Practical Assignment 2

Deadline: 21.10.2021

Please put your name here:
Name: .......

Problem 1

Encapsulate camera and primitives from main application logic (Points 5)

1. Fork the current repository
2. Study the new framework-code of
   • IShader.h, ShaderFlat.h, ShaderEyeLight.h, ShaderPhong.h
   • ILight.h, LightOmni.h
   • Scene.h and main.cpp
3. A pointer std::shared_ptr<const IPrim> hit is now contained in your Ray structure. After a ray has been successfully intersected with a primitive, store the primitive’s address in hit (if the hit ditance is smaller than ray.t).
4. In the class CScene you find a method void add(const ptr_camera_t pCamera). Change your code accordingly using the appropriate vector defined in the class. See also method CScene::getActiveCamera() to understand how the member-variable m_activeCamerashould be also initialized.
5. In the class CScene you find a method void add(const ptr_prim_t pPrim). Change your code accordingly using the appropriate vector defined in the class.
6. Rather then intersecting each primitive in the main function we will now use the bool intersect(Ray& ray) const method of the Scene class. After modification the method should iterate over all primitives, intersect them and return true or false depending on if we had a valid hit with the scene data or not.
7. The loop of main.cpp calls the CScene::RayTrace(Ray& ray) method. This method should call bool intersect(Ray& ray) and depending on a hit or not return a white or black color.

If you did everything correct the rendered image should look like:

Problem 2

The Surface-Shader Concept (Points 10 + 10)

A surface-shader is a small program that is assigned to a primitive and is responsible for computing the color of each ray hitting this primitive. For example, a flat shader might just return a constant color for a primitive, whereas another shader might compute more complex effects such as lighting, shadows, or texturing.

In this exercise, you will add some missing parts of the given basic shader framework to your ray tracer and implement two simple shaders:

1. Implement a simple flat shader. Proceed as follows:
   • The shader class has a pure virtual function Vec3f IShader::shade(const Ray& ray), which has to be implemented in all derived shaders.
   • Implement the CShaderFlat::shade(const Ray& ray) method. The method should just return the color passed in the constructor of CShaderFlat.
   • Each primitive has a pointer ptr_shader_t m_pShader, which you can obtain via IPrim::getShader() function in the new framework and a corresponding modified Constructor definition. Adjust the Constructor code appropriate. For example, our red sphere could be initialized using CSphere s1(std::make_shared<CShaderFlat>(RGB(1, 0, 0)), Vec3f(-2.0f, 1.7f, 0), 2);. As we will see later some shaders need access to the scene data (e.g for light or shadow calculations), this is why these shaders gets a reference to the scene objects (e.g. CShaderPhong).
   • Finally, if, for instance, the primitive intersected by a ray has been stored in std::shared_ptr<const IPrim> hit (you can use ray.hit = shared_from_this(); in intersection algorithms), the appropriate color can then be computed by calling hit->getShader()->shade(ray); Change your code in CScene::RayTrace(Ray& ray) such that not black or white is returned but the color from the primitive with the closest hit or the background color if a ray does not hit a primitive.
2. Implement the CShaderEyelight::shade(const Ray& ray) method in the eye light shader, which uses the angle between the incoming ray and the surface normal at the hit point to achieve a better impression of the actual primitive’s shape. The resulting color should be calculated according to: result = |cos(theta)|·color where theta is the angle between the primitive surface normal and the ray direction. As the shader now needs to know some information about the primitive (i.e. the surface normal), some modifications are necessary:
   • Implement the Vec3f CPrim::getNormal(const Ray& ray) method in all classes derived from IPrim. getNormal(const Ray& ray) should return the normalized normal of the primitive. The ray parameter passed to getNormal(const Ray& ray) should be a ray that has been successfully intersected before, so you can assume that the intersection stored in this ray corresponds to the actual primitive. For example, ray.org + ray.t * ray.dir should be a point on the primitive.
   • Implement the shading function CShaderEyelight::shade(const Ray& ray) using ray.hit->getNormal(ray) to retrieve the surface normal of the primitive. With the surface normal the above given formula can be applied. If the test scene specified in main.cpp is rendered with these two shaders it should look like:

Problem 3

Phong Shading and Point Light sources (Points 30)

In the last exercise we implemented two simple surface shaders, which do not take light sources into account. A more advanced surface shading concept, the phong shading model, utilizes light sources to increase the rendering realism and give objects a plastic like appearance. Before we can implement the CShaderPhong::shade(const Ray& ray) method in ShaderPhong.h we have to implement a simple light source.

1. Implement a point light. Proceed as follows:

   • Study the base interface class ILight. Each light source which we will derive from it has to implement an ILight::illuminate(Ray& ray) method.
   • Implement the CScene::add(const ptr_light_t pLight) method.
   • Implement the CLightOmni::illuminate(Ray& ray) method. The method should calculate the light intensity, as described in the lecture, which hits the surface point from the light source as well as the direction vector from the surface point to the light source. The direction vector will be later used for shadow computations.

2. Implement the phong illumination model

   • The value L_r returned by Vec3f CShaderPhong::Shade(const Ray& ray) const should be calculated according to:

   L_r = k_ac_aL_a + k_dc_d Σ_l=0^n-1 L_l(I_l·N)+ k_sc_s Σ_l=0^n-1 L_l(I_l·R)^k_e

   c_a: Ambient color
   c_d: Diffuse color
   c_s: Specular color (Use c_s = (1, 1, 1) for white highlights)

   k_a: Ambient coefficient
   k_d: Diffuse coefficient
   k_s: Specular coefficient
   k_e: Exponent (shine parameter)

   L_a: Ambient radiance
   L_l: Radiance arriving from light source l

   I_l: Direction to light source l
   N: Shading normal
   R: Reflected incident ray direction (must point away from the surface)

   n: Number of lights sources

Notes:

• Sometimes an incident ray may hit the backside of a surface (i.e. the shading normal points to the other side.) Then, just turn the shading normal around to face forward.
• Only consider light sources that illuminate the primitive from its front-side (i.e. I_l·N > 0).

Problem 4

Shadows (Points 20)

To add more realism to the phong model we want now to incorporate shadows into it. Proceed as follows:

• Implement the method CScene::occluded(Ray& ray) in the CScene class, which should check if something blocks the light.
• Modify CShaderPhong::shade(const Ray& ray) to check for occlusion. If everything is implemented correct your images should look like this:

Problem 5

Reflection | Recursive Raytracing (Points 25)

Now we will implement a perfect mirror material with the CShaderMirror shader. Perfect mirror has no own color and reflects all the rays, thus in constructor it takes only a reference to the scene. Please implement the CShaderMirror::shade(const Ray& ray) method. To do so you need to calculate two verctors:

• N: Shading normal
• R: Reflected incident ray direction (must point away from the surface)

and use scene object to trace recursively the reflected ray with CScene::rayTrace(Ray& ray)method.

Once you implement everything correct the rendered image should look like:

Submission

Please submit the assignment by making a pull request. Important : Please make sure that

• No extra files are submitted (except those, which were mentioned in the assignment)
• The changes were made only in those files where you were asked to write your code
• The Continiouse Integration system (appVeyor) can build the submitted code
• The rendered images are also submitted in the folder "renders"