|
Note
|
This is windows only for now since I can’t be bothered testing on multiple platforms during development. This might change in the future, but for now, it is what it is. |
A C++ library built on top of the Vulkan API(might change) to render 3D vector graphics(and also other stuff) directly without vertices. It works by shooting a pixel ray from the camera for each pixel on the screen and then checking if the ray intersects with any object in the scene. If it does, it calculates the color of the pixel based on the object’s shader. This is a way to abstract the rendering process and still remain compatible with existing 3D models since you can just copy a function that checks for intersections.
The main premise is that each object on the screen has its own shader that should be able to return things like color, normal, and a integrated code like "vre_stop" or "vre_mult" to control the way they interact with other objects.
The reason I want this is that with vertices, you can never achieve a perfect sphere, and with SDFs, you can never use traditional 3D models (at least not without a lot of overhead). This tries to be in the spirit of SDFs but with the ability to use traditional 3D models.
The advantage of this might not be seen with solid objects, but if you want a lot of magic particles or mind-bending effects, this might be the way to go. While traditional 3D rendering is good for constant objects, this approach will probably be better for dynamic objects since it is defined via functions instead of values.
I am also planning on doing something similar to this as a physics engine and a 2D rendering engine since the same principles apply. If that happens, you could probably use the same shaders for all engines, which would obviously be a huge advantage.