Prefacenode
今天有兩個東東,一個是體積煙霧,一個是封面圖dom
下一篇咱們總結項目代碼ide
Chapter 8:Volumes函數
咱們須要爲咱們的光線追蹤器添加新的物體——煙、霧,也稱爲participating media。 咱們還須要補充一個材質——次表面散射材質,它有點像物體內的濃霧。測試
體渲染一般的作法是,在體的內部有不少隨機表面,來實現散射的效果。好比一束煙能夠表示爲,在這束煙的內部任意位置,均可以存在一個面,以此來實現煙、霧spa
對於一個恆定密度體,一條光線經過其中的時候,在煙霧體中傳播的時候也會發生散射,光線在煙霧體中能傳播多遠,也是由煙霧體的密度決定的,密度越高,光線穿透性越差,光線傳播的距離也越短。從而實現煙霧的透光性。code
引用書中一張圖(光線可穿透可散射)orm
當光線經過體積時,它可能在任何點散射。 光線在任何小距離dL中散射的機率爲:blog
機率= C * dL,其中C與體積的光密度成比例。ip
對於恆定體積,咱們只須要密度C和邊界。
/// isotropic.hpp // ----------------------------------------------------- // [author] lv // [begin ] 2019.1 // [brief ] the isotropic-class for the ray-tracing project // from the 《ray tracing the next week》 // ----------------------------------------------------- #pragma once namespace rt { class isotropic :public material { public: isotropic(texture* tex) :_albedo(tex) { } virtual bool scatter(const ray& InRay, const hitInfo& info, rtvec& attenuation, ray& scattered)const override { scattered = ray(info._p, lvgm::random_unit_sphere()); attenuation = _albedo->value(info._u, info._v, info._p); return true; } private: texture * _albedo; }; } // rt namespace
這個材質的散射原理和漫反射磨砂材質的大同小異,均屬於碰撞點轉換爲新視點,沿任意方向發射新的視線,只不過漫反射的視線方向向量指向外相切球體表面,而isotropic的視線方向指向以碰撞點爲球心的單位球體表面
區別就在於
漫反射的散射光線不可能指到物體內部,它必定是散射到表面外部(視線方向指向外切球體表面)
isotropic材質的散射光線能夠沿原來的方向一往前,以此視線透光性
由於煙霧內部只是顆粒而不存在真正不可穿透的幾何實體,因此漫反射實體不可穿透,只能散射到表面外部,而煙霧可穿透
接下來咱們看一下煙霧體
/// constant_medium.hpp // ----------------------------------------------------- // [author] lv // [begin ] 2019.1 // [brief ] the constant_dedium-class for the ray-tracing project // from the 《ray tracing the next week》 // ----------------------------------------------------- #pragma once namespace rt { class constant_medium :public intersect { public: constant_medium(intersect* p, rtvar d, texture* tex); virtual bool hit(const ray& sight, rtvar t_min, rtvar t_max, hitInfo& info)const override; virtual aabb getbox()const override; private: intersect* _item; rtvar _density; //煙霧密度 material* _materialp; }; inline constant_medium::constant_medium(intersect* p, rtvar d, texture* tex) :_item(p) ,_density(d) ,_materialp(new isotropic(tex)) { } aabb constant_medium::getbox()const { return _item->getbox(); } bool constant_medium::hit(const ray& sight, rtvar t_min, rtvar t_max, hitInfo& info)const { hitInfo info1, info2; if (_item->hit(sight, -rtInf(), rtInf(), info1)) { if (_item->hit(sight, info1._t + 0.0001, rtInf(), info2)) { if (info1._t < t_min) info1._t = t_min; if (info2._t > t_max) info2._t = t_max; if (info1._t >= info2._t) return false; if (info1._t < 0) info1._t = 0; float distance_inside_boundary = (info2._t - info1._t)*sight.direction().normal(); float hit_distance = -(1 / _density)*log(lvgm::rand01()); if (hit_distance < distance_inside_boundary) { info._t = info1._t + hit_distance / sight.direction().normal(); info._p = sight.go(info._t); info._n = rtvec(1, 0, 0); // arbitrary info._materialp = _materialp; return true; } } } return false; } } // rt namespace
hit函數裏面是一些邊界合法性檢測
場景測試代碼
intersect* cornell_smoke() { intersect ** list = new intersect*[9]; int cnt = 0; material* red = new lambertian(new constant_texture(rtvec(0.65, 0.05, 0.05))); material * blue = new lambertian(new constant_texture(rtvec(0.05, 0.05, 0.73))); material* white = new lambertian(new constant_texture(rtvec(0.73, 0.73, 0.73))); material* green = new lambertian(new constant_texture(rtvec(0.12, 0.45, 0.15))); material* light = new areaLight(new constant_texture(rtvec(7, 7, 7))); list[cnt++] = new xz_rect(113, 443, 127, 432, 550, light); list[cnt++] = new flip_normal(new xz_rect(113, 443, 127, 432, 550, light)); list[cnt++] = new flip_normal(new yz_rect(0, 555, 0, 555, 555, green)); list[cnt++] = new yz_rect(0, 555, 0, 555, 0, red); list[cnt++] = new flip_normal(new xz_rect(0, 555, 0, 555, 555, white)); list[cnt++] = new xz_rect(0, 555, 0, 555, 0, white); list[cnt++] = new flip_normal(new xy_rect(0, 555, 0, 555, 555, blue)); intersect* box1 = new translate(new rotate_y(new box(rtvec(), rtvec(165, 165, 165), white), -18), rtvec(130, 0, 65)); intersect* box2 = new translate(new rotate_y(new box(rtvec(), rtvec(165, 320, 165), white), 15), rtvec(265, 0, 295)); list[cnt++] = new constant_medium(box2, 0.006, new constant_texture(rtvec(0.8, 0.58, 0.))); list[cnt++] = new constant_medium(box1, 0.008, new constant_texture(rtvec(0.9, 0.2, 0.72))); return new intersections(list, cnt); }
下面是效果:sample -> 1000
註釋 // arbitrary處爲隨機方向,以前爲(1,0,0)
我以爲改成(rand01(),rand01(),rand01())較爲合適,下面是改以後的效果
Chapter 9:A Scene Testing All New Features
最後的封面圖是這樣一個場景:
體積霧:有一個藍色的次表面散射球體,可是這個東西並無單獨實現,而是把它包裹在了一個玻璃球內。
體積霧:整個場景是由層薄薄的霧氣籠蓋着的
長方體:地面是一堆隨機長方體
玻璃球:一個做爲藍色煙霧的容器,一個是單純的玻璃球
映射紋理:地球紋理球體
過程紋理:柏林噪聲紋理球體
運動模糊球體:有一個棕色的運動球體
鏡面球體:銀白色的鏡面球
區域光照:頂部是一個長方形的區域光源
其餘未說明材質的都是漫反射
渲染器中剩下的最大缺陷是沒有陰影射線,但這就是爲何咱們容易得到焦散和次表面散射效果的緣由。
下面是渲染代碼,關於相機參數設置還需等待渲染結果出來才能公佈
VS四開,渲染了一天還沒完,我也實屬無奈
intersect* finalScene() { int nb = 20; intersect ** list = new intersect*[30]; intersect ** boxlist = new intersect*[2000]; intersect ** boxlist2 = new intersect*[2000]; material * white = new lambertian(new constant_texture(rtvec(0.73, 0.73, 0.73))); material * ground = new lambertian(new constant_texture(rtvec(0.48, 0.83, 0.53))); int b = 0; for (int i = 0; i < nb; ++i) for (int j = 0; j < nb; ++j) { rtvar w = 100; rtvar x0 = -1000 + i*w; rtvar z0 = -1000 + j*w; rtvar y0 = 0; rtvar x1 = x0 + w; rtvar y1 = 100 * (lvgm::rand01() + 0.01); rtvar z1 = z0 + w; boxlist[b++] = new box(rtvec(x0, y0, z0), rtvec(x1, y1, z1), ground); } int l = 0; list[l++] = new bvh_node(boxlist, b, 0, 1); material * light = new areaLight(new constant_texture(rtvec(10, 10, 10))); list[l++] = new xz_rect(123, 423, 147, 412, 550, light); rtvec center(400, 400, 200); list[l++] = new moving_sphere(center, center + rtvec(30, 0, 0), 0, 1, 50, new lambertian(new constant_texture(rtvec(0.7, 0.3, 0.1)))); list[l++] = new sphere(rtvec(260, 150, 45), 50, new dielectric(1.5)); list[l++] = new sphere(rtvec(0, 150, 145), 50, new metal(new constant_texture(rtvec(0.8, 0.8, 0.9)), 10.0)); intersect * boundary = new sphere(rtvec(360, 150, 145), 70, new dielectric(1.5)); list[l++] = boundary; list[l++] = new constant_medium(boundary, 0.2, new constant_texture(rtvec(0.2, 0.4, 0.9))); boundary = new sphere(rtvec(), 5000, new dielectric(1.5)); list[l++] = new constant_medium(boundary, 0.0011, new constant_texture(rtvec(1., 1., 1.))); int x, y, n; unsigned char * tex = stbi_load("earthmap.jpg", &x, &y, &n, 0); material * emat = new lambertian(new image_texture(tex, x, y)); list[l++] = new sphere(rtvec(400, 200, 400), 100, emat); texture * pertext = new noise_texture(0.1); list[l++] = new sphere(rtvec(220, 280, 300), 80, new lambertian(pertext)); int ns = 1000; for (int j = 0; j < ns; ++j) boxlist2[j] = new sphere(rtvec(165 * lvgm::rand01(), 165 * lvgm::rand01(), lvgm::rand01()), 10, white); list[l++] = new translate(new rotate_y(new bvh_node(boxlist2, ns, 0, 1), 15), rtvec(-100, 270, 395)); return new intersections(list, l); }
/-----------------------更新線-------------------------------/
對應的相機參數
效果
其中,薄霧效果過重了,霧氣參數應該小一點,大約在1e-4左右較好
vfov可能太大了,45°應該更好一點吧
鏡頭應該更靠近些
感謝您的閱讀,生活愉快~