-f-g) * (pv1->x) + f * (pv2->x) + g * (pv3->x);
1657     pout->y = (1.0f-f-g) * (pv1->y) + f * (pv2->y) + g * (pv3->y);
1658     return pout;
1659 }
1660
1661 D3DXVECTOR2* WINAPI D3DXVec2CatmullRom(D3DXVECTOR2 *pout, const D3DXVECTOR2 *pv0, const D3DXVECTOR2 *pv1, const D3DXVECTOR2 *pv2, const D3DXVECTOR2 *pv3, FLOAT s)
1662 {
1663     TRACE("pout %p, pv0 %p, pv1 %p, pv2 %p, pv3 %p, s %f\n", pout, pv0, pv1, pv2, pv3, s);
1664
1665     pout->x = 0.5f * (2.0f * pv1->x + (pv2->x - pv0->x) *s + (2.0f *pv0->x - 5.0f * pv1->x + 4.0f * pv2->x - pv3->x) * s * s + (pv3->x -3.0f * pv2->x + 3.0f * pv1->x - pv0->x) * s * s * s);
1666     pout->y = 0.5f * (2.0f * pv1->y + (pv2->y - pv0->y) *s + (2.0f *pv0->y - 5.0f * pv1->y + 4.0f * pv2->y - pv3->y) * s * s + (pv3->y -3.0f * pv2->y + 3.0f * pv1->y - pv0->y) * s * s * s);
1667     return pout;
1668 }
1669
1670 D3DXVECTOR2* WINAPI D3DXVec2Hermite(D3DXVECTOR2 *pout, const D3DXVECTOR2 *pv1, const D3DXVECTOR2 *pt1, const D3DXVECTOR2 *pv2, const D3DXVECTOR2 *pt2, FLOAT s)
1671 {
1672     FLOAT h1, h2, h3, h4;
1673
1674     TRACE("pout %p, pv1 %p, pt1 %p, pv2 %p, pt2 %p, s %f\n", pout, pv1, pt1, pv2, pt2, s);
1675
1676     h1 = 2.0f * s * s * s - 3.0f * s * s + 1.0f;
1677     h2 = s * s * s - 2.0f * s * s + s;
1678     h3 = -2.0f * s * s * s + 3.0f * s * s;
1679     h4 = s * s * s - s * s;
1680
1681     pout->x = h1 * (pv1->x) + h2 * (pt1->x) + h3 * (pv2->x) + h4 * (pt2->x);
1682     pout->y = h1 * (pv1->y) + h2 * (pt1->y) + h3 * (pv2->y) + h4 * (pt2->y);
1683     return pout;
1684 }
1685
1686 D3DXVECTOR2* WINAPI D3DXVec2Normalize(D3DXVECTOR2 *pout, const D3DXVECTOR2 *pv)
1687 {
1688     FLOAT norm;
1689
1690     TRACE("pout %p, pv %p\n", pout, pv);
1691
1692     norm = D3DXVec2Length(pv);
1693     if ( !norm )
1694     {
1695         pout->x = 0.0f;
1696         pout->y = 0.0f;
1697     }
1698     else
1699     {
1700         pout->x = pv->x / norm;
1701         pout->y = pv->y / norm;
1702     }
1703
1704     return pout;
1705 }
1706
1707 D3DXVECTOR4* WINAPI D3DXVec2Transform(D3DXVECTOR4 *pout, const D3DXVECTOR2 *pv, const D3DXMATRIX *pm)
1708 {
1709     TRACE("pout %p, pv %p, pm %p\n", pout, pv, pm);
1710
1711     pout->x = pm->u.m[0][0] * pv->x + pm->u.m[1][0] * pv->y  + pm->u.m[3][0];
1712     pout->y = pm->u.m[0][1] * pv->x + pm->u.m[1][1] * pv->y  + pm->u.m[3][1];
1713     pout->z = pm->u.m[0][2] * pv->x + pm->u.m[1][2] * pv->y  + pm->u.m[3][2];
1714     pout->w = pm->u.m[0][3] * pv->x + pm->u.m[1][3] * pv->y  + pm->u.m[3][3];
1715     return pout;
1716 }
1717
1718 D3DXVECTOR4* WINAPI D3DXVec2TransformArray(D3DXVECTOR4* out, UINT outstride, const D3DXVECTOR2* in, UINT instride, const D3DXMATRIX* matrix, UINT elements)
1719 {
1720     UINT i;
1721
1722     TRACE("out %p, outstride %u, in %p, instride %u, matrix %p, elements %u\n", out, outstride, in, instride, matrix, elements);
1723
1724     for (i = 0; i < elements; ++i) {
1725         D3DXVec2Transform(
1726             (D3DXVECTOR4*)((char*)out + outstride * i),
1727             (const D3DXVECTOR2*)((const char*)in + instride * i),
1728             matrix);
1729     }
1730     return out;
1731 }
1732
1733 D3DXVECTOR2* WINAPI D3DXVec2TransformCoord(D3DXVECTOR2 *pout, const D3DXVECTOR2 *pv, const D3DXMATRIX *pm)
1734 {
1735     D3DXVECTOR2 v;
1736     FLOAT norm;
1737
1738     TRACE("pout %p, pv %p, pm %p\n", pout, pv, pm);
1739
1740     v = *pv;
1741     norm = pm->u.m[0][3] * pv->x + pm->u.m[1][3] * pv->y + pm->u.m[3][3];
1742
1743     pout->x = (pm->u.m[0][0] * v.x + pm->u.m[1][0] * v.y + pm->u.m[3][0]) / norm;
1744     pout->y = (pm->u.m[0][1] * v.x + pm->u.m[1][1] * v.y + pm->u.m[3][1]) / norm;
1745
1746     return pout;
1747 }
1748
1749 D3DXVECTOR2* WINAPI D3DXVec2TransformCoordArray(D3DXVECTOR2* out, UINT outstride, const D3DXVECTOR2* in, UINT instride, const D3DXMATRIX* matrix, UINT elements)
1750 {
1751     UINT i;
1752
1753     TRACE("out %p, outstride %u, in %p, instride %u, matrix %p, elements %u\n", out, outstride, in, instride, matrix, elements);
1754
1755     for (i = 0; i < elements; ++i) {
1756         D3DXVec2TransformCoord(
1757             (D3DXVECTOR2*)((char*)out + outstride * i),
1758             (const D3DXVECTOR2*)((const char*)in + instride * i),
1759             matrix);
1760     }
1761     return out;
1762 }
1763
1764 D3DXVECTOR2* WINAPI D3DXVec2TransformNormal(D3DXVECTOR2 *pout, const D3DXVECTOR2 *pv, const D3DXMATRIX *pm)
1765 {
1766     const D3DXVECTOR2 v = *pv;
1767
1768     TRACE("pout %p, pv %p, pm %p\n", pout, pv, pm);
1769
1770     pout->x = pm->u.m[0][0] * v.x + pm->u.m[1][0] * v.y;
1771     pout->y = pm->u.m[0][1] * v.x + pm->u.m[1][1] * v.y;
1772     return pout;
1773 }
1774
1775 D3DXVECTOR2* WINAPI D3DXVec2TransformNormalArray(D3DXVECTOR2* out, UINT outstride, const D3DXVECTOR2 *in, UINT instride, const D3DXMATRIX *matrix, UINT elements)
1776 {
1777     UINT i;
1778
1779     TRACE("out %p, outstride %u, in %p, instride %u, matrix %p, elements %u\n", out, outstride, in, instride, matrix, elements);
1780
1781     for (i = 0; i < elements; ++i) {
1782         D3DXVec2TransformNormal(
1783             (D3DXVECTOR2*)((char*)out + outstride * i),
1784             (const D3DXVECTOR2*)((const char*)in + instride * i),
1785             matrix);
1786     }
1787     return out;
1788 }
1789
1790 /*_________________D3DXVec3_____________________*/
1791
1792 D3DXVECTOR3* WINAPI D3DXVec3BaryCentric(D3DXVECTOR3 *pout, const D3DXVECTOR3 *pv1, const D3DXVECTOR3 *pv2, const D3DXVECTOR3 *pv3, FLOAT f, FLOAT g)
1793 {
1794     TRACE("pout %p, pv1 %p, pv2 %p, pv3 %p, f %f, g %f\n", pout, pv1, pv2, pv3, f, g);
1795
1796     pout->x = (1.0f-f-g) * (pv1->x) + f * (pv2->x) + g * (pv3->x);
1797     pout->y = (1.0f-f-g) * (pv1->y) + f * (pv2->y) + g * (pv3->y);
1798     pout->z = (1.0f-f-g) * (pv1->z) + f * (pv2->z) + g * (pv3->z);
1799     return pout;
1800 }
1801
1802 D3DXVECTOR3* WINAPI D3DXVec3CatmullRom( D3DXVECTOR3 *pout, const D3DXVECTOR3 *pv0, const D3DXVECTOR3 *pv1, const D3DXVECTOR3 *pv2, const D3DXVECTOR3 *pv3, FLOAT s)
1803 {
1804     TRACE("pout %p, pv0 %p, pv1 %p, pv2 %p, pv3 %p, s %f\n", pout, pv0, pv1, pv2, pv3, s);
1805
1806     pout->x = 0.5f * (2.0f * pv1->x + (pv2->x - pv0->x) *s + (2.0f *pv0->x - 5.0f * pv1->x + 4.0f * pv2->x - pv3->x) * s * s + (pv3->x -3.0f * pv2->x + 3.0f * pv1->x - pv0->x) * s * s * s);
1807     pout->y = 0.5f * (2.0f * pv1->y + (pv2->y - pv0->y) *s + (2.0f *pv0->y - 5.0f * pv1->y + 4.0f * pv2->y - pv3->y) * s * s + (pv3->y -3.0f * pv2->y + 3.0f * pv1->y - pv0->y) * s * s * s);
1808     pout->z = 0.5f * (2.0f * pv1->z + (pv2->z - pv0->z) *s + (2.0f *pv0->z - 5.0f * pv1->z + 4.0f * pv2->z - pv3->z) * s * s + (pv3->z -3.0f * pv2->z + 3.0f * pv1->z - pv0->z) * s * s * s);
1809     return pout;
1810 }
1811
1812 D3DXVECTOR3* WINAPI D3DXVec3Hermite(D3DXVECTOR3 *pout, const D3DXVECTOR3 *pv1, const D3DXVECTOR3 *pt1, const D3DXVECTOR3 *pv2, const D3DXVECTOR3 *pt2, FLOAT s)
1813 {
1814     FLOAT h1, h2, h3, h4;
1815
1816     TRACE("pout %p, pv1 %p, pt1 %p, pv2 %p, pt2 %p, s %f\n", pout, pv1, pt1, pv2, pt2, s);
1817
1818     h1 = 2.0f * s * s * s - 3.0f * s * s + 1.0f;
1819     h2 = s * s * s - 2.0f * s * s + s;
1820     h3 = -2.0f * s * s * s + 3.0f * s * s;
1821     h4 = s * s * s - s * s;
1822
1823     pout->x = h1 * (pv1->x) + h2 * (pt1->x) + h3 * (pv2->x) + h4 * (pt2->x);
1824     pout->y = h1 * (pv1->y) + h2 * (pt1->y) + h3 * (pv2->y) + h4 * (pt2->y);
1825     pout->z = h1 * (pv1->z) + h2 * (pt1->z) + h3 * (pv2->z) + h4 * (pt2->z);
1826     return pout;
1827 }
1828
1829 D3DXVECTOR3* WINAPI D3DXVec3Normalize(D3DXVECTOR3 *pout, const D3DXVECTOR3 *pv)
1830 {
1831     FLOAT norm;
1832
1833     TRACE("pout %p, pv %p\n", pout, pv);
1834
1835     norm = D3DXVec3Length(pv);
1836     if ( !norm )
1837     {
1838         pout->x = 0.0f;
1839         pout->y = 0.0f;
1840         pout->z = 0.0f;
1841     }
1842     else
1843     {
1844         pout->x = pv->x / norm;
1845         pout->y = pv->y / norm;
1846         pout->z = pv->z / norm;
1847     }
1848
1849     return pout;
1850 }
1851
1852 D3DXVECTOR3* WINAPI D3DXVec3Project(D3DXVECTOR3 *pout, const D3DXVECTOR3 *pv, const D3DVIEWPORT9 *pviewport, const D3DXMATRIX *pprojection, const D3DXMATRIX *pview, const D3DXMATRIX *pworld)
1853 {
1854     D3DXMATRIX m;
1855
1856     TRACE("pout %p, pv %p, pviewport %p, pprojection %p, pview %p, pworld %p\n", pout, pv, pviewport, pprojection, pview, pworld);
1857
1858     D3DXMatrixIdentity(&m);
1859     if (pworld) D3DXMatrixMultiply(&m, &m, pworld);
1860     if (pview) D3DXMatrixMultiply(&m, &m, pview);
1861     if (pprojection) D3DXMatrixMultiply(&m, &m, pprojection);
1862
1863     D3DXVec3TransformCoord(pout, pv, &m);
1864
1865     if (pviewport)
1866     {
1867         pout->x = pviewport->X +  ( 1.0f + pout->x ) * pviewport->Width / 2.0f;
1868         pout->y = pviewport->Y +  ( 1.0f - pout->y ) * pviewport->Height / 2.0f;
1869         pout->z = pviewport->MinZ + pout->z * ( pviewport->MaxZ - pviewport->MinZ );
1870     }
1871     return pout;
1872 }
1873
1874 D3DXVECTOR3* WINAPI D3DXVec3ProjectArray(D3DXVECTOR3* out, UINT outstride, const D3DXVECTOR3* in, UINT instride, const D3DVIEWPORT9* viewport, const D3DXMATRIX* projection, const D3DXMATRIX* view, const D3DXMATRIX* world, UINT elements)
1875 {
1876     UINT i;
1877
1878     TRACE("out %p, outstride %u, in %p, instride %u, viewport %p, projection %p, view %p, world %p, elements %u\n",
1879         out, outstride, in, instride, viewport, projection, view, world, elements);
1880
1881     for (i = 0; i < elements; ++i) {
1882         D3DXVec3Project(
1883             (D3DXVECTOR3*)((char*)out + outstride * i),
1884             (const D3DXVECTOR3*)((const char*)in + instride * i),
1885             viewport, projection, view, world);
1886     }
1887     return out;
1888 }
1889
1890 D3DXVECTOR4* WINAPI D3DXVec3Transform(D3DXVECTOR4 *pout, const D3DXVECTOR3 *pv, const D3DXMATRIX *pm)
1891 {
1892     TRACE("pout %p, pv %p, pm %p\n", pout, pv, pm);
1893
1894     pout->x = pm->u.m[0][0] * pv->x + pm->u.m[1][0] * pv->y + pm->u.m[2][0] * pv->z + pm->u.m[3][0];
1895     pout->y = pm->u.m[0][1] * pv->x + pm->u.m[1][1] * pv->y + pm->u.m[2][1] * pv->z + pm->u.m[3][1];
1896     pout->z = pm->u.m[0][2] * pv->x + pm->u.m[1][2] * pv->y + pm->u.m[2][2] * pv->z + pm->u.m[3][2];
1897     pout->w = pm->u.m[0][3] * pv->x + pm->u.m[1][3] * pv->y + pm->u.m[2][3] * pv->z + pm->u.m[3][3];
1898     return pout;
1899 }
1900
1901 D3DXVECTOR4* WINAPI D3DXVec3TransformArray(D3DXVECTOR4* out, UINT outstride, const D3DXVECTOR3* in, UINT instride, const D3DXMATRIX* matrix, UINT elements)
1902 {
1903     UINT i;
1904
1905     TRACE("out %p, outstride %u, in %p, instride %u, matrix %p, elements %u\n", out, outstride, in, instride, matrix, elements);
1906
1907     for (i = 0; i < elements; ++i) {
1908         D3DXVec3Transform(
1909             (D3DXVECTOR4*)((char*)out + outstride * i),
1910             (const D3DXVECTOR3*)((const char*)in + instride * i),
1911             matrix);
1912     }
1913     return out;
1914 }
1915
1916 D3DXVECTOR3* WINAPI D3DXVec3TransformCoord(D3DXVECTOR3 *pout, const D3DXVECTOR3 *pv, const D3DXMATRIX *pm)
1917 {
1918     D3DXVECTOR3 out;
1919     FLOAT norm;
1920
1921     TRACE("pout %p, pv %p, pm %p\n", pout, pv, pm);
1922
1923     norm = pm->u.m[0][3] * pv->x + pm->u.m[1][3] * pv->y + pm->u.m[2][3] *pv->z + pm->u.m[3][3];
1924
1925     out.x = (pm->u.m[0][0] * pv->x + pm->u.m[1][0] * pv->y + pm->u.m[2][0] * pv->z + pm->u.m[3][0]) / norm;
1926     out.y = (pm->u.m[0][1] * pv->x + pm->u.m[1][1] * pv->y + pm->u.m[2][1] * pv->z + pm->u.m[3][1]) / norm;
1927     out.z = (pm->u.m[0][2] * pv->x + pm->u.m[1][2] * pv->y + pm->u.m[2][2] * pv->z + pm->u.m[3][2]) / norm;
1928
1929     *pout = out;
1930
1931     return pout;
1932 }
1933
1934 D3DXVECTOR3* WINAPI D3DXVec3TransformCoordArray(D3DXVECTOR3* out, UINT outstride, const D3DXVECTOR3* in, UINT instride, const D3DXMATRIX* matrix, UINT elements)
1935 {
1936     UINT i;
1937
1938     TRACE("out %p, outstride %u, in %p, instride %u, matrix %p, elements %u\n", out, outstride, in, instride, matrix, elements);
1939
1940     for (i = 0; i < elements; ++i) {
1941         D3DXVec3TransformCoord(
1942             (D3DXVECTOR3*)((char*)out + outstride * i),
1943             (const D3DXVECTOR3*)((const char*)in + instride * i),
1944             matrix);
1945     }
1946     return out;
1947 }
1948
1949 D3DXVECTOR3* WINAPI D3DXVec3TransformNormal(D3DXVECTOR3 *pout, const D3DXVECTOR3 *pv, const D3DXMATRIX *pm)
1950 {
1951     const D3DXVECTOR3 v = *pv;
1952
1953     TRACE("pout %p, pv %p, pm %p\n", pout, pv, pm);
1954
1955     pout->x = pm->u.m[0][0] * v.x + pm->u.m[1][0] * v.y + pm->u.m[2][0] * v.z;
1956     pout->y = pm->u.m[0][1] * v.x + pm->u.m[1][1] * v.y + pm->u.m[2][1] * v.z;
1957     pout->z = pm->u.m[0][2] * v.x + pm->u.m[1][2] * v.y + pm->u.m[2][2] * v.z;
1958     return pout;
1959
1960 }
1961
1962 D3DXVECTOR3* WINAPI D3DXVec3TransformNormalArray(D3DXVECTOR3* out, UINT outstride, const D3DXVECTOR3* in, UINT instride, const D3DXMATRIX* matrix, UINT elements)
1963 {
1964     UINT i;
1965
1966     TRACE("out %p, outstride %u, in %p, instride %u, matrix %p, elements %u\n", out, outstride, in, instride, matrix, elements);
1967
1968     for (i = 0; i < elements; ++i) {
1969         D3DXVec3TransformNormal(
1970             (D3DXVECTOR3*)((char*)out + outstride * i),
1971             (const D3DXVECTOR3*)((const char*)in + instride * i),
1972             matrix);
1973     }
1974     return out;
1975 }
1976
1977 D3DXVECTOR3* WINAPI D3DXVec3Unproject(D3DXVECTOR3 *pout, const D3DXVECTOR3 *pv, const D3DVIEWPORT9 *pviewport, const D3DXMATRIX *pprojection, const D3DXMATRIX *pview, const D3DXMATRIX *pworld)
1978 {
1979     D3DXMATRIX m;
1980
1981     TRACE("pout %p, pv %p, pviewport %p, pprojection %p, pview %p, pworlds %p\n", pout, pv, pviewport, pprojection, pview, pworld);
1982
1983     D3DXMatrixIdentity(&m);
1984     if (pworld) D3DXMatrixMultiply(&m, &m, pworld);
1985     if (pview) D3DXMatrixMultiply(&m, &m, pview);
1986     if (pprojection) D3DXMatrixMultiply(&m, &m, pprojection);
1987     D3DXMatrixInverse(&m, NULL, &m);
1988
1989     *pout = *pv;
1990     if (pviewport)
1991     {
1992         pout->x = 2.0f * ( pout->x - pviewport->X ) / pviewport->Width - 1.0f;
1993         pout->y = 1.0f - 2.0f * ( pout->y - pviewport->Y ) / pviewport->Height;
1994         pout->z = ( pout->z - pviewport->MinZ) / ( pviewport->MaxZ - pviewport->MinZ );
1995     }
1996     D3DXVec3TransformCoord(pout, pout, &m);
1997     return pout;
1998 }
1999
2000 D3DXVECTOR3* WINAPI D3DXVec3UnprojectArray(D3DXVECTOR3* out, UINT outstride, const D3DXVECTOR3* in, UINT instride, const D3DVIEWPORT9* viewport, const D3DXMATRIX* projection, const D3DXMATRIX* view, const D3DXMATRIX* world, UINT elements)
2001 {
2002     UINT i;
2003
2004     TRACE("out %p, outstride %u, in %p, instride %u, viewport %p, projection %p, view %p, world %p, elements %u\n",
2005         out, outstride, in, instride, viewport, projection, view, world, elements);
2006
2007     for (i = 0; i < elements; ++i) {
2008         D3DXVec3Unproject(
2009             (D3DXVECTOR3*)((char*)out + outstride * i),
2010             (const D3DXVECTOR3*)((const char*)in + instride * i),
2011             viewport, projection, view, world);
2012     }
2013     return out;
2014 }
2015
2016 /*_________________D3DXVec4_____________________*/
2017
2018 D3DXVECTOR4* WINAPI D3DXVec4BaryCentric(D3DXVECTOR4 *pout, const D3DXVECTOR4 *pv1, const D3DXVECTOR4 *pv2, const D3DXVECTOR4 *pv3, FLOAT f, FLOAT g)
2019 {
2020     TRACE("pout %p, pv1 %p, pv2 %p, pv3 %p, f %f, g %f\n", pout, pv1, pv2, pv3, f, g);
2021
2022     pout->x = (1.0f-f-g) * (pv1->x) + f * (pv2->x) + g * (pv3->x);
2023     pout->y = (1.0f-f-g) * (pv1->y) + f * (pv2->y) + g * (pv3->y);
2024     pout->z = (1.0f-f-g) * (pv1->z) + f * (pv2->z) + g * (pv3->z);
2025     pout->w = (1.0f-f-g) * (pv1->w) + f * (pv2->w) + g * (pv3->w);
2026     return pout;
2027 }
2028
2029 D3DXVECTOR4* WINAPI D3DXVec4CatmullRom(D3DXVECTOR4 *pout, const D3DXVECTOR4 *pv0, const D3DXVECTOR4 *pv1, const D3DXVECTOR4 *pv2, const D3DXVECTOR4 *pv3, FLOAT s)
2030 {
2031     TRACE("pout %p, pv0 %p, pv1 %p, pv2 %p, pv3 %p, s %f\n", pout, pv0, pv1, pv2, pv3, s);
2032
2033     pout->x = 0.5f * (2.0f * pv1->x + (pv2->x - pv0->x) *s + (2.0f *pv0->x - 5.0f * pv1->x + 4.0f * pv2->x - pv3->x) * s * s + (pv3->x -3.0f * pv2->x + 3.0f * pv1->x - pv0->x) * s * s * s);
2034     pout->y = 0.5f * (2.0f * pv1->y + (pv2->y - pv0->y) *s + (2.0f *pv0->y - 5.0f * pv1->y + 4.0f * pv2->y - pv3->y) * s * s + (pv3->y -3.0f * pv2->y + 3.0f * pv1->y - pv0->y) * s * s * s);
2035     pout->z = 0.5f * (2.0f * pv1->z + (pv2->z - pv0->z) *s + (2.0f *pv0->z - 5.0f * pv1->z + 4.0f * pv2->z - pv3->z) * s * s + (pv3->z -3.0f * pv2->z + 3.0f * pv1->z - pv0->z) * s * s * s);
2036     pout->w = 0.5f * (2.0f * pv1->w + (pv2->w - pv0->w) *s + (2.0f *pv0->w - 5.0f * pv1->w + 4.0f * pv2->w - pv3->w) * s * s + (pv3->w -3.0f * pv2->w + 3.0f * pv1->w - pv0->w) * s * s * s);
2037     return pout;
2038 }
2039
2040 D3DXVECTOR4* WINAPI D3DXVec4Cross(D3DXVECTOR4 *pout, const D3DXVECTOR4 *pv1, const D3DXVECTOR4 *pv2, const D3DXVECTOR4 *pv3)
2041 {
2042     D3DXVECTOR4 out;
2043
2044     TRACE("pout %p, pv1 %p, pv2 %p, pv3 %p\n", pout, pv1, pv2, pv3);
2045
2046     out.x = pv1->y * (pv2->z * pv3->w - pv3->z * pv2->w) - pv1->z * (pv2->y * pv3->w - pv3->y * pv2->w) + pv1->w * (pv2->y * pv3->z - pv2->z *pv3->y);
2047     out.y = -(pv1->x * (pv2->z * pv3->w - pv3->z * pv2->w) - pv1->z * (pv2->x * pv3->w - pv3->x * pv2->w) + pv1->w * (pv2->x * pv3->z - pv3->x * pv2->z));
2048     out.z = pv1->x * (pv2->y * pv3->w - pv3->y * pv2->w) - pv1->y * (pv2->x *pv3->w - pv3->x * pv2->w) + pv1->w * (pv2->x * pv3->y - pv3->x * pv2->y);
2049     out.w = -(pv1->x * (pv2->y * pv3->z - pv3->y * pv2->z) - pv1->y * (pv2->x * pv3->z - pv3->x *pv2->z) + pv1->z * (pv2->x * pv3->y - pv3->x * pv2->y));
2050     *pout = out;
2051     return pout;
2052 }
2053
2054 D3DXVECTOR4* WINAPI D3DXVec4Hermite(D3DXVECTOR4 *pout, const D3DXVECTOR4 *pv1, const D3DXVECTOR4 *pt1, const D3DXVECTOR4 *pv2, const D3DXVECTOR4 *pt2, FLOAT s)
2055 {
2056     FLOAT h1, h2, h3, h4;
2057
2058     TRACE("pout %p, pv1 %p, pt1 %p, pv2 %p, pt2 %p, s %f\n", pout, pv1, pt1, pv2, pt2, s);
2059
2060     h1 = 2.0f * s * s * s - 3.0f * s * s + 1.0f;
2061     h2 = s * s * s - 2.0f * s * s + s;
2062     h3 = -2.0f * s * s * s + 3.0f * s * s;
2063     h4 = s * s * s - s * s;
2064
2065     pout->x = h1 * (pv1->x) + h2 * (pt1->x) + h3 * (pv2->x) + h4 * (pt2->x);
2066     pout->y = h1 * (pv1->y) + h2 * (pt1->y) + h3 * (pv2->y) + h4 * (pt2->y);
2067     pout->z = h1 * (pv1->z) + h2 * (pt1->z) + h3 * (pv2->z) + h4 * (pt2->z);
2068     pout->w = h1 * (pv1->w) + h2 * (pt1->w) + h3 * (pv2->w) + h4 * (pt2->w);
2069     return pout;
2070 }
2071
2072 D3DXVECTOR4* WINAPI D3DXVec4Normalize(D3DXVECTOR4 *pout, const D3DXVECTOR4 *pv)
2073 {
2074     FLOAT norm;
2075
2076     TRACE("pout %p, pv %p\n", pout, pv);
2077
2078     norm = D3DXVec4Length(pv);
2079
2080     pout->x = pv->x / norm;
2081     pout->y = pv->y / norm;
2082     pout->z = pv->z / norm;
2083     pout->w = pv->w / norm;
2084
2085     return pout;
2086 }
2087
2088 D3DXVECTOR4* WINAPI D3DXVec4Transform(D3DXVECTOR4 *pout, const D3DXVECTOR4 *pv, const D3DXMATRIX *pm)
2089 {
2090     D3DXVECTOR4 out;
2091
2092     TRACE("pout %p, pv %p, pm %p\n", pout, pv, pm);
2093
2094     out.x = pm->u.m[0][0] * pv->x + pm->u.m[1][0] * pv->y + pm->u.m[2][0] * pv->z + pm->u.m[3][0] * pv->w;
2095     out.y = pm->u.m[0][1] * pv->x + pm->u.m[1][1] * pv->y + pm->u.m[2][1] * pv->z + pm->u.m[3][1] * pv->w;
2096     out.z = pm->u.m[0][2] * pv->x + pm->u.m[1][2] * pv->y + pm->u.m[2][2] * pv->z + pm->u.m[3][2] * pv->w;
2097     out.w = pm->u.m[0][3] * pv->x + pm->u.m[1][3] * pv->y + pm->u.m[2][3] * pv->z + pm->u.m[3][3] * pv->w;
2098     *pout = out;
2099     return pout;
2100 }
2101
2102 D3DXVECTOR4* WINAPI D3DXVec4TransformArray(D3DXVECTOR4* out, UINT outstride, const D3DXVECTOR4* in, UINT instride, const D3DXMATRIX* matrix, UINT elements)
2103 {
2104     UINT i;
2105
2106     TRACE("out %p, outstride %u, in %p, instride %u, matrix %p, elements %u\n", out, outstride, in, instride, matrix, elements);
2107
2108     for (i = 0; i < elements; ++i) {
2109         D3DXVec4Transform(
2110             (D3DXVECTOR4*)((char*)out + outstride * i),
2111             (const D3DXVECTOR4*)((const char*)in + instride * i),
2112             matrix);
2113     }
2114     return out;
2115 }
2116
2117 unsigned short float_32_to_16(const float in)
2118 {
2119     int exp = 0, origexp;
2120     float tmp = fabsf(in);
2121     int sign = (copysignf(1, in) < 0);
2122     unsigned int mantissa;
2123     unsigned short ret;
2124
2125     /* Deal with special numbers */
2126     if (isinf(in)) return (sign ? 0xffff : 0x7fff);
2127     if (isnan(in)) return (sign ? 0xffff : 0x7fff);
2128     if (in == 0.0f) return (sign ? 0x8000 : 0x0000);
2129
2130     if (tmp < (float)(1u << 10))
2131     {
2132         do
2133         {
2134             tmp *= 2.0f;
2135             exp--;
2136         } while (tmp < (float)(1u << 10));
2137     }
2138     else if (tmp >= (float)(1u << 11))
2139     {
2140         do
2141         {
2142             tmp /= 2.0f;
2143             exp++;
2144         } while (tmp >= (float)(1u << 11));
2145     }
2146
2147     exp += 10;  /* Normalize the mantissa */
2148     exp += 15;  /* Exponent is encoded with excess 15 */
2149
2150     origexp = exp;
2151
2152     mantissa = (unsigned int) tmp;
2153     if ((tmp - mantissa == 0.5f && mantissa % 2 == 1) || /* round half to even */
2154         (tmp - mantissa > 0.5f))
2155     {
2156         mantissa++; /* round to nearest, away from zero */
2157     }
2158     if (mantissa == 2048)
2159     {
2160         mantissa = 1024;
2161         exp++;
2162     }
2163
2164     if (exp > 31)
2165     {
2166         /* too big */
2167         ret = 0x7fff; /* INF */
2168     }
2169     else if (exp <= 0)
2170     {
2171         unsigned int rounding = 0;
2172
2173         /* Denormalized half float */
2174
2175         /* return 0x0000 (=0.0) for numbers too small to represent in half floats */
2176         if (exp < -11)
2177             return (sign ? 0x8000 : 0x0000);
2178
2179         exp = origexp;
2180
2181         /* the 13 extra bits from single precision are used for rounding */
2182         mantissa = (unsigned int)(tmp * (1u << 13));
2183         mantissa >>= 1 - exp; /* denormalize */
2184
2185         mantissa -= ~(mantissa >> 13) & 1; /* round half to even */
2186         /* remove 13 least significant bits to get half float precision */
2187         mantissa >>= 12;
2188         rounding = mantissa & 1;
2189         mantissa >>= 1;
2190
2191         ret = mantissa + rounding;
2192     }
2193     else
2194     {
2195         ret = (exp << 10) | (mantissa & 0x3ff);
2196     }
2197
2198     ret |= ((sign ? 1 : 0) << 15); /* Add the sign */
2199     return ret;
2200 }
2201
2202 D3DXFLOAT16 *WINAPI D3DXFloat32To16Array(D3DXFLOAT16 *pout, const FLOAT *pin, UINT n)
2203 {
2204     unsigned int i;
2205
2206     TRACE("pout %p, pin %p, n %u\n", pout, pin, n);
2207
2208     for (i = 0; i < n; ++i)
2209     {
2210         pout[i].value = float_32_to_16(pin[i]);
2211     }
2212
2213     return pout;
2214 }
2215
2216 /* Native d3dx9's D3DXFloat16to32Array lacks support for NaN and Inf. Specifically, e = 16 is treated as a
2217  * regular number - e.g., 0x7fff is converted to 131008.0 and 0xffff to -131008.0. */
2218 float float_16_to_32(const unsigned short in)
2219 {
2220     const unsigned short s = (in & 0x8000);
2221     const unsigned short e = (in & 0x7C00) >> 10;
2222     const unsigned short m = in & 0x3FF;
2223     const float sgn = (s ? -1.0f : 1.0f);
2224
2225     if (e == 0)
2226     {
2227         if (m == 0) return sgn * 0.0f; /* +0.0 or -0.0 */
2228         else return sgn * powf(2, -14.0f) * (m / 1024.0f);
2229     }
2230     else
2231     {
2232         return sgn * powf(2, e - 15.0f) * (1.0f + (m / 1024.0f));
2233     }
2234 }
2235
2236 FLOAT *WINAPI D3DXFloat16To32Array(FLOAT *pout, const D3DXFLOAT16 *pin, UINT n)
2237 {
2238     unsigned int i;
2239
2240     TRACE("pout %p, pin %p, n %u\n", pout, pin, n);
2241
2242     for (i = 0; i < n; ++i)
2243     {
2244         pout[i] = float_16_to_32(pin[i].value);
2245     }
2246
2247     return pout;
2248 }
2249
2250 /*_________________D3DXSH________________*/
2251
2252 FLOAT* WINAPI D3DXSHAdd(FLOAT *out, UINT order, const FLOAT *a, const FLOAT *b)
2253 {
2254     UINT i;
2255
2256     TRACE("out %p, order %u, a %p, b %p\n", out, order, a, b);
2257
2258     for (i = 0; i < order * order; i++)
2259         out[i] = a[i] + b[i];
2260
2261     return out;
2262 }
2263
2264 FLOAT WINAPI D3DXSHDot(UINT order, const FLOAT *a, const FLOAT *b)
2265 {
2266     FLOAT s;
2267     UINT i;
2268
2269     TRACE("order %u, a %p, b %p\n", order, a, b);
2270
2271     s = a[0] * b[0];
2272     for (i = 1; i < order * order; i++)
2273         s += a[i] * b[i];
2274
2275     return s;
2276 }
2277
2278 static void weightedcapintegrale(FLOAT *out, UINT order, FLOAT angle)
2279 {
2280     FLOAT coeff[3];
2281
2282     coeff[0] = cosf(angle);
2283
2284     out[0] = 2.0f * D3DX_PI * (1.0f - coeff[0]);
2285     out[1] = D3DX_PI * sinf(angle) * sinf(angle);
2286     if (order <= 2)
2287         return;
2288
2289     out[2] = coeff[0] * out[1];
2290     if (order == 3)
2291         return;
2292
2293     coeff[1] = coeff[0] * coeff[0];
2294     coeff[2] = coeff[1] * coeff[1];
2295
2296     out[3] = D3DX_PI * (-1.25f * coeff[2] + 1.5f * coeff[1] - 0.25f);
2297     if (order == 4)
2298         return;
2299
2300     out[4] = -0.25f * D3DX_PI * coeff[0] * (7.0f * coeff[2] - 10.0f * coeff[1] + 3.0f);
2301     if (order == 5)
2302         return;
2303
2304     out[5] = D3DX_PI * (-2.625f * coeff[2] * coeff[1] + 4.375f * coeff[2] - 1.875f * coeff[1] + 0.125f);
2305 }
2306
2307 HRESULT WINAPI D3DXSHEvalConeLight(UINT order, const D3DXVECTOR3 *dir, FLOAT radius,
2308     FLOAT Rintensity, FLOAT Gintensity, FLOAT Bintensity, FLOAT *rout, FLOAT *gout, FLOAT *bout)
2309 {
2310     FLOAT cap[6], clamped_angle, norm, scale, temp;
2311     UINT i, index, j;
2312
2313     TRACE("order %u, dir %p, radius %f, red %f, green %f, blue %f, rout %p, gout %p, bout %p\n",
2314         order, dir, radius, Rintensity, Gintensity, Bintensity, rout, gout, bout);
2315
2316     if (radius <= 0.0f)
2317         return D3DXSHEvalDirectionalLight(order, dir, Rintensity, Gintensity, Bintensity, rout, gout, bout);
2318
2319     clamped_angle = (radius > D3DX_PI / 2.0f) ? (D3DX_PI / 2.0f) : radius;
2320     norm = sinf(clamped_angle) * sinf(clamped_angle);
2321
2322     if (order > D3DXSH_MAXORDER)
2323     {
2324         WARN("Order clamped at D3DXSH_MAXORDER\n");
2325         order = D3DXSH_MAXORDER;
2326     }
2327
2328     weightedcapintegrale(cap, order, radius);
2329     D3DXSHEvalDirection(rout, order, dir);
2330
2331     for (i = 0; i < order; i++)
2332     {
2333         scale = cap[i] / norm;
2334
2335         for (j = 0; j < 2 * i + 1; j++)
2336         {
2337             index = i * i + j;
2338             temp = rout[index] * scale;
2339
2340             rout[index] = temp * Rintensity;
2341             if (gout)
2342                 gout[index] = temp * Gintensity;
2343             if (bout)
2344                 bout[index] = temp * Bintensity;
2345         }
2346     }
2347
2348     return D3D_OK;
2349 }
2350
2351 FLOAT* WINAPI D3DXSHEvalDirection(FLOAT *out, UINT order, const D3DXVECTOR3 *dir)
2352 {
2353     const FLOAT dirxx = dir->x * dir->x;
2354     const FLOAT dirxy = dir->x * dir->y;
2355     const FLOAT dirxz = dir->x * dir->z;
2356     const FLOAT diryy = dir->y * dir->y;
2357     const FLOAT diryz = dir->y * dir->z;
2358     const FLOAT dirzz = dir->z * dir->z;
2359     const FLOAT dirxxxx = dirxx * dirxx;
2360     const FLOAT diryyyy = diryy * diryy;
2361     const FLOAT dirzzzz = dirzz * dirzz;
2362     const FLOAT dirxyxy = dirxy * dirxy;
2363
2364     TRACE("out %p, order %u, dir %p\n", out, order, dir);
2365
2366     if ((order < D3DXSH_MINORDER) || (order > D3DXSH_MAXORDER))
2367         return out;
2368
2369     out[0] = 0.5f / sqrtf(D3DX_PI);
2370     out[1] = -0.5f / sqrtf(D3DX_PI / 3.0f) * dir->y;
2371     out[2] = 0.5f / sqrtf(D3DX_PI / 3.0f) * dir->z;
2372     out[3] = -0.5f / sqrtf(D3DX_PI / 3.0f) * dir->x;
2373     if (order == 2)
2374         return out;
2375
2376     out[4] = 0.5f / sqrtf(D3DX_PI / 15.0f) * dirxy;
2377     out[5] = -0.5f / sqrtf(D3DX_PI / 15.0f) * diryz;
2378     out[6] = 0.25f / sqrtf(D3DX_PI / 5.0f) * (3.0f * dirzz - 1.0f);
2379     out[7] = -0.5f / sqrtf(D3DX_PI / 15.0f) * dirxz;
2380     out[8] = 0.25f / sqrtf(D3DX_PI / 15.0f) * (dirxx - diryy);
2381     if (order == 3)
2382         return out;
2383
2384     out[9] = -sqrtf(70.0f / D3DX_PI) / 8.0f * dir->y * (3.0f * dirxx - diryy);
2385     out[10] = sqrtf(105.0f / D3DX_PI) / 2.0f * dirxy * dir->z;
2386     out[11] = -sqrtf(42.0f / D3DX_PI) / 8.0f * dir->y * (-1.0f + 5.0f * dirzz);
2387     out[12] = sqrtf(7.0f / D3DX_PI) / 4.0f * dir->z * (5.0f * dirzz - 3.0f);
2388     out[13] = sqrtf(42.0f / D3DX_PI) / 8.0f * dir->x * (1.0f - 5.0f * dirzz);
2389     out[14] = sqrtf(105.0f / D3DX_PI) / 4.0f * dir->z * (dirxx - diryy);
2390     out[15] = -sqrtf(70.0f / D3DX_PI) / 8.0f * dir->x * (dirxx - 3.0f * diryy);
2391     if (order == 4)
2392         return out;
2393
2394     out[16] = 0.75f * sqrtf(35.0f / D3DX_PI) * dirxy * (dirxx - diryy);
2395     out[17] = 3.0f * dir->z * out[9];
2396     out[18] = 0.75f * sqrtf(5.0f / D3DX_PI) * dirxy * (7.0f * dirzz - 1.0f);
2397     out[19] = 0.375f * sqrtf(10.0f / D3DX_PI) * diryz * (3.0f - 7.0f * dirzz);
2398     out[20] = 3.0f / (16.0f * sqrtf(D3DX_PI)) * (35.0f * dirzzzz - 30.f * dirzz + 3.0f);
2399     out[21] = 0.375f * sqrtf(10.0f / D3DX_PI) * dirxz * (3.0f - 7.0f * dirzz);
2400     out[22] = 0.375f * sqrtf(5.0f / D3DX_PI) * (dirxx - diryy) * (7.0f * dirzz - 1.0f);
2401     out[23] = 3.0f * dir->z * out[15];
2402     out[24] = 3.0f / 16.0f * sqrtf(35.0f / D3DX_PI) * (dirxxxx - 6.0f * dirxyxy + diryyyy);
2403     if (order == 5)
2404         return out;
2405
2406     out[25] = -3.0f/ 32.0f * sqrtf(154.0f / D3DX_PI) * dir->y * (5.0f * dirxxxx - 10.0f * dirxyxy + diryyyy);
2407     out[26] = 0.75f * sqrtf(385.0f / D3DX_PI) * dirxy * dir->z * (dirxx - diryy);
2408     out[27] = sqrtf(770.0f / D3DX_PI) / 32.0f * dir->y * (3.0f * dirxx - diryy) * (1.0f - 9.0f * dirzz);
2409     out[28] = sqrtf(1155.0f / D3DX_PI) / 4.0f * dirxy * dir->z * (3.0f * dirzz - 1.0f);
2410     out[29] = sqrtf(165.0f / D3DX_PI) / 16.0f * dir->y * (14.0f * dirzz - 21.0f * dirzzzz - 1.0f);
2411     out[30] = sqrtf(11.0f / D3DX_PI) / 16.0f * dir->z * (63.0f * dirzzzz - 70.0f * dirzz + 15.0f);
2412     out[31] = sqrtf(165.0f / D3DX_PI) / 16.0f * dir->x * (14.0f * dirzz - 21.0f * dirzzzz - 1.0f);
2413     out[32] = sqrtf(1155.0f / D3DX_PI) / 8.0f * dir->z * (dirxx - diryy) * (3.0f * dirzz - 1.0f);
2414     out[33] = sqrtf(770.0f / D3DX_PI) / 32.0f * dir->x * (dirxx - 3.0f * diryy) * (1.0f - 9.0f * dirzz);
2415     out[34] = 3.0f / 16.0f * sqrtf(385.0f / D3DX_PI) * dir->z * (dirxxxx - 6.0f * dirxyxy + diryyyy);
2416     out[35] = -3.0f/ 32.0f * sqrtf(154.0f / D3DX_PI) * dir->x * (dirxxxx - 10.0f * dirxyxy + 5.0f * diryyyy);
2417
2418     return out;
2419 }
2420
2421 HRESULT WINAPI D3DXSHEvalDirectionalLight(UINT order, const D3DXVECTOR3 *dir, FLOAT Rintensity, FLOAT Gintensity, FLOAT Bintensity, FLOAT *Rout, FLOAT *Gout, FLOAT *Bout)
2422 {
2423     FLOAT s, temp;
2424     UINT j;
2425
2426     TRACE("Order %u, Vector %p, Red %f, Green %f, Blue %f, Rout %p, Gout %p, Bout %p\n", order, dir, Rintensity, Gintensity, Bintensity, Rout, Gout, Bout);
2427
2428     s = 0.75f;
2429     if ( order > 2 )
2430         s += 5.0f / 16.0f;
2431     if ( order > 4 )
2432         s -= 3.0f / 32.0f;
2433     s /= D3DX_PI;
2434
2435     D3DXSHEvalDirection(Rout, order, dir);
2436     for (j = 0; j < order * order; j++)
2437     {
2438         temp = Rout[j] / s;
2439
2440         Rout[j] = Rintensity * temp;
2441         if ( Gout )
2442             Gout[j] = Gintensity * temp;
2443         if ( Bout )
2444             Bout[j] = Bintensity * temp;
2445     }
2446
2447     return D3D_OK;
2448 }
2449
2450 HRESULT WINAPI D3DXSHEvalHemisphereLight(UINT order, const D3DXVECTOR3 *dir, D3DXCOLOR top, D3DXCOLOR bottom,
2451     FLOAT *rout, FLOAT *gout, FLOAT *bout)
2452 {
2453     FLOAT a[2], temp[4];
2454     UINT i, j;
2455
2456     TRACE("order %u, dir %p, rout %p, gout %p, bout %p\n", order, dir, rout, gout, bout);
2457
2458     D3DXSHEvalDirection(temp, 2, dir);
2459
2460     a[0] = (top.r + bottom.r) * 3.0f * D3DX_PI;
2461     a[1] = (top.r - bottom.r) * D3DX_PI;
2462     for (i = 0; i < order; i++)
2463         for (j = 0; j < 2 * i + 1; j++)
2464             if (i < 2)
2465                 rout[i * i + j] = temp[i * i + j] * a[i];
2466             else
2467                 rout[i * i + j] = 0.0f;
2468
2469     if (gout)
2470     {
2471         a[0] = (top.g + bottom.g) * 3.0f * D3DX_PI;
2472         a[1] = (top.g - bottom.g) * D3DX_PI;
2473         for (i = 0; i < order; i++)
2474             for (j = 0; j < 2 * i + 1; j++)
2475                 if (i < 2)
2476                     gout[i * i + j] = temp[i * i + j] * a[i];
2477                 else
2478                     gout[i * i + j] = 0.0f;
2479     }
2480
2481     if (bout)
2482     {
2483         a[0] = (top.b + bottom.b) * 3.0f * D3DX_PI;
2484         a[1] = (top.b - bottom.b) * D3DX_PI;
2485         for (i = 0; i < order; i++)
2486             for (j = 0; j < 2 * i + 1; j++)
2487                 if (i < 2)
2488                     bout[i * i + j] = temp[i * i + j] * a[i];
2489                 else
2490                     bout[i * i + j] = 0.0f;
2491     }
2492
2493     return D3D_OK;
2494 }
2495
2496 HRESULT WINAPI D3DXSHEvalSphericalLight(UINT order, const D3DXVECTOR3 *dir, FLOAT radius,
2497     FLOAT Rintensity, FLOAT Gintensity, FLOAT Bintensity, FLOAT *rout, FLOAT *gout, FLOAT *bout)
2498 {
2499     D3DXVECTOR3 normal;
2500     FLOAT cap[6], clamped_angle, dist, temp;
2501     UINT i, index, j;
2502
2503     TRACE("order %u, dir %p, radius %f, red %f, green %f, blue %f, rout %p, gout %p, bout %p\n",
2504         order, dir, radius, Rintensity, Gintensity, Bintensity, rout, gout, bout);
2505
2506     if (order > D3DXSH_MAXORDER)
2507     {
2508         WARN("Order clamped at D3DXSH_MAXORDER\n");
2509         order = D3DXSH_MAXORDER;
2510     }
2511
2512     if (radius < 0.0f)
2513         radius = -radius;
2514
2515     dist = D3DXVec3Length(dir);
2516     clamped_angle = (dist <= radius) ? D3DX_PI / 2.0f : asinf(radius / dist);
2517
2518     weightedcapintegrale(cap, order, clamped_angle);
2519     D3DXVec3Normalize(&normal, dir);
2520     D3DXSHEvalDirection(rout, order, &normal);
2521
2522     for (i = 0; i < order; i++)
2523         for (j = 0; j < 2 * i + 1; j++)
2524         {
2525             index = i * i + j;
2526             temp = rout[index] * cap[i];
2527
2528             rout[index] = temp * Rintensity;
2529             if (gout)
2530                 gout[index] = temp * Gintensity;
2531             if (bout)
2532                 bout[index] = temp * Bintensity;
2533         }
2534
2535     return D3D_OK;
2536 }
2537
2538 FLOAT * WINAPI D3DXSHMultiply2(FLOAT *out, const FLOAT *a, const FLOAT *b)
2539 {
2540     FLOAT ta, tb;
2541
2542     TRACE("out %p, a %p, b %p\n", out, a, b);
2543
2544     ta = 0.28209479f * a[0];
2545     tb = 0.28209479f * b[0];
2546
2547     out[0] = 0.28209479f * D3DXSHDot(2, a, b);
2548     out[1] = ta * b[1] + tb * a[1];
2549     out[2] = ta * b[2] + tb * a[2];
2550     out[3] = ta * b[3] + tb * a[3];
2551
2552     return out;
2553 }
2554
2555 FLOAT * WINAPI D3DXSHMultiply3(FLOAT *out, const FLOAT *a, const FLOAT *b)
2556 {
2557     FLOAT t, ta, tb;
2558
2559     TRACE("out %p, a %p, b %p\n", out, a, b);
2560
2561     out[0] = 0.28209479f * a[0] * b[0];
2562
2563     ta = 0.28209479f * a[0] - 0.12615662f * a[6] - 0.21850968f * a[8];
2564     tb = 0.28209479f * b[0] - 0.12615662f * b[6] - 0.21850968f * b[8];
2565     out[1] = ta * b[1] + tb * a[1];
2566     t = a[1] * b[1];
2567     out[0] += 0.28209479f * t;
2568     out[6] = -0.12615662f * t;
2569     out[8] = -0.21850968f * t;
2570
2571     ta = 0.21850968f * a[5];
2572     tb = 0.21850968f * b[5];
2573     out[1] += ta * b[2] + tb * a[2];
2574     out[2] = ta * b[1] + tb * a[1];
2575     t = a[1] * b[2] +a[2] * b[1];
2576     out[5] = 0.21850968f * t;
2577
2578     ta = 0.21850968f * a[4];
2579     tb = 0.21850968f * b[4];
2580     out[1] += ta * b[3] + tb * a[3];
2581     out[3]  = ta * b[1] + tb * a[1];
2582     t = a[1] * b[3] + a[3] * b[1];
2583     out[4] = 0.21850968f * t;
2584
2585     ta = 0.28209480f * a[0] + 0.25231326f * a[6];
2586     tb = 0.28209480f * b[0] + 0.25231326f * b[6];
2587     out[2] += ta * b[2] + tb * a[2];
2588     t = a[2] * b[2];
2589     out[0] += 0.28209480f * t;
2590     out[6] += 0.25231326f * t;
2591
2592     ta = 0.21850969f * a[7];
2593     tb = 0.21850969f * b[7];
2594     out[2] += ta * b[3] + tb * a[3];
2595     out[3] += ta * b[2] + tb * a[2];
2596     t = a[2] * b[3] + a[3] * b[2];
2597     out[7] = 0.21850969f * t;
2598
2599     ta = 0.28209479f * a[0] - 0.12615663f * a[6] + 0.21850969f * a[8];
2600     tb = 0.28209479f * b[0] - 0.12615663f * b[6] + 0.21850969f * b[8];
2601     out[3] += ta * b[3] + tb * a[3];
2602     t = a[3] * b[3];
2603     out[0] += 0.28209479f * t;
2604     out[6] -= 0.12615663f * t;
2605     out[8] += 0.21850969f * t;
2606
2607     ta = 0.28209479f * a[0] - 0.18022375f * a[6];
2608     tb = 0.28209479f * b[0] - 0.18022375f * b[6];
2609     out[4] += ta * b[4] + tb * a[4];
2610     t = a[4] * b[4];
2611     out[0] += 0.28209479f * t;
2612     out[6] -= 0.18022375f * t;
2613
2614     ta = 0.15607835f * a[7];
2615     tb = 0.15607835f * b[7];
2616     out[4] += ta * b[5] + tb * a[5];
2617     out[5] += ta * b[4] + tb * a[4];
2618     t = a[4] * b[5] + a[5] * b[4];
2619     out[7] += 0.15607834f * t;
2620
2621     ta = 0.28209479f * a[0] + 0.09011186f * a[6] - 0.15607835f * a[8];
2622     tb = 0.28209479f * b[0] + 0.09011186f * b[6] - 0.15607835f * b[8];
2623     out[5] += ta * b[5] + tb * a[5];
2624     t = a[5] * b[5];
2625     out[0] += 0.28209479f * t;
2626     out[6] += 0.09011186f * t;
2627     out[8] -= 0.15607835f * t;
2628
2629     ta = 0.28209480f * a[0];
2630     tb = 0.28209480f * b[0];
2631     out[6] += ta * b[6] + tb * a[6];
2632     t = a[6] * b[6];
2633     out[0] += 0.28209480f * t;
2634     out[6] += 0.18022376f * t;
2635
2636     ta = 0.28209479f * a[0] + 0.09011186f * a[6] + 0.15607835f * a[8];
2637     tb = 0.28209479f * b[0] + 0.09011186f * b[6] + 0.15607835f * b[8];
2638     out[7] += ta * b[7] + tb * a[7];
2639     t = a[7] * b[7];
2640     out[0] += 0.28209479f * t;
2641     out[6] += 0.09011186f * t;
2642     out[8] += 0.15607835f * t;
2643
2644     ta = 0.28209479f * a[0] - 0.18022375f * a[6];
2645     tb = 0.28209479f * b[0] - 0.18022375f * b[6];
2646     out[8] += ta * b[8] + tb * a[8];
2647     t = a[8] * b[8];
2648     out[0] += 0.28209479f * t;
2649     out[6] -= 0.18022375f * t;
2650
2651     return out;
2652 }
2653
2654 FLOAT * WINAPI D3DXSHMultiply4(FLOAT *out, const FLOAT *a, const FLOAT *b)
2655 {
2656     FLOAT ta, tb, t;
2657
2658     TRACE("out %p, a %p, b %p\n", out, a, b);
2659
2660     out[0] = 0.28209479f * a[0] * b[0];
2661
2662     ta = 0.28209479f * a[0] - 0.12615663f * a[6] - 0.21850969f * a[8];
2663     tb = 0.28209479f * b[0] - 0.12615663f * b[6] - 0.21850969f * b[8];
2664     out[1] = ta * b[1] + tb * a[1];
2665     t = a[1] * b[1];
2666     out[0] += 0.28209479f * t;
2667     out[6] = -0.12615663f * t;
2668     out[8] = -0.21850969f * t;
2669
2670     ta = 0.21850969f * a[3] - 0.05839917f * a[13] - 0.22617901f * a[15];
2671     tb = 0.21850969f * b[3] - 0.05839917f * b[13] - 0.22617901f * b[15];
2672     out[1] += ta * b[4] + tb * a[4];
2673     out[4] = ta * b[1] + tb * a[1];
2674     t = a[1] * b[4] + a[4] * b[1];
2675     out[3] = 0.21850969f * t;
2676     out[13] = -0.05839917f * t;
2677     out[15] = -0.22617901f * t;
2678
2679     ta = 0.21850969f * a[2] - 0.14304817f * a[12] - 0.18467439f * a[14];
2680     tb = 0.21850969f * b[2] - 0.14304817f * b[12] - 0.18467439f * b[14];
2681     out[1] += ta * b[5] + tb * a[5];
2682     out[5] = ta * b[1] + tb * a[1];
2683     t = a[1] * b[5] + a[5] * b[1];
2684     out[2] = 0.21850969f * t;
2685     out[12] = -0.14304817f * t;
2686     out[14] = -0.18467439f * t;
2687
2688     ta = 0.20230066f * a[11];
2689     tb = 0.20230066f * b[11];
2690     out[1] += ta * b[6] + tb * a[6];
2691     out[6] += ta * b[1] + tb * a[1];
2692     t = a[1] * b[6] + a[6] * b[1];
2693     out[11] = 0.20230066f * t;
2694
2695     ta = 0.22617901f * a[9] + 0.05839917f * a[11];
2696     tb = 0.22617901f * b[9] + 0.05839917f * b[11];
2697     out[1] += ta * b[8] + tb * a[8];
2698     out[8] += ta * b[1] + tb * a[1];
2699     t = a[1] * b[8] + a[8] * b[1];
2700     out[9] = 0.22617901f * t;
2701     out[11] += 0.05839917f * t;
2702
2703     ta = 0.28209480f * a[0] + 0.25231326f * a[6];
2704     tb = 0.28209480f * b[0] + 0.25231326f * b[6];
2705     out[2] += ta * b[2] + tb * a[2];
2706     t = a[2] * b[2];
2707     out[0] += 0.28209480f * t;
2708     out[6] += 0.25231326f * t;
2709
2710     ta = 0.24776671f * a[12];
2711     tb = 0.24776671f * b[12];
2712     out[2] += ta * b[6] + tb * a[6];
2713     out[6] += ta * b[2] + tb * a[2];
2714     t = a[2] * b[6] + a[6] * b[2];
2715     out[12] += 0.24776671f * t;
2716
2717     ta = 0.28209480f * a[0] - 0.12615663f * a[6] + 0.21850969f * a[8];
2718     tb = 0.28209480f * b[0] - 0.12615663f * b[6] + 0.21850969f * b[8];
2719     out[3] += ta * b[3] + tb * a[3];
2720     t = a[3] * b[3];
2721     out[0] += 0.28209480f * t;
2722     out[6] -= 0.12615663f * t;
2723     out[8] += 0.21850969f * t;
2724
2725     ta = 0.20230066f * a[13];
2726     tb = 0.20230066f * b[13];
2727     out[3] += ta * b[6] + tb * a[6];
2728     out[6] += ta * b[3] + tb * a[3];
2729     t = a[3] * b[6] + a[6] * b[3];
2730     out[13] += 0.20230066f * t;
2731
2732     ta = 0.21850969f * a[2] - 0.14304817f * a[12] + 0.18467439f * a[14];
2733     tb = 0.21850969f * b[2] - 0.14304817f * b[12] + 0.18467439f * b[14];
2734     out[3] += ta * b[7] + tb * a[7];
2735     out[7] = ta * b[3] + tb * a[3];
2736     t = a[3] * b[7] + a[7] * b[3];
2737     out[2] += 0.21850969f * t;
2738     out[12] -= 0.14304817f * t;
2739     out[14] += 0.18467439f * t;
2740
2741     ta = -0.05839917f * a[13] + 0.22617901f * a[15];
2742     tb = -0.05839917f * b[13] + 0.22617901f * b[15];
2743     out[3] += ta * b[8] + tb * a[8];
2744     out[8] += ta * b[3] + tb * a[3];
2745     t = a[3] * b[8] + a[8] * b[3];
2746     out[13] -= 0.05839917f * t;
2747     out[15] += 0.22617901f * t;
2748
2749     ta = 0.28209479f * a[0] - 0.18022375f * a[6];
2750     tb = 0.28209479f * b[0] - 0.18022375f * b[6];
2751     out[4] += ta * b[4] + tb * a[4];
2752     t = a[4] * b[4];
2753     out[0] += 0.28209479f * t;
2754     out[6] -= 0.18022375f * t;
2755
2756     ta = 0.15607835f * a[7];
2757     tb = 0.15607835f * b[7];
2758     out[4] += ta * b[5] + tb * a[5];
2759     out[5] += ta * b[4] + tb * a[4];
2760     t = a[4] * b[5] + a[5] * b[4];
2761     out[7] += 0.15607835f * t;
2762
2763     ta = 0.22617901f * a[3] - 0.09403160f * a[13];
2764     tb = 0.22617901f * b[3] - 0.09403160f * b[13];
2765     out[4] += ta * b[9] + tb * a[9];
2766     out[9] += ta * b[4] + tb * a[4];
2767     t = a[4] * b[9] + a[9] * b[4];
2768     out[3] += 0.22617901f * t;
2769     out[13] -= 0.09403160f * t;
2770
2771     ta = 0.18467439f * a[2] - 0.18806319f * a[12];
2772     tb = 0.18467439f * b[2] - 0.18806319f * b[12];
2773     out[4] += ta * b[10] + tb * a [10];
2774     out[10] = ta * b[4] + tb * a[4];
2775     t = a[4] * b[10] + a[10] * b[4];
2776     out[2] += 0.18467439f * t;
2777     out[12] -= 0.18806319f * t;
2778
2779     ta = -0.05839917f * a[3] + 0.14567312f * a[13] + 0.09403160f * a[15];
2780     tb = -0.05839917f * b[3] + 0.14567312f * b[13] + 0.09403160f * b[15];
2781     out[4] += ta * b[11] + tb * a[11];
2782     out[11] += ta * b[4] + tb * a[4];
2783     t = a[4] * b[11] + a[11] * b[4];
2784     out[3] -= 0.05839917f * t;
2785     out[13] += 0.14567312f * t;
2786     out[15] += 0.09403160f * t;
2787
2788     ta = 0.28209479f * a[0] + 0.09011186f * a[6] - 0.15607835f * a[8];
2789     tb = 0.28209479f * b[0] + 0.09011186f * b[6] - 0.15607835f * b[8];
2790     out[5] += ta * b[5] + tb * a[5];
2791     t = a[5] * b[5];
2792     out[0] += 0.28209479f * t;
2793     out[6] += 0.09011186f * t;
2794     out[8] -= 0.15607835f * t;
2795
2796     ta = 0.14867701f * a[14];
2797     tb = 0.14867701f * b[14];
2798     out[5] += ta * b[9] + tb * a[9];
2799     out[9] += ta * b[5] + tb * a[5];
2800     t = a[5] * b[9] + a[9] * b[5];
2801     out[14] += 0.14867701f * t;
2802
2803     ta = 0.18467439f * a[3] + 0.11516472f * a[13] - 0.14867701f * a[15];
2804     tb = 0.18467439f * b[3] + 0.11516472f * b[13] - 0.14867701f * b[15];
2805     out[5] += ta * b[10] + tb * a[10];
2806     out[10] += ta * b[5] + tb * a[5];
2807     t = a[5] * b[10] + a[10] * b[5];
2808     out[3] += 0.18467439f * t;
2809     out[13] += 0.11516472f * t;
2810     out[15] -= 0.14867701f * t;
2811
2812     ta = 0.23359668f * a[2] + 0.05947080f * a[12] - 0.11516472f * a[14];
2813     tb = 0.23359668f * b[2] + 0.05947080f * b[12] - 0.11516472f * b[14];
2814     out[5] += ta * b[11] + tb * a[11];
2815     out[11] += ta * b[5] + tb * a[5];
2816     t = a[5] * b[11] + a[11] * b[5];
2817     out[2] += 0.23359668f * t;
2818     out[12] += 0.05947080f * t;
2819     out[14] -= 0.11516472f * t;
2820
2821     ta = 0.28209479f * a[0];
2822     tb = 0.28209479f * b[0];
2823     out[6] += ta * b[6] + tb * a[6];
2824     t = a[6] * b[6];
2825     out[0] += 0.28209479f * t;
2826     out[6] += 0.18022376f * t;
2827
2828     ta = 0.09011186f * a[6] + 0.28209479f * a[0] + 0.15607835f * a[8];
2829     tb = 0.09011186f * b[6] + 0.28209479f * b[0] + 0.15607835f * b[8];
2830     out[7] += ta * b[7] + tb * a[7];
2831     t = a[7] * b[7];
2832     out[6] += 0.09011186f * t;
2833     out[0] += 0.28209479f * t;
2834     out[8] += 0.15607835f * t;
2835
2836     ta = 0.14867701f * a[9] + 0.18467439f * a[1] + 0.11516472f * a[11];
2837     tb = 0.14867701f * b[9] + 0.18467439f * b[1] + 0.11516472f * b[11];
2838     out[7] += ta * b[10] + tb * a[10];
2839     out[10] += ta * b[7] + tb * a[7];
2840     t = a[7] * b[10] + a[10] * b[7];
2841     out[9] += 0.14867701f * t;
2842     out[1] += 0.18467439f * t;
2843     out[11] += 0.11516472f * t;
2844
2845     ta = 0.05947080f * a[12] + 0.23359668f * a[2] + 0.11516472f * a[14];
2846     tb = 0.05947080f * b[12] + 0.23359668f * b[2] + 0.11516472f * b[14];
2847     out[7] += ta * b[13] + tb * a[13];
2848     out[13] += ta * b[7]+ tb * a[7];
2849     t = a[7] * b[13] + a[13] * b[7];
2850     out[12] += 0.05947080f * t;
2851     out[2] += 0.23359668f * t;
2852     out[14] += 0.11516472f * t;
2853
2854     ta = 0.14867701f * a[15];
2855     tb = 0.14867701f * b[15];
2856     out[7] += ta * b[14] + tb * a[14];
2857     out[14] += ta * b[7] + tb * a[7];
2858     t = a[7] * b[14] + a[14] * b[7];
2859     out[15] += 0.14867701f * t;
2860
2861     ta = 0.28209479f * a[0] - 0.18022375f * a[6];
2862     tb = 0.28209479f * b[0] - 0.18022375f * b[6];
2863     out[8] += ta * b[8] + tb * a[8];
2864     t = a[8] * b[8];
2865     out[0] += 0.28209479f * t;
2866     out[6] -= 0.18022375f * t;
2867
2868     ta = -0.09403160f * a[11];
2869     tb = -0.09403160f * b[11];
2870     out[8] += ta * b[9] + tb * a[9];
2871     out[9] += ta * b[8] + tb * a[8];
2872     t = a[8] * b[9] + a[9] * b[8];
2873     out[11] -= 0.09403160f * t;
2874
2875     ta = -0.09403160f * a[15];
2876     tb = -0.09403160f * b[15];
2877     out[8] += ta * b[13] + tb * a[13];
2878     out[13] += ta * b[8] + tb * a[8];
2879     t = a[8] * b[13] + a[13] * b[8];
2880     out[15] -= 0.09403160f * t;
2881
2882     ta = 0.18467439f * a[2] - 0.18806319f * a[12];
2883     tb = 0.18467439f * b[2] - 0.18806319f * b[12];
2884     out[8] += ta * b[14] + tb * a[14];
2885     out[14] += ta * b[8] + tb * a[8];
2886     t = a[8] * b[14] + a[14] * b[8];
2887     out[2] += 0.18467439f * t;
2888     out[12] -= 0.18806319f * t;
2889
2890     ta = -0.21026104f * a[6] + 0.28209479f * a[0];
2891     tb = -0.21026104f * b[6] + 0.28209479f * b[0];
2892     out[9] += ta * b[9] + tb * a[9];
2893     t = a[9] * b[9];
2894     out[6] -= 0.21026104f * t;
2895     out[0] += 0.28209479f * t;
2896
2897     ta = 0.28209479f * a[0];
2898     tb = 0.28209479f * b[0];
2899     out[10] += ta * b[10] + tb * a[10];
2900     t = a[10] * b[10];
2901     out[0] += 0.28209479f * t;
2902
2903     ta = 0.28209479f * a[0] + 0.12615663f * a[6] - 0.14567312f * a[8];
2904     tb = 0.28209479f * b[0] + 0.12615663f * b[6] - 0.14567312f * b[8];
2905     out[11] += ta * b[11] + tb * a[11];
2906     t = a[11] * b[11];
2907     out[0] += 0.28209479f * t;
2908     out[6] += 0.12615663f * t;
2909     out[8] -= 0.14567312f * t;
2910
2911     ta = 0.28209479f * a[0] + 0.16820885f * a[6];
2912     tb = 0.28209479f * b[0] + 0.16820885f * b[6];
2913     out[12] += ta * b[12] + tb * a[12];
2914     t = a[12] * b[12];
2915     out[0] += 0.28209479f * t;
2916     out[6] += 0.16820885f * t;
2917
2918     ta =0.28209479f * a[0] + 0.14567312f * a[8] + 0.12615663f * a[6];
2919     tb =0.28209479f * b[0] + 0.14567312f * b[8] + 0.12615663f * b[6];
2920     out[13] += ta * b[13] + tb * a[13];
2921     t = a[13] * b[13];
2922     out[0] += 0.28209479f * t;
2923     out[8] += 0.14567312f * t;
2924     out[6] += 0.12615663f * t;
2925
2926     ta = 0.28209479f * a[0];
2927     tb = 0.28209479f * b[0];
2928     out[14] += ta * b[14] + tb * a[14];
2929     t = a[14] * b[14];
2930     out[0] += 0.28209479f * t;
2931
2932     ta = 0.28209479f * a[0] - 0.21026104f * a[6];
2933     tb = 0.28209479f * b[0] - 0.21026104f * b[6];
2934     out[15] += ta * b[15] + tb * a[15];
2935     t = a[15] * b[15];
2936     out[0] += 0.28209479f * t;
2937     out[6] -= 0.21026104f * t;
2938
2939     return out;
2940 }
2941
2942 static void rotate_X(FLOAT *out, UINT order, FLOAT a, FLOAT *in)
2943 {
2944     out[0] = in[0];
2945
2946     out[1] = a * in[2];
2947     out[2] = -a * in[1];
2948     out[3] = in[3];
2949
2950     out[4] = a * in[7];
2951     out[5] = -in[5];
2952     out[6] = -0.5f * in[6] - 0.8660253882f * in[8];
2953     out[7] = -a * in[4];
2954     out[8] = -0.8660253882f * in[6] + 0.5f * in[8];
2955     out[9] = -a * 0.7905694842f * in[12] + a * 0.6123724580f * in[14];
2956
2957     out[10] = -in[10];
2958     out[11] = -a * 0.6123724580f * in[12] - a * 0.7905694842f * in[14];
2959     out[12] = a * 0.7905694842f * in[9] + a * 0.6123724580f * in[11];
2960     out[13] = -0.25f * in[13] - 0.9682458639f * in[15];
2961     out[14] = -a * 0.6123724580f * in[9] + a * 0.7905694842f * in[11];
2962     out[15] = -0.9682458639f * in[13] + 0.25f * in[15];
2963     if (order == 4)
2964         return;
2965
2966     out[16] = -a * 0.9354143739f * in[21] + a * 0.3535533845f * in[23];
2967     out[17] = -0.75f * in[17] + 0.6614378095f * in[19];
2968     out[18] = -a * 0.3535533845f * in[21] - a * 0.9354143739f * in[23];
2969     out[19] = 0.6614378095f * in[17] + 0.75f * in[19];
2970     out[20] = 0.375f * in[20] + 0.5590170026f * in[22] + 0.7395099998f * in[24];
2971     out[21] = a * 0.9354143739f * in[16] + a * 0.3535533845f * in[18];
2972     out[22] = 0.5590170026f * in[20] + 0.5f * in[22] - 0.6614378691f * in[24];
2973     out[23] = -a * 0.3535533845f * in[16] + a * 0.9354143739f * in[18];
2974     out[24] = 0.7395099998f * in[20] - 0.6614378691f * in[22] + 0.125f * in[24];
2975     if (order == 5)
2976         return;
2977
2978     out[25] = a * 0.7015607357f * in[30] - a * 0.6846531630f * in[32] + a * 0.1976423711f * in[34];
2979     out[26] = -0.5f * in[26] + 0.8660253882f * in[28];
2980     out[27] = a * 0.5229125023f * in[30] + a * 0.3061861992f * in[32] - a * 0.7954951525f * in[34];
2981     out[28] = 0.8660253882f * in[26] + 0.5f * in[28];
2982     out[29] = a * 0.4841229022f * in[30] + a * 0.6614378691f * in[32] + a * 0.5728219748f * in[34];
2983     out[30] = -a * 0.7015607357f * in[25] - a * 0.5229125023f * in[27] - a * 0.4841229022f * in[29];
2984     out[31] = 0.125f * in[31] + 0.4050463140f * in[33] + 0.9057110548f * in[35];
2985     out[32] = a * 0.6846531630f * in[25] - a * 0.3061861992f * in[27] - a * 0.6614378691f * in[29];
2986     out[33] = 0.4050463140f * in[31] + 0.8125f * in[33] - 0.4192627370f * in[35];
2987     out[34] = -a * 0.1976423711f * in[25] + a * 0.7954951525f * in[27] - a * 0.5728219748f * in[29];
2988     out[35] = 0.9057110548f * in[31] - 0.4192627370f * in[33] + 0.0624999329f * in[35];
2989 }
2990
2991 FLOAT* WINAPI D3DXSHRotate(FLOAT *out, UINT order, const D3DXMATRIX *matrix, const FLOAT *in)
2992 {
2993     FLOAT alpha, beta, gamma, sinb, temp[36], temp1[36];
2994
2995     TRACE("out %p, order %u, matrix %p, in %p\n", out, order, matrix, in);
2996
2997     out[0] = in[0];
2998
2999     if ((order > D3DXSH_MAXORDER) || (order < D3DXSH_MINORDER))
3000         return out;
3001
3002     if (order <= 3)
3003     {
3004         out[1] = matrix->u.m[1][1] * in[1] - matrix->u.m[2][1] * in[2] + matrix->u.m[0][1] * in[3];
3005         out[2] = -matrix->u.m[1][2] * in[1] + matrix->u.m[2][2] * in[2] - matrix->u.m[0][2] * in[3];
3006         out[3] = matrix->u.m[1][0] * in[1] - matrix->u.m[2][0] * in[2] + matrix->u.m[0][0] * in[3];
3007
3008         if (order == 3)
3009         {
3010             FLOAT coeff[]={
3011                 matrix->u.m[1][0] * matrix->u.m[0][0], matrix->u.m[1][1] * matrix->u.m[0][1],
3012                 matrix->u.m[1][1] * matrix->u.m[2][1], matrix->u.m[1][0] * matrix->u.m[2][0],
3013                 matrix->u.m[2][0] * matrix->u.m[2][0], matrix->u.m[2][1] * matrix->u.m[2][1],
3014                 matrix->u.m[0][0] * matrix->u.m[2][0], matrix->u.m[0][1] * matrix->u.m[2][1],
3015                 matrix->u.m[0][1] * matrix->u.m[0][1], matrix->u.m[1][0] * matrix->u.m[1][0],
3016                 matrix->u.m[1][1] * matrix->u.m[1][1], matrix->u.m[0][0] * matrix->u.m[0][0], };
3017
3018             out[4] = (matrix->u.m[1][1] * matrix->u.m[0][0] + matrix->u.m[0][1] * matrix->u.m[1][0]) * in[4];
3019             out[4] -= (matrix->u.m[1][0] * matrix->u.m[2][1] + matrix->u.m[1][1] * matrix->u.m[2][0]) * in[5];
3020             out[4] += 1.7320508076f * matrix->u.m[2][0] * matrix->u.m[2][1] * in[6];
3021             out[4] -= (matrix->u.m[0][1] * matrix->u.m[2][0] + matrix->u.m[0][0] * matrix->u.m[2][1]) * in[7];
3022             out[4] += (matrix->u.m[0][0] * matrix->u.m[0][1] - matrix->u.m[1][0] * matrix->u.m[1][1]) * in[8];
3023
3024             out[5] = (matrix->u.m[1][1] * matrix->u.m[2][2] + matrix->u.m[1][2] * matrix->u.m[2][1]) * in[5];
3025             out[5] -= (matrix->u.m[1][1] * matrix->u.m[0][2] + matrix->u.m[1][2] * matrix->u.m[0][1]) * in[4];
3026             out[5] -= 1.7320508076f * matrix->u.m[2][2] * matrix->u.m[2][1] * in[6];
3027             out[5] += (matrix->u.m[0][2] * matrix->u.m[2][1] + matrix->u.m[0][1] * matrix->u.m[2][2]) * in[7];
3028             out[5] -= (matrix->u.m[0][1] * matrix->u.m[0][2] - matrix->u.m[1][1] * matrix->u.m[1][2]) * in[8];
3029
3030             out[6] = (matrix->u.m[2][2] * matrix->u.m[2][2] - 0.5f * (coeff[4] + coeff[5])) * in[6];
3031             out[6] -= (0.5773502692f * (coeff[0] + coeff[1]) - 1.1547005384f * matrix->u.m[1][2] * matrix->u.m[0][2]) * in[4];
3032             out[6] += (0.5773502692f * (coeff[2] + coeff[3]) - 1.1547005384f * matrix->u.m[1][2] * matrix->u.m[2][2]) * in[5];
3033             out[6] += (0.5773502692f * (coeff[6] + coeff[7]) - 1.1547005384f * matrix->u.m[0][2] * matrix->u.m[2][2]) * in[7];
3034             out[6] += (0.2886751347f * (coeff[9] - coeff[8] + coeff[10] - coeff[11]) - 0.5773502692f *
3035                   (matrix->u.m[1][2] * matrix->u.m[1][2] - matrix->u.m[0][2] * matrix->u.m[0][2])) * in[8];
3036
3037             out[7] = (matrix->u.m[0][0] * matrix->u.m[2][2] + matrix->u.m[0][2] * matrix->u.m[2][0]) * in[7];
3038             out[7] -= (matrix->u.m[1][0] * matrix->u.m[0][2] + matrix->u.m[1][2] * matrix->u.m[0][0]) * in[4];
3039             out[7] += (matrix->u.m[1][0] * matrix->u.m[2][2] + matrix->u.m[1][2] * matrix->u.m[2][0]) * in[5];
3040             out[7] -= 1.7320508076f * matrix->u.m[2][2] * matrix->u.m[2][0] * in[6];
3041             out[7] -= (matrix->u.m[0][0] * matrix->u.m[0][2] - matrix->u.m[1][0] * matrix->u.m[1][2]) * in[8];
3042
3043             out[8] = 0.5f * (coeff[11] - coeff[8] - coeff[9] + coeff[10]) * in[8];
3044             out[8] += (coeff[0] - coeff[1]) * in[4];
3045             out[8] += (coeff[2] - coeff[3]) * in[5];
3046             out[8] += 0.86602540f * (coeff[4] - coeff[5]) * in[6];
3047             out[8] += (coeff[7] - coeff[6]) * in[7];
3048         }
3049
3050         return out;
3051     }
3052
3053     if (fabsf(matrix->u.m[2][2]) != 1.0f)
3054     {
3055         sinb = sqrtf(1.0f - matrix->u.m[2][2] * matrix->u.m[2][2]);
3056         alpha = atan2f(matrix->u.m[2][1] / sinb, matrix->u.m[2][0] / sinb);
3057         beta = atan2f(sinb, matrix->u.m[2][2]);
3058         gamma = atan2f(matrix->u.m[1][2] / sinb, -matrix->u.m[0][2] / sinb);
3059     }
3060     else
3061     {
3062         alpha = atan2f(matrix->u.m[0][1], matrix->u.m[0][0]);
3063         beta = 0.0f;
3064         gamma = 0.0f;
3065     }
3066
3067     D3DXSHRotateZ(temp, order, gamma, in);
3068     rotate_X(temp1, order, 1.0f, temp);
3069     D3DXSHRotateZ(temp, order, beta, temp1);
3070     rotate_X(temp1, order, -1.0f, temp);
3071     D3DXSHRotateZ(out, order, alpha, temp1);
3072
3073     return out;
3074 }
3075
3076 FLOAT * WINAPI D3DXSHRotateZ(FLOAT *out, UINT order, FLOAT angle, const FLOAT *in)
3077 {
3078     UINT i, sum = 0;
3079     FLOAT c[5], s[5];
3080
3081     TRACE("out %p, order %u, angle %f, in %p\n", out, order, angle, in);
3082
3083     order = min(max(order, D3DXSH_MINORDER), D3DXSH_MAXORDER);
3084
3085     out[0] = in[0];
3086
3087     for (i = 1; i < order; i++)
3088     {
3089         UINT j;
3090
3091         c[i - 1] = cosf(i * angle);
3092         s[i - 1] = sinf(i * angle);
3093         sum += i * 2;
3094
3095         out[sum - i] = c[i - 1] * in[sum - i];
3096         out[sum - i] += s[i - 1] * in[sum + i];
3097         for (j = i - 1; j > 0; j--)
3098         {
3099             out[sum - j] = 0.0f;
3100             out[sum - j] = c[j - 1] * in[sum - j];
3101             out[sum - j] += s[j - 1] * in[sum + j];
3102         }
3103
3104         if (in == out)
3105             out[sum] = 0.0f;
3106         else
3107             out[sum] = in[sum];
3108
3109         for (j = 1; j < i; j++)
3110         {
3111             out[sum + j] = 0.0f;
3112             out[sum + j] = -s[j - 1] * in[sum - j];
3113             out[sum + j] += c[j - 1] * in[sum + j];
3114         }
3115         out[sum + i] = -s[i - 1] * in[sum - i];
3116         out[sum + i] += c[i - 1] * in[sum + i];
3117     }
3118
3119     return out;
3120 }
3121
3122 FLOAT* WINAPI D3DXSHScale(FLOAT *out, UINT order, const FLOAT *a, const FLOAT scale)
3123 {
3124     UINT i;
3125
3126     TRACE("out %p, order %u, a %p, scale %f\n", out, order, a, scale);
3127
3128     for (i = 0; i < order * order; i++)
3129         out[i] = a[i] * scale;
3130
3131     return out;
3132 }