This repository contains most of the projects that I've done in OpenGL.
I learned OpenGL during the summer of 2022. I tried multiple different tutorials at first, but I couldn't manage to stick with any of them. The one by Yan Chernikov was good to get started, especially in regards to using Visual Studio, but I find that he talks too much: The first quarter to a third of the videos is often just spent talking. However, his videos on C++ did help me understand the language better. And he did introduce me to ImGui.
Eventually, I followed the tutorials by Joey de Vries, which I have decided not to include in this repository. I have done the Getting started, Lighting, and Model Loading sections (though I didn't like the last section because I didn't understand it much), and stopped at Framebuffers in the Advanced OpenGL section.
One thing to know, is that I didn't really agree with the tutorials making classes for each components such as shaders and camera, but I did make a class that has a VAO, a VBO, and a shader program. It can also load textures and framebuffers.
Inspired by this video, I thought to use that as a gift for a relative, as I often do with these projects. I managed to make it work in a surprisingly short amount of time; setting up controls with ImGui took longer.
I originally did this project in October 2022 in Visual Studio, using vcpkg for the first time. I later remade it using CMake in July 2023, adding a stereoscopic version of it: I got interested in stereograms some years back (January 2020 I believe), and wanted to view the scene in 3D without glasses. After a day or two of testing (which I didn't include in this repository), it turned out that you merely needed to draw everything twice, once while the camera is translated to the left then once to the right (or the other way around), with the viewport taking half the screen.
You can right click to free the cursor, and left click to bring it back. You can also zoom in and out, though it's slower using a mouse.
This project uses vcpkg
I first did this project using OpenCV, but this time I did it in OpenGL in December 2022.
In V1, I use another framebuffer for the fading and blurring effects.
Explanation of the main loop:
First, the agents are drawn on the created framebuffer to use it as a texture for the default framebuffer, which applies the fading and blurring effects.
The results are sent back to the created framebuffer with glBlitFramebuffer(), then the pixels are read from the default one and stored into a dynamic array, which is used to update the agents' position.
The downside of this version is that, since the ImGui window part of the OpenGL context, its pixels also get transferred to the framebuffer texture, so it messes with the simulation.
While faster than its OpenCV counterpart, I wasn't that impressed by the simulation's speed. I decided to try my hand at compute shaders, since from what I had seen, they seemed to be exactly what was needed for this kind of simulation.
I found some tutorials, though the one that I think helped me most was this one.
If the starting position is the center or the corner(s), the program starts out at 60 fps, but as the agents take more of the screen space, it drops to around 10 (at least for me). No matter how I dispatch the workgroup, I couldn't make it go faster. Though I suppose that, if we compare the difference in the number of agents compare to V1, this one is better.
This project was done using Visual Studio, CMake, and vcpkg
I first tried doing Conway's Game of Life using WebGL (think of it as OpenGL but on the Web) because I wanted to be able to do stuff for relatives that are on MacOS.
Andrew Adamson's tutorials were an extremely great ressource to learn WebGL (or even the regular OpenGL).
Unfortunately, due to my lack of knowledge in JavaScript and asynchronous programming, I failed (I figured out months later that I only needed to use a function with RequestAnimationFrame() instead of a while loop).
I made the Cellular_automata folder to try and do, well... cellular automata. I had realized that the double framebuffer setup that I used for Slime_mold/V1 could be used for a lot of cellular automata and other simulations with simple rules, not just Conway's Game of Life (which I made here). Inspired once again by Sebastian Lague and one of his videos, I tried Reaction-Diffusion, but for the life of me, I couldn't manage it! Whether in be here, or using OpenCV (either with C++ or Python/Cython). I did not include my attempt in the repository.
GPU_Info just makes a text file with your GPU information on it.
This project was done using vcpkg
So far, I had only been doing OpenGL projects using C++, but I wanted to see if I could do one only in C.
Inspired by this video, I decided to do the same thing in openGL.
This is the also the first project that I compiled for MacOS. Something annoying that I had to figured out (and I couldn't find anything on the internet about), was that you had to double the viewport size on the Mac for some reason (something with the dpi maybe?).
Unfortunately, my Mac isn't running the project properly (maybe because of an update or age, I don't know), so I can't guarantee that it will work.
There are things that I didn't include in this repository:
- Tutorial: I already mentioned that one, it contains stuff that I made following Joey de Vries' tutorials. You can still see files that I used in the ressources folder.
- 2D_physics: I didn't actually do anything with this, except prepare a dummy OpenGL project. I instead used OpenCV, but the most I did was bounce a red ball on the screen. With multiple, I couldn't even make them bounce against one another. I quickly abandoned it since it was the start of the 2022-2023 school year.
- Particle_Life: after the Slime mold simulation and Conway's Game of Life, I wanted to try my hand at other systems with emergent behavior. I attempted to do Particle Life, both with the double framebuffer setup, and with a computer shader, but the results were underwhelming. I also tried Boids, or the simulation of flocking behavior, and while the result was better, it still wasn't that great.
- Stereoscopy: This was just a folder where I tested how to do stereoscopy. I was satisfied after I changed a program from the tutorial to be in 3D when crossing your eyes.
- Langton's ant: I've talked about this one in my OpenCV repository. I couldn't manage to do it here, only using Python and OpenCV.