Skip to main content

Featured

Learning To Work With Python Frameworks: FLASK vs DJANGO

                 Image:  https://www.bitcot.com/flask-vs-django/ Breakdown I had experience building with Python and other APIs, but I’d never worked with a web framework. I wanted to learn how a full-stack web application fits together, so I learned two of the most popular Python ones: Flask first, as the small and simple option, then Django, as the larger one with more built in. This post breaks down what I learned from each, and how they compare. Tools Python 3.14 VS Code Virtual Environments Flask Django The Use Cases For Each Flask Built for simple projects, with a small core and a minimal set of features. Easy to install and quick to get started with. Fast and lightweight, but not designed for high-workload projects. Built on a synchronous model, so each worker handles one request at a time. Handling several requests at once means running multiple workers or threads. Best for quick projects and prototyping, like getting a ...

Learning Systems Through Building As A Solo Developer - Part 1

Breakdown

This is a breakdown of the tech demo I’m building as a solo developer in Unreal Engine 5.8, which aims to push rendering quality with MegaLights and Lumen. On the character side, I captured a likeness with KeenTools, brought it into a MetaHuman with Mesh to MetaHuman, dressed it in parametric cloth, and tested facial capture from my webcam through Live Link Hub. On the rendering side, I’m running ray traced Lumen alongside MegaLights, which is the main reason I’m on 5.8. The gameplay side is where I’m learning the most. I prototype movement mechanics like pick ups, slow motion, super jumps and dashes, which led me into C++ and Blueprints, overlay states in the Animation Blueprint, integrating my character into the Game Animation Sample Project, and building a widget and interaction system for picking up items.

The Tools

  • Blender
  • Unreal Engine 5.8

MetaHuman Pipeline Experiment With KeenTools

I used KeenTools in Blender to capture a likeness from a face, then sent the model to Unreal Engine 5.8 and brought it in with Mesh to MetaHuman. From there, I created a generic MetaHuman and fed the imported KeenTools model into it, then started polishing the likeness with the MetaHuman sculpting tools. It was tricky, but the goal was never a 1-to-1 match.

Parametric Cloth

Parametric clothing is built to adapt to the body underneath it, so when the proportions change, the garment resizes to fit instead of being remodelled by hand. I had to learn how parametric cloth works, along with some cloth simulation, before getting Epic Games assets from Fab and adding them to my generic MetaHuman. I then started adjusting the body proportions with the clothes on, and the clothes updated accordingly. Clipping did show up later, once I integrated the character into the GASP system, where parts of the clothing clip and move out of sync with the animation when playing with the player controller. That’s a story for a later section.

Facial Capture

I exported an optimised MetaHuman. It’s still meant for a game, but the demo is meant to push rendering quality with MegaLights & Lumen. I recorded footage with my webcam and ingested it with Live Link Hub, then exported the animation data and used it as an animation asset. Because a MetaHuman is made of separate pieces, I played it back on the exported MetaHuman’s head skeletal mesh. This is still an exploratory phase, so the aim was to see the capture in action under certain scenarios and find out whether it would be viable for a cinematic in the tech demo. The result was better than I expected.


Megalights with Lumen

MegaLights and Lumen solve different problems: MegaLights handles direct lighting, and Lumen handles indirect lighting. MegaLights lets me place a large number of dynamic, shadowed lights, because its cost depends on screen resolution and hardware rather than on how many lights are in the scene. Lumen then handles the other half, global illumination and reflections, which is how that light bounces around and fills the shadows. I enabled hardware ray tracing and running Lumen in its ray traced mode alongside MegaLights. MegaLights is the main reason I’m using Unreal Engine 5.8, because of it being production-ready.


Basic Game Design

I’m unintentionally learning game design by building the demo as a solo developer. I don’t understand it at an academic level, so my approach is research through design: instead of studying the theory first, I learn by making things and seeing how they play. I started with movement mechanics, because they dictate how the game part of the demo is supposed to play, and built a pick up system, slow motion, super jumps and dashes. Each one began as a small prototype, a quick way of asking whether an idea works and what it wants to become. They all live in the same character, so the prototypes don’t stay separate for long. Testing one beside another shows which ideas support each other and which don’t, and that’s how loose ideas coalesce into a design.

Gameplay Programming

Gameplay programming has been my biggest learning curve. As someone who builds tools and plugins, I’m comfortable with less complex languages at a smaller scale, where results are predictable. Gameplay is a different challenge, because I had to get comfortable with C++ and Blueprints, and with managing every gameplay mechanic and how they intertwine. On top of that, I’m trying to keep my Blueprints legible, reduce technical debt and build things that scale, both in terms of being able to scale and how to do it. That feels like a mammoth task, and I’m still missing some of the knowledge, and the proper jargon, to do it well.

Game Animation Sample Project

I migrated the Game Animation Sample Project (GASP) into my own project, since it had all the goodies set up. I slotted my custom MetaHuman into the CMC (Character Movement Component) Blueprint. I had issues migrating, so I couldn’t use the widget’s automatic switch to swap out the sample’s existing Kellan MetaHuman. Instead, I created a child Blueprint of the CMC parent Blueprint and added my custom character components to it manually. The components were then out of sync during gameplay and animation, which caused the clipping I mentioned earlier. I fixed it by binding the head, clothing and hair to the Body with a Set Leader Pose Component node, with the Body skeletal mesh component as the leader.


Animation Blueprints & Overlay States

I learned how to build on top of the existing Animation Blueprint by adding new states in the Anim Graph. I blended a pistol state and a rifle state with the original idle animation using Layered Blend per Bone and Blend Poses, and isolated the bones for each weapon so that only those parts of the animation are kept. Finally, I bound keys in the character Blueprint to switch between the idle, rifle and pistol states.

Widgets & Interaction Systems

I started with the UI, building a simple one with a Widget Blueprint, which is Unreal’s asset for interfaces. My widget prompts the player to pick up an item. For the interaction itself, I learned how a Blueprint Actor Component can be used with a Blueprint Interface. An Actor Component is a reusable bundle of logic and data that attaches to any Actor, with no transform or visuals of its own, which makes it a good home for behaviour like interaction that I only want to write once. A Blueprint Interface is a list of function names with no logic, where every Blueprint that implements it decides what those functions do, so I can call the functions in the Actor Component without needing to know, or cast to, the class of whatever I’m interacting with. I then placed the Actor Component in a Blueprint Actor which I created as the pick up item.

What I Learned

My biggest lesson is that research through design works for me. I don’t understand game design at an academic level, but prototyping the movement mechanics showed me how they fit together. Gameplay programming has been the steepest learning curve. As someone who builds tools and plugins, I’m used to less complex languages at a smaller scale, so keeping Blueprints legible, reducing technical debt and building things that scale is a challenge, and I’m still missing some of the jargon.

Conclusion

The demo isn’t finished. Facial capture isn’t implemented yet, and there’s plenty of work left on keeping my Blueprints clean and scalable. But the pipeline is working end to end: a custom MetaHuman, in cloth, moving through the Game Animation Sample Project with my own mechanics, lit with MegaLights and Lumen. The facial capture test was better than I expected, which makes it a promising candidate for a cinematic. Building it as a solo developer pushed me past the comfortable, predictable side of technical art and into game design and gameplay programming, which is where the learning has been.


Comments