diff options
Diffstat (limited to 'indra/newview/noise.h')
-rw-r--r-- | indra/newview/noise.h | 416 |
1 files changed, 208 insertions, 208 deletions
diff --git a/indra/newview/noise.h b/indra/newview/noise.h index b3efad73c5..ae819cf542 100644 --- a/indra/newview/noise.h +++ b/indra/newview/noise.h @@ -1,25 +1,25 @@ -/** +/** * @file noise.h * @brief Perlin noise routines for procedural textures, etc * * $LicenseInfo:firstyear=2000&license=viewerlgpl$ * Second Life Viewer Source Code * Copyright (C) 2010, Linden Research, Inc. - * + * * This library is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; * version 2.1 of the License only. - * + * * This library is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. - * + * * You should have received a copy of the GNU Lesser General Public * License along with this library; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - * + * * Linden Research, Inc., 945 Battery Street, San Francisco, CA 94111 USA * $/LicenseInfo$ */ @@ -37,65 +37,65 @@ F32 noise3(float *vec); inline F32 bias(F32 a, F32 b) { - return (F32)pow(a, (F32)(log(b) / log(0.5f))); + return (F32)pow(a, (F32)(log(b) / log(0.5f))); } inline F32 gain(F32 a, F32 b) { - F32 p = (F32) (log(1.f - b) / log(0.5f)); - - if (a < .001f) - return 0.f; - else if (a > .999f) - return 1.f; - if (a < 0.5f) - return (F32)(pow(2 * a, p) / 2.f); - else - return (F32)(1.f - pow(2 * (1.f - a), p) / 2.f); + F32 p = (F32) (log(1.f - b) / log(0.5f)); + + if (a < .001f) + return 0.f; + else if (a > .999f) + return 1.f; + if (a < 0.5f) + return (F32)(pow(2 * a, p) / 2.f); + else + return (F32)(1.f - pow(2 * (1.f - a), p) / 2.f); } inline F32 turbulence2(F32 *v, F32 freq) { - F32 t, vec[2]; - - for (t = 0.f ; freq >= 1.f ; freq *= 0.5f) { - vec[0] = freq * v[0]; - vec[1] = freq * v[1]; - t += noise2(vec)/freq; - } - return t; + F32 t, vec[2]; + + for (t = 0.f ; freq >= 1.f ; freq *= 0.5f) { + vec[0] = freq * v[0]; + vec[1] = freq * v[1]; + t += noise2(vec)/freq; + } + return t; } inline F32 turbulence3(F32 *v, F32 freq) { - F32 t, vec[3]; - - for (t = 0.f ; freq >= 1.f ; freq *= 0.5f) { - vec[0] = freq * v[0]; - vec[1] = freq * v[1]; - vec[2] = freq * v[2]; - t += noise3(vec)/freq; -// t += fabs(noise3(vec)) / freq; // Like snow - bubbly at low frequencies -// t += sqrt(fabs(noise3(vec))) / freq; // Better at low freq -// t += (noise3(vec)*noise3(vec)) / freq; - } - return t; + F32 t, vec[3]; + + for (t = 0.f ; freq >= 1.f ; freq *= 0.5f) { + vec[0] = freq * v[0]; + vec[1] = freq * v[1]; + vec[2] = freq * v[2]; + t += noise3(vec)/freq; +// t += fabs(noise3(vec)) / freq; // Like snow - bubbly at low frequencies +// t += sqrt(fabs(noise3(vec))) / freq; // Better at low freq +// t += (noise3(vec)*noise3(vec)) / freq; + } + return t; } inline F32 clouds3(F32 *v, F32 freq) { - F32 t, vec[3]; - - for (t = 0.f ; freq >= 1.f ; freq *= 0.5f) { - vec[0] = freq * v[0]; - vec[1] = freq * v[1]; - vec[2] = freq * v[2]; - //t += noise3(vec)/freq; -// t += fabs(noise3(vec)) / freq; // Like snow - bubbly at low frequencies -// t += sqrt(fabs(noise3(vec))) / freq; // Better at low freq - t += (noise3(vec)*noise3(vec)) / freq; - } - return t; + F32 t, vec[3]; + + for (t = 0.f ; freq >= 1.f ; freq *= 0.5f) { + vec[0] = freq * v[0]; + vec[1] = freq * v[1]; + vec[2] = freq * v[2]; + //t += noise3(vec)/freq; +// t += fabs(noise3(vec)) / freq; // Like snow - bubbly at low frequencies +// t += sqrt(fabs(noise3(vec))) / freq; // Better at low freq + t += (noise3(vec)*noise3(vec)) / freq; + } + return t; } /* noise functions over 1, 2, and 3 dimensions */ @@ -121,230 +121,230 @@ static void init(void); #define lerp_m(t, a, b) ( a + t * (b - a) ) #define setup_noise(i,b0,b1,r0,r1)\ - t = vec[i] + N;\ - b0 = (lltrunc(t)) & BM;\ - b1 = (b0+1) & BM;\ - r0 = t - lltrunc(t);\ - r1 = r0 - 1.f; + t = vec[i] + N;\ + b0 = (lltrunc(t)) & BM;\ + b1 = (b0+1) & BM;\ + r0 = t - lltrunc(t);\ + r1 = r0 - 1.f; inline void fast_setup(F32 vec, U8 &b0, U8 &b1, F32 &r0, F32 &r1) { - S32 t_S32; - - r1 = vec + NF32; - t_S32 = lltrunc(r1); - b0 = (U8)t_S32; - b1 = b0 + 1; - r0 = r1 - t_S32; - r1 = r0 - 1.f; + S32 t_S32; + + r1 = vec + NF32; + t_S32 = lltrunc(r1); + b0 = (U8)t_S32; + b1 = b0 + 1; + r0 = r1 - t_S32; + r1 = r0 - 1.f; } inline F32 noise1(const F32 arg) { - int bx0, bx1; - F32 rx0, rx1, sx, t, u, v, vec[1]; + int bx0, bx1; + F32 rx0, rx1, sx, t, u, v, vec[1]; - vec[0] = arg; - if (gNoiseStart) { - gNoiseStart = 0; - init(); - } + vec[0] = arg; + if (gNoiseStart) { + gNoiseStart = 0; + init(); + } - setup_noise(0, bx0,bx1, rx0,rx1); + setup_noise(0, bx0,bx1, rx0,rx1); - sx = s_curve(rx0); + sx = s_curve(rx0); - u = rx0 * g1[ p[ bx0 ] ]; - v = rx1 * g1[ p[ bx1 ] ]; + u = rx0 * g1[ p[ bx0 ] ]; + v = rx1 * g1[ p[ bx1 ] ]; - return lerp_m(sx, u, v); + return lerp_m(sx, u, v); } inline F32 fast_at2(F32 rx, F32 ry, F32 *q) { - return rx * (*q) + ry * (*(q + 1)); + return rx * (*q) + ry * (*(q + 1)); } inline F32 fast_at3(F32 rx, F32 ry, F32 rz, F32 *q) { - return rx * (*q) + ry * (*(q + 1)) + rz * (*(q + 2)); + return rx * (*q) + ry * (*(q + 1)) + rz * (*(q + 2)); } inline F32 noise3(F32 *vec) { - U8 bx0, bx1, by0, by1, bz0, bz1; - S32 b00, b10, b01, b11; - F32 rx0, rx1, ry0, ry1, rz0, rz1, *q, sy, sz, a, b, c, d, t, u, v; - S32 i, j; - - if (gNoiseStart) { - gNoiseStart = 0; - init(); - } - - fast_setup(*vec, bx0,bx1, rx0,rx1); - fast_setup(*(vec + 1), by0,by1, ry0,ry1); - fast_setup(*(vec + 2), bz0,bz1, rz0,rz1); - - i = p[ bx0 ]; - j = p[ bx1 ]; - - b00 = p[ i + by0 ]; - b10 = p[ j + by0 ]; - b01 = p[ i + by1 ]; - b11 = p[ j + by1 ]; - - t = s_curve(rx0); - sy = s_curve(ry0); - sz = s_curve(rz0); - - q = g3[ b00 + bz0 ]; - u = fast_at3(rx0,ry0,rz0,q); - q = g3[ b10 + bz0 ]; - v = fast_at3(rx1,ry0,rz0,q); - a = lerp_m(t, u, v); - - q = g3[ b01 + bz0 ]; - u = fast_at3(rx0,ry1,rz0,q); - q = g3[ b11 + bz0 ]; - v = fast_at3(rx1,ry1,rz0,q); - b = lerp_m(t, u, v); - - c = lerp_m(sy, a, b); - - q = g3[ b00 + bz1 ]; - u = fast_at3(rx0,ry0,rz1,q); - q = g3[ b10 + bz1 ]; - v = fast_at3(rx1,ry0,rz1,q); - a = lerp_m(t, u, v); - - q = g3[ b01 + bz1 ]; - u = fast_at3(rx0,ry1,rz1,q); - q = g3[ b11 + bz1 ]; - v = fast_at3(rx1,ry1,rz1,q); - b = lerp_m(t, u, v); - - d = lerp_m(sy, a, b); - - return lerp_m(sz, c, d); + U8 bx0, bx1, by0, by1, bz0, bz1; + S32 b00, b10, b01, b11; + F32 rx0, rx1, ry0, ry1, rz0, rz1, *q, sy, sz, a, b, c, d, t, u, v; + S32 i, j; + + if (gNoiseStart) { + gNoiseStart = 0; + init(); + } + + fast_setup(*vec, bx0,bx1, rx0,rx1); + fast_setup(*(vec + 1), by0,by1, ry0,ry1); + fast_setup(*(vec + 2), bz0,bz1, rz0,rz1); + + i = p[ bx0 ]; + j = p[ bx1 ]; + + b00 = p[ i + by0 ]; + b10 = p[ j + by0 ]; + b01 = p[ i + by1 ]; + b11 = p[ j + by1 ]; + + t = s_curve(rx0); + sy = s_curve(ry0); + sz = s_curve(rz0); + + q = g3[ b00 + bz0 ]; + u = fast_at3(rx0,ry0,rz0,q); + q = g3[ b10 + bz0 ]; + v = fast_at3(rx1,ry0,rz0,q); + a = lerp_m(t, u, v); + + q = g3[ b01 + bz0 ]; + u = fast_at3(rx0,ry1,rz0,q); + q = g3[ b11 + bz0 ]; + v = fast_at3(rx1,ry1,rz0,q); + b = lerp_m(t, u, v); + + c = lerp_m(sy, a, b); + + q = g3[ b00 + bz1 ]; + u = fast_at3(rx0,ry0,rz1,q); + q = g3[ b10 + bz1 ]; + v = fast_at3(rx1,ry0,rz1,q); + a = lerp_m(t, u, v); + + q = g3[ b01 + bz1 ]; + u = fast_at3(rx0,ry1,rz1,q); + q = g3[ b11 + bz1 ]; + v = fast_at3(rx1,ry1,rz1,q); + b = lerp_m(t, u, v); + + d = lerp_m(sy, a, b); + + return lerp_m(sz, c, d); } /* F32 noise3(F32 *vec) { - int bx0, bx1, by0, by1, bz0, bz1, b00, b10, b01, b11; - F32 rx0, rx1, ry0, ry1, rz0, rz1, *q, sy, sz, a, b, c, d, t, u, v; - S32 i, j; + int bx0, bx1, by0, by1, bz0, bz1, b00, b10, b01, b11; + F32 rx0, rx1, ry0, ry1, rz0, rz1, *q, sy, sz, a, b, c, d, t, u, v; + S32 i, j; - if (gNoiseStart) { - gNoiseStart = 0; - init(); - } + if (gNoiseStart) { + gNoiseStart = 0; + init(); + } - setup_noise(0, bx0,bx1, rx0,rx1); - setup_noise(1, by0,by1, ry0,ry1); - setup_noise(2, bz0,bz1, rz0,rz1); + setup_noise(0, bx0,bx1, rx0,rx1); + setup_noise(1, by0,by1, ry0,ry1); + setup_noise(2, bz0,bz1, rz0,rz1); - i = p[ bx0 ]; - j = p[ bx1 ]; + i = p[ bx0 ]; + j = p[ bx1 ]; - b00 = p[ i + by0 ]; - b10 = p[ j + by0 ]; - b01 = p[ i + by1 ]; - b11 = p[ j + by1 ]; + b00 = p[ i + by0 ]; + b10 = p[ j + by0 ]; + b01 = p[ i + by1 ]; + b11 = p[ j + by1 ]; - t = s_curve(rx0); - sy = s_curve(ry0); - sz = s_curve(rz0); + t = s_curve(rx0); + sy = s_curve(ry0); + sz = s_curve(rz0); #define at3(rx,ry,rz) ( rx * q[0] + ry * q[1] + rz * q[2] ) - q = g3[ b00 + bz0 ] ; u = at3(rx0,ry0,rz0); - q = g3[ b10 + bz0 ] ; v = at3(rx1,ry0,rz0); - a = lerp_m(t, u, v); + q = g3[ b00 + bz0 ] ; u = at3(rx0,ry0,rz0); + q = g3[ b10 + bz0 ] ; v = at3(rx1,ry0,rz0); + a = lerp_m(t, u, v); - q = g3[ b01 + bz0 ] ; u = at3(rx0,ry1,rz0); - q = g3[ b11 + bz0 ] ; v = at3(rx1,ry1,rz0); - b = lerp_m(t, u, v); + q = g3[ b01 + bz0 ] ; u = at3(rx0,ry1,rz0); + q = g3[ b11 + bz0 ] ; v = at3(rx1,ry1,rz0); + b = lerp_m(t, u, v); - c = lerp_m(sy, a, b); + c = lerp_m(sy, a, b); - q = g3[ b00 + bz1 ] ; u = at3(rx0,ry0,rz1); - q = g3[ b10 + bz1 ] ; v = at3(rx1,ry0,rz1); - a = lerp_m(t, u, v); + q = g3[ b00 + bz1 ] ; u = at3(rx0,ry0,rz1); + q = g3[ b10 + bz1 ] ; v = at3(rx1,ry0,rz1); + a = lerp_m(t, u, v); - q = g3[ b01 + bz1 ] ; u = at3(rx0,ry1,rz1); - q = g3[ b11 + bz1 ] ; v = at3(rx1,ry1,rz1); - b = lerp_m(t, u, v); + q = g3[ b01 + bz1 ] ; u = at3(rx0,ry1,rz1); + q = g3[ b11 + bz1 ] ; v = at3(rx1,ry1,rz1); + b = lerp_m(t, u, v); - d = lerp_m(sy, a, b); + d = lerp_m(sy, a, b); - return lerp_m(sz, c, d); + return lerp_m(sz, c, d); } */ static void normalize2(F32 v[2]) { - F32 s; + F32 s; - s = 1.f/(F32)sqrt(v[0] * v[0] + v[1] * v[1]); - v[0] = v[0] * s; - v[1] = v[1] * s; + s = 1.f/(F32)sqrt(v[0] * v[0] + v[1] * v[1]); + v[0] = v[0] * s; + v[1] = v[1] * s; } static void normalize3(F32 v[3]) { - F32 s; + F32 s; - s = 1.f/(F32)sqrt(v[0] * v[0] + v[1] * v[1] + v[2] * v[2]); - v[0] = v[0] * s; - v[1] = v[1] * s; - v[2] = v[2] * s; + s = 1.f/(F32)sqrt(v[0] * v[0] + v[1] * v[1] + v[2] * v[2]); + v[0] = v[0] * s; + v[1] = v[1] * s; + v[2] = v[2] * s; } static void init(void) { - // we want repeatable noise (e.g. for stable terrain texturing), so seed with known value - srand(42); - int i, j, k; - - for (i = 0 ; i < B ; i++) { - p[i] = i; - - g1[i] = (F32)((rand() % (B + B)) - B) / B; - - for (j = 0 ; j < 2 ; j++) - g2[i][j] = (F32)((rand() % (B + B)) - B) / B; - normalize2(g2[i]); - - for (j = 0 ; j < 3 ; j++) - g3[i][j] = (F32)((rand() % (B + B)) - B) / B; - normalize3(g3[i]); - } - - while (--i) { - k = p[i]; - p[i] = p[j = rand() % B]; - p[j] = k; - } - - for (i = 0 ; i < B + 2 ; i++) { - p[B + i] = p[i]; - g1[B + i] = g1[i]; - for (j = 0 ; j < 2 ; j++) - g2[B + i][j] = g2[i][j]; - for (j = 0 ; j < 3 ; j++) - g3[B + i][j] = g3[i][j]; - } - - // reintroduce entropy - srand(time(NULL)); // Flawfinder: ignore + // we want repeatable noise (e.g. for stable terrain texturing), so seed with known value + srand(42); + int i, j, k; + + for (i = 0 ; i < B ; i++) { + p[i] = i; + + g1[i] = (F32)((rand() % (B + B)) - B) / B; + + for (j = 0 ; j < 2 ; j++) + g2[i][j] = (F32)((rand() % (B + B)) - B) / B; + normalize2(g2[i]); + + for (j = 0 ; j < 3 ; j++) + g3[i][j] = (F32)((rand() % (B + B)) - B) / B; + normalize3(g3[i]); + } + + while (--i) { + k = p[i]; + p[i] = p[j = rand() % B]; + p[j] = k; + } + + for (i = 0 ; i < B + 2 ; i++) { + p[B + i] = p[i]; + g1[B + i] = g1[i]; + for (j = 0 ; j < 2 ; j++) + g2[B + i][j] = g2[i][j]; + for (j = 0 ; j < 3 ; j++) + g3[B + i][j] = g3[i][j]; + } + + // reintroduce entropy + srand(time(NULL)); // Flawfinder: ignore } #undef B |