Thomas Charier is Art Director at Happy Volcano, and an accomplished 3D artist specializing in animation and character art. Here, he discusses his recent collaboration with the Substance 3D team, creating a triptych of stylized lamp images.
Initially, I was contracted by the Substance team to create a stylized animation using the materials available on the Substance 3D Assets platform. The ultimate goals for this type of project are typically twofold: it creates content allowing the Substance team to showcase the level of quality that can be attained with their resources, and there’s also a ‘research’ component – the 3D Assets team gets to ‘stress-test’ their resources in a real-world use case, essentially to see how well their assets meet the needs of 3D artists, and to identify any potential areas of improvement. On my side, my work at Happy Volcano tends to involve a range of stylized projects, and I was eager to see how Painter’s pipeline and tools can help NPR (non-photorealistic) 3D artists. Substance 3D Painter is already well known as a great tool for photorealistic texturing, but what about for stylized textures? The collaboration seemed a good fit.
Concepting and Modeling
For this project, I ultimately chose to cover the overall 3D process of creating three animated lamps, from the design phase to the final 3D animation. For my own creative brief, I decided that each lamp scene must have a unique theme, and use its own specific color scheme.
I start with concepting. And, for me, concepting and sketching ideas usually represents 50% or more of the overall project. Only when all the artistic rules and details are established can I begin production. The more information I have, the better.
I think it’s fair to say that I have a very structured approach to creating artwork – I prefer to plan every single step of the process very meticulously from the start. And I document every step of the creating process as I work. This allows me to see at a glance the work carried out so far, and what still remains to be done. And yet, for this particular project, this structured approach would not prove ideal. This project would force me out of my comfort zone, and cause me to reevaluate and challenge my skills and habits as a 3D artist.
For this project, I began by finding references lamps from the Art Nouveau style, freestanding lamps as well as wall fixtures. I had previously worked on such a lamp as a sculpting experiment, and I circled back around to that concept now. That older sculpture did require a rework – recycling old ideas and projects is always a good design experiment, I find; taste evolves over time, and design flaws reveal themselves sooner or later.
The Art Nouveau style can be very detailed, which doesn’t necessarily leave a lot of space for the texturing process. And so one of the challenging points during the modeling phase was to create a simple version of this style. To do this, I tweaked the sculpt in Blender, added thickness to the mesh, and also created a custom environment to support the prop.
I hit upon the idea of creating a color triptych, three variations of the same scene that would remain within the same overall color palette. The back wall section behind the lamp, and the waterfall bottom plate, would allow me to create three unique scenes. And so I had to keep the environment really simple; this would allow me to transform the shapes into something really different for each theme. I also kept very firmly in mind that the final stage of this project would be to animate it. And so I made sure that the lamp had an organic aspect about it – this, I find, makes the animation phase much more convenient.
I kept the final mesh quite low-poly, as I thought it might be interesting to use real-time renderers at the end of the process. Keeping the mesh low-poly would make that a lot faster and easier, overall.
Testing out Textures
To begin the texturing process, I typically start out by finding some good sources of inspiration, and a strong color palette. For this project I first checked the stylized material starter kit in the Substance 3D Assets library to see if any of the available materials could fit the scene. Some of the materials available seemed like an obvious fit – some of the metallic parts and liquid materials, for instance. But, of course, at this early stage every choice remains hypothetical.
Once I’d made a selection, it was time to start applying the materials. This is still part of the experimentation phase – the goal here is to test out scale, and values, so that the initial range of hypothetical possibilities can start to be trimmed down into something more definite. And here, for sure, it’s important to have a strong design document. Without this, the endless texturing possibilities can be quite overwhelming, and it’s easy to lose your way.
In my texturing process I always start with the surfacing work, such as applying normal map details. Each surface is different; each one can have varying degrees of grunge. Then I play around with colors and try to make the whole thing harmonious.
One of the big advantages with the Substance resources, I find, is that I can rapidly test some very complex materials, and I don’t really need to paint at all. It’s a great resource for design and iteration.
Working with Color
I rendered the first draft with a custom light rig in Cycles, and the result wasn’t really what I expected. I discovered that the rendering setup, such as the lights and the shading, can conflict with stylized textures. In this case, the shading and smooth lighting didn’t really help the textures.
Finding the best combination of colors for each scene, and to maintain the harmony within the triptych overall, became a real struggle. There were just too many possibilities. I ended up carrying out a lot of color experiments – really, a lot of experiments – to test things out. I think that this approach would have been easier if each scene within the triptych had had its own specific animation or custom lamp model. In my case, however, I’d determined for artistic reasons that I’d change only the textures from scene to scene. It wasn’t a great situation.
To help establish a color scheme, I was fortunate enough to be able to work with Anaïs Lamelliere, Color & Material Designer on the Substance team; Anaïs would work with me to push my triptych idea further towards the creation of a unique universe for each scene. I found it fascinating to discover the precise process behind the definition of a color scheme. Establishing harmony between soft and contrasting colors had been a mystery for me, until that point. With the help of Anaïs, we defined a new design document. And yet I felt quite submerged by all these new colors that I would typically never get to play around with.
A Creative Impasse
The project got stuck at this point. A part of this was because some real-life obligations came along that impacted how much time I could devote to it. But, largely, it was a creative block. I mentioned above that I tend to have a very structured approach to creating artwork – this is what allows me to work efficiently. But here, in this stylized lamp project, my familiar, proven work processes weren’t providing me with good results.
The project would ultimately move quite far away from the initial brief that I’d decided with the Substance team. And this is something that happens from time to time – the result of a creative project must fit the initial brief, but the brief must also fit the technical possibilities of the project. In this case, I remained in close contact with the Substance team throughout this blockage, and we discussed the best approach both in creative terms, and in terms of fulfilling the overall mission objective of showcasing the stylized materials available in the Substance 3D asset collection.
In retrospect, I think it was also an issue that there was quite a lot of repetition of materials from scene to scene within the triptych. There wasn’t a sufficiently satisfying degree of variation from one image to the next. Of course, I was working on this stage of the project in Q3 of 2021; the range of stylized materials on Substance 3D Assets was smaller then. Since then, more materials have been added; I think this wouldn’t be a problem, now.
I’d originally envisioned the project as requiring a straightforward, settled process; now, in an attempt to find a solution, I was being forced to dive into a complex, experimental approach. This impasse continued for a good couple of months.
The Stylized Lamp workflow
Modeling: Blender > Texturing: Substance 3D Assets / Substance 3D Painter > Test renders: Blender Cycles > Base color painting: Substance 3D Painter > Animation: Blender > Final renders: Blender Eevee.
At some point, I jettisoned much of the considerable planning that I’d carried out so far. This wasn’t at all easy to do – 3D creation is quite flexible if you manage your pipeline properly, but making changes to the core concept of a 3D project has a huge impact on the following steps, and on the overall creative freshness. The deeper you get into a project, the less freedom you have to take a few steps backwards.
My main issue was discarding my earlier triptych color ideas and thinking about something completely different and unique. It took me some time to refresh my brain as I’d been so focused on a specific idea for such a long time.
Every day I approached new ideas, with no precise artistic target, just a fuzzy idea of the direction in which I hoped things might move. It was a really tedious process, as none of the details were planned out.
When I’m stuck in a creative dead end, I usually push my comfort zone by trying more dangerous design approaches. Sometimes, this pays off. Here, I accepted that I would only resolve this issue by largely restarting the texturing process from scratch, with an approach that was outside my skill set at that time. I’d always been curious to try base color only texturing; this project, I realized, offered an opportunity to explore this technique.
This change in approach had a huge impact in untangling the knot of problems that were halting progress on the lamp project. I returned to the color design document, and settled on three specific themes: Antique Light, Candy Light, and Blossom Light. And then I began my base color texturing work.
Base Color Painting
Base color texturing is a little like converting a conventional 2D illustration into a 3D mesh. Every color and every detail must be included in the painting process – including lights. Essentially, you’re painting light into the scene, in addition to every other detail. Which means that you can’t count on the renderer to make the scene look good.
I started experimenting with this method, creating a few scenes to test out my ideas, and discovering new texturing techniques at each step. I’d found the idea of completely reworking the project very discouraging, and a big drain on morale – but here, I was learning something new, and this gave me a huge motivational boost.
One of my early 3D illustration experiments.
I really love this style because it frees the artist from any technical aspect during the texturing process. It’s also cheap to render and makes art direction easier. You just have to follow the sketches. This technique has been used on a few very cool projects, like the VR short film Dream Droplets.
I found that hand painting the textures in this way was the ideal solution to push the stylized look of my scenes. This approach required me to paint almost everything, which made the project even more challenging.
If you break down a base color project, you’ll find that it’s all about color gradients. It’s not a good idea to create these gradients by hand; this kind of thing can quickly become a nightmare. Instead, I experimented with Painter’s amazing generators and filters until found a way to convert simple masks into smooth gradients and details. Then I used a lot of anchors to drive some masks across the whole texture set.
The more I painted by hand, the more I found new ways to create complex textures. Notably, I discovered light baking, which allowed me to art direct all the highlights and shadows in the scene. It’s the same kind of thing I might accomplish with a standard light rig, but with a little twist – this way, I can paint light.
This discovery was the key to the whole project. Taking this approach, I also used the Smudge tool to distort the highlights, and to create a noisy dotted gradient, faking the same sort of thing you’d see on a hammered surface. This all required more work than using procedural shaders and materials, but it allowed me to tweak every detail of the render, just as I’d normally do in post-production.
Animating in Blender
Aside from the texturing work, I used Blender to do everything in this project. It’s a really versatile software tool; it can do a lot of different things together. We’d established from the start that animation would be one of the most important elements of this project, and that it would contribute a lot to the overall storytelling, so I found that Blender was the ideal choice. I’m a big fan of animation loops, and I thought that this would be the perfect direction for this project.
The idea was to unlock and reveal the lamp while making it grow out from the wall. Each of the three scenes has the same animation.
First, I set up a rough pass based on 2D Grease Pencil animations. Essentially, I anticipated the key moments by sketching some poses, keeping in mind the ratio and the silhouettes. It was kind of like creating an advanced storyboard, at this stage.
Then I compiled the key poses to make an animatic, and finally created inbetweens to have a solid base before animating in 3D. I find that making an initial animation in 2D is a great way to guarantee that I’ll focus on the end result, and that I won’t let the technical aspects of working in 3D restrict the animation. I animate at 8 frames per second, which allows me to really focus on the key poses and to art direct every single frame.
As a final polish pass, I set up VFX Grease Pencil animations on top of my overall animation. I adjust these extra details later, with the sound design. In the end, the 3D animation process is quite fast when you follow ready-made 2D timings.
Since I’d baked all of my light sources in the textures, I didn’t need to set up a custom light rig, as I typically would in most of my other work. I used Eevee as my final renderer. Blender’s real-time render engine gives me instant feedback in the viewport; this makes it much faster for me to check the final result, so that I can tweak the music to match the animation.
I asked Etap – a 3D artist, composer, and a good friend – to help me create some music loops that I’d use as a base ambient layer. For the three different scenes, we opted for a quiet, medieval mood, a candy factory background, and a spirit blossom vibe.
The background music uses some bells and drum echoes to illustrate the plain, empty environment. The peaceful atmosphere is established by the use of wooden xylophone, harp, and piano. After the first pass to synchronize the music with the visual, I used some SFX to illustrate some key explosions and bring more power to the overall animation.
Selecting the right color palette, and the texturing phase of this project, proved to be unexpectedly challenging. The solution was ultimately a shift in mindset, and embracing a willingness to learn new skills. It was an effective reminder that, no matter what level you are as a 3D artist, you need to remain willing to drop any rigid habits you might have, to better give your creativity free rein.
The choice to use a base color painting approach to texture the lamp scenes was pivotal to this project. By process of experimentation, and learning, I was able to gain precise control over the final appearance of the lamp scenes, while remaining with the brief of giving the project a stylized look, all in a completely unanticipated way. I’m delighted with the final result of the project; I feel that I’ve leveled up in some small way as an artist, as well.