Sculpting Nature-Inspired Bas-Reliefs on a Stone Wall in a Moody Ancient Forest Environment

Introduction

My name is Ilya Oskanyan, and I'm a Concept Artist currently working in the film industry. My main focus is environment and architectural concept art – designing spaces and building environments. Most of the projects I'm currently working on are under NDA, so unfortunately I can't say much.

My first job in the industry was in game development, where I worked on Steel Hunters as a Junior Concept Artist.

I heard about concept art as a profession in 2020 after reading an article about Jama Jurabaev. Despite my long-standing love for films, I had honestly never thought much about someone designing these worlds before they appeared on screen. Discovering concept art was a real turning point for me.

Before moving into concept art, I worked in live event production as a designer, focusing on set and stage design. At that point, I simply wanted to improve my skills, but the more I explored visual development and concept art, the more I realized that this was the direction I wanted to pursue professionally.

One of my first major learning experiences was a course by Mark Kolobaev, where I got a broad introduction to different software and image-making workflows. Another important learning experience was The Art Nest. The course was focused on building a strong, targeted portfolio, but it also gave me a lot in terms of storytelling and drawing. 

Getting Started with the Heart of the Forest Project

The simple idea behind Heart of the Forest had been in my head for quite a long time. I wanted to create some kind of gate or portal, although I didn't have a specific story behind it at first. I imagined it as something similar to an entrance leading to a boss arena in a Souls-like game, a mysterious and monumental structure hidden somewhere.

One thing I definitely wanted to design was bas-relief. I've always been fascinated by sculptural reliefs, so incorporating them into the architecture was one of the main ideas from the beginning. During the early stages, I explored several different directions, including battle scenes and horse statues placed on pedestals, before gradually moving toward the final design.

My approach to references is fairly straightforward. I usually start with a broad PureRef board and throw in almost everything I find that could potentially relate to the idea. Once the concept becomes clearer, I clean up the board and remove references that no longer feel relevant. The board continues to evolve throughout production as I add new references whenever I need to figure out something visually.

I generally try to rely on real photography and film stills, especially for lighting, atmosphere, materials, and natural environments. However, sometimes I add other artists' artworks when I find something useful for my project.

Blockout

Once the main idea was established, I started exploring it through simple blockouts and rough sketches. For the first 3D version, I kept everything extremely simple: two walls, a large arch, and a staircase. At this stage, I only wanted to establish the main shapes and overall spatial structure.

When looking for the right composition, I usually create a lot of cameras, ten to fifteen angles and lenses. I usually experiment with a pretty wide range of focal lengths, from around 20mm to 200mm, depending on the scene. For this particular image, I mainly explored those between 20mm and 50mm and eventually settled on 24mm. A longer lens would have compressed the scene too much and made it feel less monumental.

While doing this, I always keep a close eye on the horizon line, the rule of thirds, and the overall balance of the image. I start thinking about these even at the blockout stage. Sometimes I divide the image into two halves. If I can remove one half and nothing changes, the composition is bad.

My main principle is to work from general parts to specific ones. I start with large shapes and focus on silhouettes, scale, and the relationship between the main elements. Once the composition starts working, I add secondary forms and eventually smaller details.

The workflow constantly moves between 3D and 2D. Once I've found the perfect camera angles and a general lighting direction, I take screenshots from Blender into Photoshop and sketch over them. This allows me to quickly test ideas before spending time building them properly in 3D.

Building the Bas-Reliefs

After the sketches and early exploration, creating the bas-reliefs became the most challenging part of the project. I had to figure out the workflow while working on it.

I built all four compositions in ZBrush, including the trees, sculpted tree canopies, stars, backgrounds, animals, and characters. I then used the Bas Relief function to create height maps for each composition.

Whenever possible, I used the existing assets as a starting point – it's how I approached most of the animals. I also used a generic character and sent it to Mixamo for rigging and posing.

The friezes were built as geometry because I wanted them to feel heavier and more architectural. I modeled the frames, decorative bands, and other larger elements, while some of the base ornament came from Textures.com. The ornament on the pedestals was created separately.

The upper parts of the diptych were created using height maps with the Displace modifier in Blender. For the lower reliefs, I used height maps to create separate geometry, which allowed me to move and arrange all the different elements individually.

Modeling & Assets

I try to avoid building everything from scratch. If I can find a good base mesh, scan, or existing asset that can be adapted to the project, I usually start with that.

The deer is a good example. I found a suitable base model on Sketchfab and adapted it. I modified the proportions and silhouette to make it feel more powerful, adjusted the angle of the head to break the original symmetry, and sculpted additional fur in ZBrush.

The antlers were one of the most important design elements. They were based on the same pair, which I reused with small modifications and changes in scale. I also used assets from Quixel, including roots and pinecones, together with custom-made elements for the small totems attached to the antlers.

I'm not really a character artist, but I try to include characters when they can impact the storytelling and mood. It's not just a journey of the stick salesman.

The bas-reliefs tell a story about the union between humans and nature. The main character is a huntress arriving at a hidden sanctuary, where she meets the spirit of the forest, who grants her a spirit animal. The huntress and the statues were essentially based on the same character model in different poses. For the character workflow, I mainly used Character Creator and Marvelous Designer. I started with a base mesh in an A-pose, created the outfit in Marvelous Designer, and used posed versions from Character Creator as morph targets.

I already had some clothing assets, so I didn't need to build everything from scratch. After that, I brought the assets into ZBrush for additional adjustments.

For the statues, I remeshed and simplified the character, removing smaller details and making the surface much rougher. They were meant to feel like stone interpretations of the original character rather than detailed living figures.

Overall, one of the biggest time-saving principles was simply choosing the right starting point for every asset and combining the existing material with custom work where it was actually needed.

Unwrapping

I'm still a Concept Artist, so I try not to overthink topology. I still like to keep my meshes reasonably clean. They don't need to be perfectly optimized or production-ready, but I prefer working with geometry that is manageable.

My UV workflow is also fairly simple. Most of the time, I unwrap assets directly in Blender using projection methods such as Cube Projection or Project From View.

The goal is simply to get a functional result that allows me to apply textures and masks where needed without spending too much time on technical preparation.

Overall, I try to keep this part of the workflow as simple as possible and spend more time on design, composition, and storytelling.

Texturing

My texturing workflow is quite simple. I don't really enjoy building overcomplicated node setups, so I try to keep my materials straightforward. Most of the time, I use a standard Principled BSDF shader with albedo, roughness, normal, and sometimes bump maps.

My main sources of textures are Quixel and Textures.com. I usually treat them as a starting point rather than a finished solution and adapt them depending on the needs of the scene.

There were no particularly complex materials in Heart of the Forest. Most of the scene was built using simple shaders.

The bas-reliefs are a good example. I used a basic stone material built around a Principled BSDF shader and standard texture maps. I also used a mask to emphasize and slightly brighten the edges and facets of the relief.

The sculpted details themselves did a lot of the work. Since the bas-reliefs already contained a lot of information in their geometry, the material didn't need to create that complexity artificially.

I also often use cavity and edge masks extracted from ZBrush or generated with the Mask Tools add-on in Blender to emphasize specific areas of my geometry.

For me, believable materials don't necessarily come from having the most complicated node setup. Since I paint over in the final stage, I don't need to solve every surface detail perfectly in 3D. The goal is to create a convincing foundation for the render and then push whatever needs more attention in 2D.

Forest & Vegetation

The forest was just as important to me as the wall with the reliefs. Getting the right feeling of the forest was one of the key aspects of the project.

My first version didn't work. It didn't feel like the forest I wanted. It needed to feel ancient, powerful, and enormous while also framing the scene and guiding the viewer's eye toward the sanctuary.

I was even ready to sacrifice part of the relief wall if necessary to make the forest work better. The overall composition and feeling of the scene were more important than showing every element I had created.

One of the main principles at this stage was value grouping. I organized vegetation around the larger shapes: mainly the massive tree trunks and the path leading toward the entrance.

I started with a simple plane and sculpted it to create more natural terrain. At the same time, I shaped the path into a subtle S-curve, helping to lead the character and the viewer through the image. I also made a quick value sketch in Photoshop to establish the placement of the major tree trunks and overall framing before returning to Blender and rebuilding that composition in 3D.

I spent quite a lot of time on this stage because the composition was already close to final. The trees – their size, positioning, and silhouette – all had to support the composition and the mood.

For the vegetation itself, Geo-Scatter was extremely helpful. I used it to create around 80% of the smaller vegetation and ground cover. The larger and more noticeable elements, such as the ferns, were placed manually. I also added fallen tree trunks and stumps, which helped create the feeling of a very old forest.

The goal was to build the right hierarchy of shapes – big/medium/small – and depth: large tree trunks creating the frame, vegetation supporting those shapes, and the path leading toward the sanctuary.

Lighting

Lighting was one of the most important parts of the project for me. I think light is one of the main tools for both storytelling and realism.

The main illumination came from two directional lights. The first one had a temperature of around 6500K, while the second was warmer, around 4500K. They were placed at slightly different angles to strengthen the separation between illuminated and shadowed areas.

I also used the contrast between warm and cool light to reinforce the composition and guide the viewer's eye through the scene.

Besides the two main lights, I used around ten additional light sources throughout the environment. Some helped bring out the depth of the bas-reliefs, while others created rim light around the statues or added separation between different elements.

The paint-over stage is also a large part of the polishing process in my work.

I use several passes from Blender: beauty, AO, glossy, and always a mist pass.

1. Depth and Atmosphere

I usually start by working on depth and atmosphere. The mist pass is especially useful for controlling how different parts of the scene fade into distance and establishing a stronger sense of depth.

2. Textures and Photobashing

The next stage is focused on textures. In this particular project, I photobashed the tree trunks and ivy, and also used the Mixer brush for additional refinement.

3. Lighting

After that, I continue working on the lighting. I add extra highlights, strengthen some light areas, and adjust the overall contrast.

4. Final Realism Pass

By this stage, I usually merge most of the previous layers and continue working on a single image. This is where I add more color variation and work on smaller details in focal areas. In this project, that included small chips, cracks, imperfections, and irregularities. One of the goals of the 2D process is to make the overall image feel less like a 3D render.

Finally, I add adjustment and effect layers, including additional lighting masks, High Pass, overall noise, a small amount of blur, sharpening, chromatic aberration, and effects or particles depending on what the image needs.

You can make similar adjustments in Blender's compositor, but I prefer doing it in Photoshop because I can paint directly over the image while making all the other adjustments. I like to think of my workflow as a 70/30 rule – roughly 70% 3D work and 30% 2D work. 

Takeaways

The project took me around three weeks to complete while also working a full-time job.

Taking on new tasks is always a challenge and a step outside of your comfort zone. You can easily run into problems you weren't prepared for, especially when you're trying to solve something you haven't done before.

The bas-reliefs were a good example of that. I had to figure out the workflow as I went, and there is always a possibility that the result won't work the way you imagined. But I think that is also an important part of personal improvement: trying things you haven't done before and accepting that not every problem will have an obvious solution from the beginning.

For me, one of the biggest challenges in personal projects often happens before production even starts: choosing what to work on. I think it's important to choose subjects that are genuinely close to you. If you love fantasy and castles, it probably doesn't make much sense to force yourself to make mechs the main focus of your personal project.

Today, there are countless workflows, tools, and courses available to artists, and it's easy to become too focused on finding the perfect process. But I don't think the tools are the most important part. Creative process shouldn't be controlled by them.

Tools should help, not define what you're capable of creating. If a workflow starts limiting your ideas, it's probably time to rethink the workflow rather than the idea.

As Feng Zhu said, "Just draw!"

Ilya Oskanyan, Concept Artist

Interview conducted by Gloria Levine