Building a Creepy 3D Fantasy Environment Inspired by Greek Mythology
Helena Hansen took us through the creation of the Blighted Harbour project, an atmospheric fantasy scene based on the River Styx and made as her first Blender project.
Introduction
My name is Helena Hansen, and I have been doing 3D for just about 10 years now. I have been doing traditional art in the form of painting and drawing for practically my whole life before starting 3D, a switch which mainly came from curiosity about how game graphics are made.
I think curiosity is a great word to describe this project as a whole. I love creating environments and interesting spaces that feel believable and like they have a story to tell. A big part of my process is figuring out how to introduce movement and dynamics into my work, and how those elements can make a space feel more alive.
Taking it from a still image to a living, moving space is possibly my favorite part of the entire pipeline when I create environments like Blighted Harbour.
Reference Gathering and Inspiration
For Blighted Harbour, my initial inspiration came from the River Styx in Greek mythology. As the project developed, it evolved into a run-down harbour set in a creepy fantasy world, with a broken lighthouse in the very back of the scene. I've always had a bit of a soft spot for weird and creepy characters, and I wanted it to be reflected in this project.
It served as a great foundation for learning Unreal Engine's own rigging pipeline and creating something that perhaps felt a bit more dynamic, compared to my previous work. In general, it was the curiosity behind how control rigs are made and how they are animated inside Unreal Engine that ended up being the driving force behind the project. It allowed me to learn something new, along with using Blender and SpeedTree for the first time.
I typically don't work from a concept, as I enjoy working out the composition and focal point on my own. Instead, I more so create a moodboard of images that inspire the color scheme, atmosphere, and shape language, rather than translating any one image directly into 3D.
I will then break down the materials and assets and find more specific references for those, so I can see how damage, wear and tear, size, and construction really work, for example, with the boat in Blighted Harbour.
Blocking Out and Modeling
As I didn't use a concept for this project, I treat my blockout much more like a sketch. Just as I would draw rough lines on a piece of paper and refine them as I am drawing, I approach my blockout the same way — it's merely a method for me to get the idea I'm having in my head into Unreal Engine in a quick and often messy way.
What I keep an eye on here is mostly composition and the scale between all of the assets, rather than worrying about it looking refined. The goal is simply translating the idea in my head and from my moodboard into something a bit more tangible in an Unreal Engine scene.
I personally like getting a skybox and some rough lighting in early on to establish the mood and atmosphere, but as you can see in the progress image, this changes drastically along the way, which isn't unusual for the way that I like to work on my personal projects.
I used Blender for the first time on this project, which came with a bit of a learning curve. I've been using Maya since I first started learning 3D, so switching to a different modeling software naturally meant that creating some of the assets took me longer than usual. Luckily, I feel like the project served as a good place to start learning how to navigate new software, as I wanted it to have quite a few standout assets, but without it being overwhelming.
The main asset where I used Blender throughout the modeling pipeline was the boat. I started by blocking it out in Blender before taking it into ZBrush for sculpting, then brought it back into Blender for retopology and UVs. The boat was a great asset to start with as it pushed me to get comfortable with Blender while still incorporating tools I was already very familiar with, and it gave me enough knowledge in the software to tackle the rest of the assets.
After modeling, sculpting, and creating the low-poly versions of my assets, I would take them into Substance 3D Painter for texturing. This is where having a lot of references for the material itself, as well as how it wears and ages, becomes especially useful. A piece of advice I would give when texturing is to not be afraid of color-picking from your reference images, as it can be a useful way of checking whether your values and colors are in the right range of what you are trying to achieve.
I also like to check my textured assets against one another in a neutrally lit scene inside Unreal Engine. This helps ensure that the materials work well together in context, rather than only looking good individually.
Blighted Harbour contains a lot of wood, and I wanted to make sure that it didn't feel like every wooden asset had been cut from the exact same tree. At the same time, I didn't want all of the assets to sit within the same value range, as it can worsen the readability of the scene. I wanted the boat to be the centerpiece, so I gave it slightly brighter values to help it stand out from its surroundings while still feeling like it belongs within the same environment.
A bit of shader work can also help out a lot. As all of the houses and the docks are assembled from the same wood kit, the values were all the same. In order to create some visual interest and break up the scene, I created a gradient shader for those assets so that I could help unify everything and ground the houses properly in the environment. In the images below, you can see an exaggerated version where I have cranked up the values, and compared with the final renders, where the effect is much more subtle.
This project was the first time I used SpeedTree. In the past, I have assembled all my foliage inside my modeling software or used Houdini to help assemble or generate it, but SpeedTree saved me quite a bit of time once I got the hang of it. I still used Houdini for some simple cloth and rope simulations in the scene, but I am very pleased with how the foliage turned out despite using a new workflow.
For the character, I went through the same pipeline as with the boat, where I blocked it out in Blender and sculpted it in ZBrush. The rig and animation were done inside Unreal Engine, which was a fun challenge. The tools felt intuitive to use, and there are a lot of great materials and tutorials online that explain it much better than I possibly could, so I won't go into detail about it here — but I can highly recommend exploring the tools a bit if you are considering adding some movement to your environments too.
Lighting and Composition
Lighting and composition are some of the most enjoyable parts of any project to me. I will start out by mentioning that the skybox I used in the project is from Velarion on Fab, and the water is Easy Waterscape by William Faucher. I think both contributed well to the final quality of the scene, and obviously the skybox has an impact on the lighting itself.
I always start by figuring out the skylight in my scene. It serves as a good lighting base and also helps convey a time of day and lay the foundation for the atmosphere. Only afterward will I start tweaking the directional light intensity, direction, and color to my liking. Utilizing both the skylight and the directional light, I try to ensure that my shadows aren't too dark and that nothing is overexposed. In the end, I place smaller lights to accompany the emissives, or what I call artificial lights that are just there for highlights, such as on the fish.
Doing the lighting is also essentially where I think the whole environment comes together and where I can really tell if my work has any glaring composition issues. I very deliberately ensure that the brightest points of the image are where I want the viewer to look, and if my values are off in texturing, the focal point isn't going to be supported in the way that I need it to.
I usually go in and check my values here by making what I intend to be my main render black and white, in order to make sure that I don't have any unintentional dark areas or overexposed spots. I always try to be incredibly intentional with my color palette when it comes to my work and its composition. If anything needs major rework, this is usually also where it will get difficult to fulfill the vision I'm trying to achieve if anything is off.
The color palette for Blighted Harbour is very muted and earthy with a blue tint to it, whereas the limited saturated pops of orange are placed along the depth of the image to drive the viewer's eyes towards the broken lighthouse in the back.
This is also where I ensure that my composition looks good and check whether there is anything I need to change in order to support my intentions. I always try to split my renderings up into a clear foreground, middle ground, and background already from the blockout, and I want the lighting and volumetrics to support this. In the image below, you can see how I split it up with the asset placement.
The green here indicates the foreground, blue the middle ground, and the lighter green the background, whereas the red is the focal point. I personally prefer to do all of my image processing inside Unreal Engine, so there are no additional touch-ups done outside of the engine.
I usually render out an image and put it into my photo editing software and then copy the values into the post-processing inside Unreal Engine to replicate the changes I want. This way I know that my videos and images will look consistent, and all I have to do is add my logo.
Conclusion
This project took me approximately 2 months to finish, not counting roughly a month where I didn't touch it. My main challenges with the project definitely came from learning new software and tools at the same time. The composition was also a bit tricky for me at first, as the background is much more of an open horizon than I am used to.
I managed to learn some new skills which will help my work going forward, and I look forward to refining those skills more. I am happy that I chose to do an environment that is fairly repetitive and doesn't require a whole bunch of assets, as that would have taken a lot longer in new software and might have been quite overwhelming. A huge part of doing art for me is finding a balance between motivation and satisfying my curiosity for new workflows, but ultimately, I think curiosity as a driving force in our personal work is only going to allow us to grow.
My advice to new artists is to find and chase that curiosity behind how things work, whether it's just wondering how a gradient shader is made, or how a small looping animation is done in Unreal Engine. Sometimes we can surprise ourselves with what new skills we can pick up from just being curious about how something works and wanting to try it out.