Scifi energy sword

Rendered in UE5 at 4k

Breakdown

References

This was a collab project with the amazing Lucas Steavens (full project here on his Artstation), a fellow student in the Concept Art Dept at CG Spectrum. Lucas was looking for someone to collab with for a more ambitious project and I decided to contact him and from there we hit it off. I'm big into scifi hard surface designs and personally that was my only requirement. When he started working on it we talked a lot about how to make the design more impactful and also make a functional design at the end. During this process he proposed a lot of different variants of the whole sword, fusion core and even sheathe design for me to choose and also give my own input on how we can go further on the particular design selected and at the end of this back and forth we came up with these two swords and 3 sheaths designs and as seen on my renders I went for the 1st sheathe.

So about the design of the sword the idea was that the swords belong to a gladiator who narrowly survived his first fight with nothing but a shovel he found lying around. Upon becoming a champion gladiator he requested duel blades, specifying that the old shovel be included in his new weapons. The fusion core hosts a bunch of nanobots which moves through the shovel and to the blade for self healing capabilities which keep the sword at 100% durability at all times. Gameplay wise this isn't ideal but the nanobots can be a form of resource or even "mana" to repair your sword.

As you've probably noticed I only have one sword (main) out of the two swords (off-hand and main). The idea was to do both and render both in the same scene but due to time constraints, at 2-3 weeks left before the deadline of the Rookies, I realized I couldn't do the two at the same time to their full potential and just focused on the main weapon instead.

Blockout

I want to preface that everything was modeled and UVed in Blender. I started with a blockout to establish the overall look, shape and size then slowly but surely started adding details and then got to the final result. To speed up my workflow in Blender I use dedicated hard surface add ons like boxcutter, mesh machine and also just some quality of life improvement add ons like hard ops

Mid Poly, UVs and Baking

At first I wanted to go with the normal high to low poly workflow and I actually did both but upon doing some test bakes in Substance I realized that the pipes on the high poly (that had subdivision on) and the low poly pipes (with no subdivision) caused some baking issues because of some shape difference (see image below)

High Poly on the left, low poly on the right

When I realized this I pondered for a bit because I did not want to add subdivision on my low poly (I'm already at 231k with the sheathe on my model) and I came to the conclusion that I don't really need this tiny detail to be curved as I wanted at first and just ditched the subdivision on the low poly. In any case a player won't really come this close to the asset and see this so it's no big deal.

For the rest of the model pieces I used the bevel modifier blender has with which you can change the bevel segment and amount non destructively and going mid poly I used 2 bevel segments. I generally use 3 for the high poly and 1 will not give adequate results so 2 segments was the answer.

For the UVs I decided to go with a UDIM workflow to give the highest amount of UV space to most pieces to achieve extreme quality at the end. I divided my whole sword into 5 materials/shaders. Metal for all the metallic parts, Rubber for all the rubber parts, Emissive for every emissive parts on the sword (the blade, pipes etc), Sheathe for the sheathe and Clasp for the clasp of the sheathe and all these took 9 UV squares in total (as seen above)

Here we have all the parts with the Metal material/shader applied onto them and each UV square is either a piece or numerous pieces unwrapped together to achieve maximum texture quality at the end.

I also tried to keep the texel density (TD) of each part relatively close to each other and my TD ranges from 83.31px/cm to 181.57px/cm which is quite a lot different. Almost a 100px/cm difference and honestly (and realistically too) I could have joined more pieces and unwrapped them together but this was for showcase purposes.

Sword has been modeled and unwrapped and it's time to export it to substance for baking.  But actually before unwrapping I ensured I applied all my modifiers in Blender (blender offers a non-destructive workflow through the use of modifiers). I kind of discovered it the hard way with this project that I need to apply modifiers (especially the bevel modifier) before placing seems and unwrapping because if I place the seams, unwrap and then apply the bevel modifier, the seams shift and it's no more where I wanted to be exactly and now the UVs are not right and thus causes some bake issues in Substance.

Black spots on mesh after baking due to unwanted seam placement caused by bevel modifier, especially on the AO map

After this little intermission, I triangulated my whole mesh before exporting it as fbx and then bring it into Substance Painter and baked it with the following settings:

After I finished texturing I exported the textures with the Unreal Engine 4 packed preset in Substance Painter. From there I created a new Unreal project (I was using the latest version of Unreal Engine 5 which was 5.4.1), imported my sword and my sheathe as two separate fbx files, imported my textures as virtual textures and started playing with it.

Here are all the materials I used for the sword. I learned about master materials and material instances and noticed it would be perfect for my project so I created a material instance for each material that had an emissive part associated with it. Since I wanted to control the color and emissive strength, I created a simple node setup linked with the emissive map (the highlighted node on the second image) of each master material with an emission map and parametrize it so that in the material instance (last image), I have two parameters I can play it and it updates on the fly! It's actually nothing too complex but I was proud of myself of doing it and see the results live!

Lighting, Rendering and Niagara Effects

The main aim of this project was for me to learn the basics of Unreal Engine 5 and implement it in my workfkow. I can say it was a huge success and I achieved even more than I had planned to do with it. Stuff like lumen, nanite and the fancy niagara to achieve some cool effects. Since it was my very first time opening Unreal I got help from tutorials and the very best that helped me with setting up my scene and understanding lights in Unreal was this tutorial by William Faucher. It covers setting up some settings for lumen and global illumination, how lights work in Unreal, baked vs movable lights and much more. Armed with this new knowledge I setup my swords and rect light. For the camera I used a 16:9 DSLR because I heard it provides the best quality, a focal length ranging from 85mm to 200mm (but more like I used either 85mm, 105mm or 200mm camera depending on the shot) and the following render settings:

Concerning the niagara lightning effect used for the blade I asked helped from my peers at CG Spectrum and someone was kind enough to lend me a hand and did a fantastic job with it. I still can't wrap my head around it as it was my first time seeing it in action but at least once it was setup I got to play with the parameters and understood a bit more about the system in general. Definitely excited to learn more about it.

As for nanite I didn't know it was that easy to set up. Right click on the fbx in the content drawer, Nanite, and then tick the box to turn it into a nanite mesh. For this project I actually didn't need to use nanite but just for the fun of it I did. With my sword being at 231k which is quite a lot (those damn tubes), I don't know how useable it would be in an actual game as a hero prop but nanite is supposed to be the fix. Not many games has been released with Unreal Engine 5. The only to date I'm aware of is Senua's Saga: Hellblade II and it probably doesn't use all the crazy tech that has been released as of recent with Unreal 5.4. All this to say that we have yet to see a game title utilizing these tools to their full potential so a lot of questions still lie unanswered and these titles are probably only getting developed now and we will see them in the next 4-5 yrs. So it's very valuable to learn about this new tech and that's definitely what I'm planning to do.