Stylized Campfire Post Mortem

So I wanted to do something calmer, darker and a bit creepier after my last project. Something with muted colors, a dark atmosphere and hopefully set at night. Unlike the last project which ended up being a desperate scramble to the finish, this project was a rather relaxed rush to completion across every week, where I was dogged more by technical limitations rather than inexperience. Its due to this that instead of taking this day by day I’m going to break it down into the individual subsections.

So everything that’s made 100% out of wood in this project shares the same texture map. After the last project where I remembered on the second to last day that draw calls were a thing, with this project I went in ahead of time with every prop on my list broken up into small chunks to determine who is sharing their textures with who. The trees (which warrant their own sections) and the stumps and logs. This is where I ran into my first technical difficulty. I wanted to do a mix of Hand Painting and PBR which is fine, substance painter exists specifically for this. However for some reason, substance painter has started running horribly on my rig when attempting to do traditional layered painting. So I let these pieces determine how I would proceed in order to account for this limitation. I used wide swathes of similar colors to give them a vibrant, however still dead feeling. I was inspired by how cute animal crossing made all their wooden naturey furniture. The actual sculpting process involved a lot of reference which is what ended up taking these props from a 5 to a 10.

The tent itself is a different beast. In my mind, adventurers were taking these tents with them, so I opted to make them a different color to make them seem more foreign. I think the effect turned out rather nice. I went for a scaled pattern for the tops as what would be stronger to fashion a tent out of than dragon scales? I got a lot of mixed responses on that decision. My girlfriend personally didn’t enjoy the idea, but my buddies who actively play fantasy games like World of Warcraft were onboard with how cool it seemed. I suppose the unusual texture could be off-putting. I personally like the effect however this asset got a little rushed. I actually wanted to use the cloth effect to attach the scales to the grommets. However in practice this effect looked both confusing and underwhelming and my deadline for the prop was approaching, due to this I cut my losses and accepted on a pretty nice looking prop.

So I made most of these today actually. By the time these had come around the scene was all assembled, lit, and the tone established. So I knew these props in particular needed to stand out against the rather dim and gloomy lighting of the campfire. I opted to crank the roughness down to make sure they caught more of the orange and blue light from the fire and the sky respectively. They’re all operating on a 256x256 map, however they all have their own. That’s 4 extra draw calls (as all the chains share the same map). I was actually rather happy with the end product on all but the sword, I really kicked it into high gear and they came out rather okay.

In my opinion the sword looks very nice reflecting the moonlight.

So, welcome to my own personal nightmare. I’ve never made leaves before, I’ve actually avoided making trees so I wouldn’t have to make leaves. However you can’t grow if you don’t struggle and man, did I struggle. So we’ll talk about the grass, then we’ll get to the leaves.

Now then, let’s move onto the true struggle, leaves. First of all I didn’t even know how to place them accurately in a way that wouldn’t look super video gamey. That’s not to say that super basic trees aren’t nice or fun to look at, but I wanted to get that oppressively safe forest canopy. So how did I put this together?

I made a particle system in Blender, and prayed to the dickens that it worked. Basically I made a single leaf billboard, then I made it hair on some spheres I was using as placeholder leaves. From there I converted all the hair leaf particles to meshes, turned them all into single user objects, and then joined em all together, and the effect is pretty bad.

The normals were all screwed up as you can see above. But that’s not even them in their worst state, that’s them in their best. I got them to look presentable at specific angles for the sake of the deadline but I’m going to need to do more research into properly making game leaves. I used a method that I saw somebody do forever ago where they made the leaves all point their normals out towards the faces of a sphere but when your base mesh itself is subpar, there’s only so much that can do.

There’s just so much no that comes out of my eyes when I look at these. I’ll have to note this failure and seek to do better in my next projects with trees.

So the atmosphere of this piece was actually a fun accident when I was putting the lowpoly in in place of the graybox. I decided to reuse the undertale cave rocks for the fireplace since some of the rocks were under 100 polygons and wouldn’t hit my budget.
I wasn’t happy with the look the built in unreal atmopshere/fog could create so I decided “I’ve got time, I’ll write my own.” I spent a couple on the post process material that gave it this hazy, dark and blue feeling, which then established my pallet of Orange and Blue. They’re the default contrasting colors for video games so it kind of makes sense.

I then spent the next 3 days just playing around with the composition, the location of the rock faces, trees, what to frame, how to frame, etc. I think the overall flow is good but the colors were a bit lacking in the end. I had never lit a night scene and night in a lot of games I was referencing feels like it’s either a blue filter or pitch black darkness. I tried to meet in the middle and I can’t tell if it was a success or a failure. I’m leaning towards success as it gives everything this kind of moody vibe.

The light shafts were a particular issue as my material had this weird habit of showing up on the rim of the mesh for some odd reason. So I took the time to troubleshoot it and ended up just completely rewriting the material from scratch.

The fire in my opinion is another big flop, while it is fine within still images it animates in this extremely bizzare and unnatural way. This is more than likely due to my inexperience with Niagara, I have no false pretenses about my abilities with VFX, however it cannot hurt me to learn more about Unreal’s “new” experimental particle system. Initially to match the kind of angular style I have going, I had intended on using Vertex Displacement Animation from Houdini to create the fire. This was a massive point of folly. You cannot learn Houdini in two weeks. You cannot learn Houdini in a month. Houdini is like an Ostrich Egg, in that it takes a lot of deliberate and directed force to crack. I’ll be learning it in my spare time but trying to pick it up properly for this project was willfully ignorant at best.

Do not be afraid of bright lighting in a dark scene, the contrast will help direct the viewers attention.

I need to learn more about general VFX for scenes that involve fire/light shafts/particles as a whole.

Lighting can be Bright and Dark at the same time, and I need to use that contrast in these dimmer scenes to make the scene actually pop.

That warning Susbtance Designer gave me about my GPU requires immediate attention.

Rock sculpting can be a lot more destructive than I normally made it, and it can make some more interesting shapes because of it.

Going too crazy with the post process materials can break the camera. I had an ongoing issue where the camera had a blue bar routinely along its right hand side. To this day I do not know why.

Texture optimization, sizing, and UV packing all happened much faster than on my last project

The Props were all through the highpoly phase within two days, a dramatic increase compared to the last project. However I am still not fast enough for the industry, but my time is reducing. Matter of time and practice.

My hand painting, I feel improved through this project. It went very fast and was a lot of fun when Substance was working properly.

I reached out for critique from an artist I respect, got slapped upside the head on the brightness and iterated.

I iterated much faster on my ideas.

I managed to selfcrit what was and what wasn’t working.

I got through it without suffering any mental breaks.

I managed to work through life issues.

I tackled particle systems with a particle editor I am not familiar with.

I learned a bit of Houdini.

EDIT: I also need to use a mesh for scale next time. I did some reflecting and that’s something I was kicking myself over.

Related posts:

Taking a step back to unwind after a hefty working day is quite satisfying, especially if one greeted by Slatan Coffee Tea Chocolate humbling hospitality. Located in the well-known neighborhood…

Nowadays most organizations manage, analyze and leverage big data in any given industry. One of the most promising areas where big data analytics can be applied is healthcare. Healthcare analytics…

Taking a screenshot programmatically in an Android Application can be very easy. It will be very helpful for developers when user reports something and send a visual form of that screen when a bug…