I’ve been thinking about starting this blog for a while now. However, I’ve been stalling because I’m not sure which direction I want to take here. Do I want to blog for an audience, or just as a logbook for myself? Do I include explanations to make things easier to read? Do I write a big introduction post with all history up to this point, or just add pieces where necessary? I guess I want to please every possible audience, but is that even realistic? Do I stall on these questions for another week, or just start right now?
Let me take you on a journey where I experiment with bleeding edge technology in videogames. The screenshots in this post have been taken in-engine, using the graphics code I have been working on for the past 3 years.
My name is Ramon Stefano. I’ve been modding and creating games since I was a kid. I started with drawing on paper, switched to 3d modeling and texturing when I was 12 years old, and then started coding. I’ve studied game design & application development in college, worked full time in racing games & API development, and have been working on my passion project, a racing simulator, for many years.
I like film/cinematography and want to create immersive, magical experiences in videogames that make you forget about the real world.
To do this, I want to bring photorealistic graphics to videogames using real time path tracing. This road from the very first videogames to photorealism, with graphics and graphics cards improving every year, is a journey we’ll only get to experience once, and I want to be a part of it.
I like to experiment with bleeding edge technology. The kind of things that are still in active research, things that are not ready, not stable, and won’t appear in mainstream videogames for the next few years.
This summer, I’ve been preparing the technology for a little side project called Tiny Adventurers, a game about clay characters coming to life and having to face challenges to survive.
Here’s what I have in mind:
- Path traced graphics
- Ray traced audio
- Speech to text
- Text to speech
- LLM trained on personality, sensory and interaction data
To start off, to achieve path traced graphics, we’re going to need a somewhat modern graphics card.
I will be doing all the graphics programming on this RTX 3080.
The RTX 3080 has the necessary RT and CUDA cores on board which we need to do path tracing and LLM inference in real time.
We’re also going to need a VR headset. I’ve picked the Quest 1 over the Quest 2 because it’s supposed to have better colors. I prefer realistic colors over resolution.
And of course, papers, lectures and books. I got myself this nice hardcover of Ray Tracing Gems II which includes a lot of key ingredients towards building a real time path tracer.
I dropped everything maths related in school because I wanted to become a guitar teacher. Unfortunately, graphics is all about maths and statistics. Thanks to TU Wien & Utrecht University, I’ve been studying and watching lectures to get back up to speed.
I want the characters to have some kind of personality. The past few years I’ve been reading books, watching interviews and listening to podcasts to learn about personality disorders, but I don’t know a lot about personality types. So, I decided to pick up this book about personality types called “Surrounded By Idiots”, a dutch book for a change.
If you’ve come this far, thanks for reading. The next post will most likely be a technical deep dive into either graphics or character behavior.