Building a simple ray tracing project I followed the book Raytracing in a Weekend by Peter Shirley and here is what I've done to complete the project!!
The code generates a ppm file something like this P3 512 288 255 0 120 214 0 120 214 0 120 214 0 120 214
P3 says that colors are in ASCII format
the line 512 288 is the dimension of the 2d image 512px * 288px with aspect of 16:9
the 255 in the 3rd line indicates the limit of the colors
and after these initializers, we have the actual colors for each pixel
After creating ray struct (I preferred struct for most of my codes, it's memory friendly) here is the final image, a simple gradient from white to sky blue
I implemented a simple sphere, or a circle for now, and it's similar to Japan's flag
Now here was the fun part, I implemented the normal vector and normal map and it gave me a beautiful mixture of blue, purple and green, like in uv coordinates
After implementing normals, it was time to get into the actual work, by defining the hitable and hitable_list structs I could inherit from it for the objects that are hitable, like sphere! I also created another big sphere as ground
So after implementing all those stuff, it was time for creating diffuse color and light, but not a complex light, a simple one, for this I needed a randomly generated number between 0 and 1, unfortunately I couldn't find a built in cpp or c method for that, so I created mine, that clamps the value between 0 and 1 for demonstration I have two outputs, one with random generated number and one with no random generated number
I love this chapter, here I implemented a metallic material with fuzziness, the basic idea is that when rays coming from all directions hit a certain point on a surface, the randomness in reflecting them indicates the metallic attribute of an object!!
So here I did something wrong in my code, so some of the rays that have a certain attribute should be reflected from the surface and some of them should refract, well, here I refract the rays by mistake, and I got emission material!!! like a glowing sun or lamp
Okay, something like water or glass or even air do not reflect the light or refract like other materials, these objects have a formula n * sin(theta) = n' * sin(theta') which says that when a ray of light enters a dielectric space like water, it changes its direction, and there are some complicated stuff here but I achieved this after several attempts!
and here is the final result for this chapter!
Amazing right?
So the next chapter talked about camera, because before coming to this chapter the camera was handled in the main.cpp file, so here I separated the camera and put it in another struct
There is something strange right? why the images have a shear effect applied to them? A LITTLE STUPID MISTAKE WHICH TOOK ONE WEEK TO FIGURE OUT!
So I ignored the shear effect that applied to my output and said to myself: Yeah, this is fine!! So after this I implemented depth of field, in summary it blurs the parts of image we don't focus on (worst explanation of depth of field!)
A beautiful green image!!! . . . .
after adjusting the aperture value and camera position and rotation, I got this
Okay here we should spawn many spheres with random materials, see my beautiful image
yeah, the stupid mistake got me here, remember when I told you I ignored the stupid little mistake?
So after days of refactoring code and searching in my codes, I finally got another result, I call it a Nightmare!!
I was losing my sanity, after I realized that the Cross product in my vector class is causing this bug, so basically here is the formula for vector 3 cross product
C.x = A.yB.z - A.zB.y C.y = A.zB.x - A.xB.z C.z = A.xB.y - A.yB.x
so in first line, instead of what is written, I wrote: C.x = A.y*B.z - A.z*B.x
You didn't notice it either?? the last element should be B.y, not B.x!!!
So I fixed this and no more shearing!!
here is the final images and results with 3 spheres
