Dead End: Stylized Environment Exploration Part 2

Production (Second Part)

- Texturing in Substance Painter

Substance Painter has largely enhanced asset texturing workflow. By combining smart materials and smart masks, quick iterations of base textures can be mocked up and then enhanced with clever use of the supporting tools.

When texturing, it’s important to consider the density of asset unwraps, also known as texel density. A rule taught is 1024 map = 1 meter of surface area so the candles in Figure 12 (see in the first part of the article) would most likely be 256, whereas the barrel would more likely be a 1024 map.

Baking is a more dynamic process too, including the option to utilize material IDs which enables us to mask smart material edits on particular assets of a mesh (nuts and bolts, wood, wires, etc).

Prior to baking, setting the low poly asset to 1 smoothing group eliminates any post-baking artifacts as demonstrated in Figures 13 and 14. As I understand, it is forcing the high poly normal maps to lead the vector information, but I could be wrong here.

Once baked, use a fill layer to add the curvature map and set it to Multiply with 50% opacity and the AO to Overlay with 50% opacity. This really makes the base material pop. It’s important to layer these above any additional base layers.

Substance Painter also has the ability to showcase opacity and emissive maps in real-time during editing. This came in very handy when producing the candles and lantern assets to ensure their believability. By adding masked fill layers and using the Paint tools, it was possible to develop a natural gradient of color whilst keeping true to the stylized art direction (Figure 15).

Quixel Megascans was another resource briefly utilized in this project to develop trees because the original sculpted tree wasn’t doing the scene justice. The process here was very straightforward. Find an appropriate dead branch to use modularly and create a conifer using alpha cards downloaded from the Megascans library. The problem which then followed was that alpha details were too high realistic as they were made with photogrammetry. Adding a surface blur in Photoshop and painting a thicker mask helped produce a stylized look.

- Substance Designer

Substance Designer is a perfect tool to produce tiling materials. The plan with the level was to create a courtyard and surrounding woodland and there needed to be some base for this. With that, 3 materials were developed: linear cobblestones, center cobblestones, and grassy mud. Not fully familiar with Substance Designer node development, I referred to the stylized material developer Alysson Soares also known as Kalyson and followed his YouTube resources. 

I’d say the tiling materials are one the weakest point in the project and going back, I should have really developed variants to height blend between each other. It would also have helped to use some POM decals for larger cracks or concaves. What did help was using tertiary assets, such as rocks and leaves, to break the ground up after the tiling materials had been applied.

- Level Design

With the majority of the assets and materials ready and in the engine, it was important to move away from the original map and begin working towards the larger-scaled design. The first iteration involved the overall layout and the addition of an exponential height fog for the misty atmosphere (Figure 18).

A key takeaway from Figures 18 and 19 is the introduction of color through lights and shape language to direct the focal point of the scene and the player's interaction with the environment. Originally, the scene was quite flat, so it was important to break this up. Vistas in the background created by the landscape tools added a world beyond the scene, while small chapels on the mounds helped break the foreground up. What I really wanted to push was this almost creepy feeling where the ground was alive, mutating, and eating everything up.

- World-Building and Lighting 2.0

Figure 20 shows the next stages of the world-building process. Here, there was abundant use of limited assets to create a lived-in world. The chapel in the distance has been altered to be a much grander-sized building promoting the concept of power. The forest has been thickened using the foliage tools to incorporate a messy and wild collage of graves, shrubs, grass, and trees. In the distance, there are pockets of areas of interest to spur the player to explore (those places could contain exciting loot or secrets).

Everything in the area has been purposely placed to ensure each corner had something visually interesting and appealing while fitting their own composition and story.

Figure 16 also demonstrates the understanding of using light and color to define a theme and mood as well as directing the player's attention. More prop lights are used to highlight points of interest, the looming red building in the distance support san evil presence and the bluish-green lights at the entrance of the chapels suggest ghost-like magic. Additionally, all these elements support the initial Halloween-based color palettes and use them in their most complementary forms.

Don’t be afraid to use plugins or pre-made assets, too. The scene, as said before, utilized some Megascans assets, pls Fluid Ninja for the fire, Smart Spline Generator for the vines, fog cards were taken from a UE4 starter tutorial scene and the sounds and music - from a variety of sources.

What’s important to understand about this process is that it is literally a case of going back and forth and iterating on everything. For this project, I always thought about the gameplay area. It’s good to have specific angles for your camera to focus on, but as a game/level artist, I want to make sure to answer the following questions: If players were to enter this place, what would they expect? What would they want? Is it exciting? Does it promote wonder and awe? And so on.

VR really enhanced this thought process for me. I’d simulate, place my headset on and bam! I’d be in, fully immersed, walking around and exploring. My imagination would run wild through endless possibilities and expectations. It was a very powerful development process.

Porting to VR

- The Process

Porting to Virtual Reality (VR) has never been easier. The first step in this process is to create a Virtual Reality base project. From there, it’s a case of migrating the original map to the content folder of the new VR project.

After adding a Motion Controller Pawn to the world, it’s important to ensure two things are done. The first is to disable the Auto Possess feature, otherwise, the function will spawn its own Player Spawn Mesh at the last location of the viewport camera, which is extremely disorientating in VR.

The second thing is to make sure that in the Player Height Setup within the BP of the Pawn, the output of Occulus HMD is actually named Occulus HMD. The bug related to this is an offset spawn which again is disorientating and frustrating to recalibrate.

The final step is to set up the playable boundaries. The VR system has a navigation method of playing. Add Nav Mesh Bounds Volume to the scene. This defines the playable area for the VR participant. What’s great about this is that you can use multiple volumes together to create a versatile and explorable level as seen in Figure 20.

Quick Tip: Press “P” to view the Nav Mesh Bounds.

Post-Processing

- VFX

When you’re not a VFX artist or weak in a particular area, it’s always an option to use tools and systems that will help build whatever it is you need.

For this project, an in-engine program called FluidNinja VFX Tools was used to create the stylized fire seen in Figure 21. This tool also allowed exploration of smoke, swarms, water, magical items, plus it has a built-in editor for quick asset generation. 

What’s useful about this tool is that a lot of the effects are baked into a 2D format and applied to planes making them extremely cost-efficient compared to complex particle systems.

- SFX

Adding sound brings another dimension to a simulated world as it pulls another sense into the experience which drives immersion. As this project was made for educational purposes, using public domain soundscapes was the most appropriate option. There’re many free resources and sounds available for download or purchase.

The most important consideration when creating SFX is 3D space and 360-degree sounds. This includes any panning sounds a person may experience within different areas of the level.

The sounds to be included were:

• Night-time creatures (bats/owls/crows/foxes)
• Spooky musical note
• Forest soundscape
• Demonic/ghostly sounds

What was quickly implemented in the scene was attenuation radius which allowed the sound to fade in and out depending on the player’s position relative to the core sound location. Layering these up in the world, as seen in Figure 22, really enhanced the experience and added a layer of believability to the final environment.

- Post-Process

Post-processing is ultimately where the final stages of magic occur. With little information regarding the specifics of this, the idea was to go straight in and make any changes that seemed visually compelling.

Some key changes were made to the overall color grading, the values of saturation, and white-to dark areas. I also changed the “Film settings” by altering the toe and slope.

The best advice I received for this process from a peer was to just experiment with different combinations because each project differs in its mood, theme, and emotion.

Feedback

- Strengths and Weaknesses

This project has been a huge challenge. Environment artists create believable worlds through creative workflow and technical understanding which means a balance between both sides of the brain.

The biggest strength of this project is its consistent research, the iterative and evolving project development and the overall collaboration of development skills and techniques. I have worked on lots of things I wasn’t comfortable with such as Substance Designer, ZBrush sculpting and better material development in Substance Painter. Additionally, I considered how I might be tasked with an environment at a studio and would have to work from various concepts, a blockout and/or a theme. Developing a full-scale area with an extremely limited pallet was also challenging. Yet, by doing all of this, I have come out as a much stronger and developed artist and will be able to enter the same development areas with a bit more confidence and the mindset to explore further.

In that sense, the entire project was a huge success and accomplishment.

What the project lacks is technical practice. Utilizing tessellation techniques and material height blends with vertex painting would have really aided the final level just like some POM Decals. Another area for improvement would be project management. Overall, the scene came out excellent but with a bit more consideration regarding the needed assets and their production, some of the things would have been built and optimized differently. For example, all the metal fences, gates, etc. could have been built, unwrapped onto the same UV, and textured at once. 

One more area for improvement is the chapels: using the same chapel again and again and again is repetitive. Utilizing decals, trim sheets, etc. for details and damage would’ve helped differentiate them. A broken chapel or a chapel the player could enter would’ve really broken the environment up nicely. This is something I’ll definitely consider for my next piece.

I also think it's really useful to understand the optimization process better. For me, the light complexity was overdrawn and I’ll be looking into its optimization in future projects. Foliage is on the list, too: lowering the complexity of the cards and using planes for single blades of grass was much more optimal than opacity masks.

- Final Thoughts

To conclude, I think this project has been a great success. With clear progression from start to finish and an overall demonstration of skills and techniques utilized in the modern game industry, it professionally showcases exactly what it originally was meant to do: create an immersive level for a player whilst following a strict art direction and the goal of producing a playable space.

Additional areas of improvement mentioned before would have pushed the project further and are definitely something that will be explored in the future, but I can’t complain about what I have achieved.

Thank you so much for reading! Feel free to add me on social media or drop a comment on ArtStation. It would be great to get your feedback.

Have a great day! 

If you missed the first part, read it here. It covers the theory, visual research, blockout, the first lighting pass, and sculpting.

Stephen Seress-Smith, 3D Environment Artist