, by Louise Melin

A Material Exploration of Joseon-era Korea

Louise Melin breaks down selected materials from her Joseon-era Korea collection, and discusses her approach to learning Designer.

  • Interview

Hey! I’m Louise and I’m a self-taught 3D artist specializing in materials and textures. I’m a philosophy teacher by profession, though I’ve also taken to spending my evenings creating materials with Designer. Those two activities aren’t typically linked, I feel, and I’ll talk a little below about how that happened. But, for now, I’ll leap into discussing one of my material collections, an exploration of some of the materials present in the Joseon era in Korea.

Exploring Technical and Symbolic Perspectives

Before beginning this collection, I’d previously created one other big collection of materials, my Lord of the Rings sets, as well as some standalone materials – I think smaller projects are important, because they allow you to take a step back and experiment with less weight on your shoulders. But I was eager to start another collection, and searching for a new project.

At that time, I was watching Netflix’s Kingdom, a Korean drama filled with magnificent textures. I liked the idea of learning about another culture through artistic exploration, so I started planning sets and gathering references. This project would ultimately take me four months to complete, and involve the creation of dozens of materials, each one with its own challenges and appeal.

I spent a lot of time reading about Korean culture and history. Each object needed to be understood both from a technical point of view (how was it made?) and from a symbolic point of view (what did it mean?). Those two approaches are crucial and complementary. The first relies mainly on reference images, whereas the second is more about immersing oneself in the culture: why did they use those specific colors to paint their ceilings? Why did their rank belts always have 20 plates? How heavy was the Chinese influence? And so on. This cultural knowledge of your setting is important, I feel, and brings meaning to your work. Like this, you aren’t just manipulating pixels on a screen – you’re continuing ancient traditions in your own way. This awareness also contributes to the quality of your materials: for example, if you understand that the famous moon jars are part of a wider ceramic current which reflects the spiritual shift from Buddhism to Confucianism, then you better understand their whiteness and their simplicity better.

For instance, moon jars symbolize frugality and humility. In considering the question, ‘What did it mean?’ we’d answer that their slight asymmetry and imperfections are part of their beauty. Approaching the question, ‘How was it made?’ we’d see that the two halves of the jar are assembled just before firing, and thus the stains are often noticeable where the two parts have been joined.

Let’s look at some specific materials. I created a lot of materials for the collection; here I’ve selected some examples that I feel are the most representative of my workflow. I’ll explain briefly how I approach material creation, then detail the processes involved.

My Approach to Material Creation

Keep the final scene in sight

The first thing I think about when tackling a material is the environment it will be a part of. This approach may not fit every project, but I try to keep in mind that a material is not an end in itself. It’s easy to lose sight of it when you’re working in Designer, tweaking nodes and playing with your HDRI in the real-time viewer. Your material may look good at that moment, but it’s only when you load it into your final scene that you spot any issues, realize what needs adjustments, and so on. So quite early in the creation process, I open Blender to test my material, to see how things go.

It’s always best when materials interact with one another in some way: creating contrast, casting shadows, making light bounce around them, and so on.

Break down complexity

One question I’ve frequently been asked when I share my WIPs is how I can be patient enough to build very complex materials. The answer is in the question: patience! Though this also requires optimization and careful planning. You should always keep in mind that complexity is nothing but a combination of simple elements. Once you get those elements right, you just have to combine them. Designer’s node-based workflow is perfect for that.

If you try to spot all the repeating elements in this picture, you’ll see that there are a lot of them (thankfully!)

The Embroideries

There are two important embroidered pieces in my project. The first is the dragon roundel worn by kings and queens; the second is the rank badge worn by civil or military officials. Both are beautiful pieces that took a long time to reproduce.

Let’s look at the dragon patch. The first thing we should be searching for is repetition. Is there any uniformity we could build upon? In this case… not much. The only elements that really repeat themselves are the stylized leaves and clouds around the dragon. But the dragon itself is made of unique elements that can’t really be reproduced. Even the scale pattern is hard to tile from a single unit because of its warping and twisting (this wasn’t a problem for the other dragons I made, however; I’ll discuss this later).

The second thing we should notice is the thread flow: this is what really made this material a challenge. First, you have the filling embroidery at the back. This was fairly easy to make because it’s either circular or unidirectional, so I could simply rely on a tile sampler and a shape mapper node. And then there’s the top embroidery, with golden threads that follow the dragon’s silhouette using a couching technique.

To achieve that effect and to have control over what gets on top of what, I made the dragon embroidery on a subgraph with an output for each part of the body.

This is how it looks for the head part:

It may look intimidating, but it’s actually pretty simple. I made one fiber as a building block for all the embroidery, using an approach sometimes used by Pauline Boiteux.

I then used this fiber as an input for tile samplers and splatter circulars.

Here you can see that I made the head’s edge using a warped splatter circular.

The same is used for the nose: swirls and warps gave me the result I wanted.

All I had to do was to repeat this process for each part of the dragon, each time looking for the best way to direct the flow of the thread. Sometimes a splatter circular was the answer, sometimes it worked best with a tile sampler and a swirl. Once the body was done I could reassemble all the pieces in my main graph, using levels to control the height hierarchy.

The albedo is fairly simple on this one. Quick tip, though: I like to use a leveled version of my curvature map, mask random parts of it with a Perlin noise, and use it to highlight bits of the embroidery by blending the map with itself in add mode.

This material was a little laborious to make, but I learned a lot from it – and so, when I needed to make the second embroidered patch, I knew what to expect and what mistakes to avoid. This second patch is called a hyungbae (literally “front and back”) and was much easier to make.

First notice how all the threads followed roughly the same direction in this reference image; there isn’t much curving and twisting here.

http://www.metmuseum.org/art/collection/search/49599

Second, the shapes are rather flat and independent of the thread flow. This means we can deal with them separately. This time though I realized that making everything inside Designer wasn’t always the most efficient approach, especially for ornamental shapes that don’t need much tweaking. So I drew almost all the shapes by hand, using a simple layer system to export each part of the embroidery separately. For the cranes I only needed two bitmaps: the main shape and the edges. It would have taken me hours to do this with shape nodes, 2D transforms, warps etc., and the result would have been less accurate. Instead, those two masks took me a few minutes to draw.

But once you have your flat masks, how do you convert them into interesting height information?

First, you need to give them a bit of depth by either blurring them or beveling them. This will give your height map a gentle bump. At this point though the actual embroidery is still missing; you need to add your threads. And to do that, all you need is a tile sampler:

I could spend hours talking about this fabulous node, so let me just explain why it’s interesting when dealing with fabrics and embroideries. Fabrics can be understood as particle systems of threads, and the tile sampler as a controller for particle systems: it drives the placement of the particles, their size, their grayscale values, and so on. The rotation controller is one that I use the most. It can take two inputs: a vector input (hence my normal node) and a grayscale input. Here I plugged in a heavily blurred version of my mask to make the threads spread out from the center of the cranes.

I then used this as a base, adding lighter values to create bumps on certain parts, darkening the places where I wanted the edges to appear.

And here’s how it looks (above)! Pretty cool, huh?

Here, a simple checker node gave me this great result for my lower ornament:

and here are other results I got when simply playing with it, trying out different inputs, testing things for the sheer pleasure of it.

The Celadon Bottle

Working on ceramics with Substance Designer, now called Designer, was probably my favorite part of the project. First because I’ve been passionate about pottery for a long time, and second because Designer proved to be perfect to deal with this type of material. What I also liked about ceramics is that they bring variety to your workflow. For these objects the height and normal information are not as important as for other textures. Roughness and albedo maps, on the other hand, are crucial; they need a very subtle approach.

Let’s talk about the albedo first. For a piece as ancient as this little bottle (my reference was more than 8 centuries old) there are two things to consider: the original celadon color which was created during the firing, and the oxidation, wear and stains caused by the passage of time. So I made two passes:

The first pass is a very subtle mix between two shades of greyish green: if you can barely see it, that’s good! It takes some effort not to overkill things because Designer often tempts you to go over the top (adding too much noise, or too much damage, or too much contrast, etc.).

The second pass is for the oxidation and dirt stains. For this one I opted for a BnW spots 2; it has that crispness that works well for small-scale specks. I only randomized it by adding a Perlin as a mask, something I often do in order to break the library grunge maps into something new and less recognizable.

Then there are the cracks… but, wait a minute, where should the ornaments go? If you pay attention, you’ll notice that the crazing must go on top of the painted leaves, and that’s because when the ceramic cools down it’s the upper layer that makes the rest crack, then covering the ornaments and colors underneath. In my opinion, this is one of the most fascinating phenomena in pottery. It creates a unique pattern and adds depth to a surface that’s otherwise flat and smooth – each crack refracts the light in a specific way. This is why we have to make the ornaments first.

For this I relied on my usual bitmap approach. I drew simple leaves by hand, imported them, then altered them in Designer with light warps and slope blurs. But what makes it work is this simple trick: I took my mask, blurred it and leveled it, then blended it with the non-blurred input. It’s so simple and yet so effective: look at how the painting seems to seep into the ceramic! That’s exactly what happens in real life, and I was really happy to have found a similar effect in Designer.

I could then add my cracks and various noises on top of it. To give the material that slight graininess that can be seen under the glazing I made two passes of gaussian spots, the first one dark, and the second one lighter with a slight offset.

I used that same technique with the cracks, faking depth by offsetting my noise with a lighter value.

Now let’s talk about the roughness a bit. It’s what really makes the object feel like ceramic, and Designer provides us with a very interesting tool to achieve that: the clearcoat outputs! Clearcoat basically adds a transparent layer on top of your original material. It can be used to achieve a great variety of effects, including ceramic glazing. But it doesn’t just add a shiny cover, it also carries its own normal map, thus creating a very interesting system of layers. For this specific piece I chipped the glazing by blending the normal and the roughness map of the clearcoat with a Dirt 2 and a moisture noise

I was careful not to go too dark with the roughness, keeping in mind that this piece had a nice, weathered patina that made it a little rough.

Overall, I think that the clearcoat maps offer tons of possibilities for textures: they can be used for varnished materials, obviously, but you can be even more creative with them: you can apply them on top of transparent pieces, you can use them for organic materials such as food or flesh, create frost/ice layers, experiment with them on fabrics, and so on. As usual with Designer, the sky’s the limit!

The Dragons

Let me finish this breakdown with one of my dragon pieces. I textured a lot of dragons for this project and ended up finding a kind of routine to approach them.

This time I couldn’t really rely on my bitmap workflow: those sculpted dragons are not as flat as embroidery or paint; they have too much height information. So I took my time and built each part of the dragon inside Designer with warps and blurs.

The good thing with this approach is that you can preview in real time how your displacement goes, something your eye usually can’t pick up looking only at the grayscale map.

The main body parts were multiplied with this scale pattern that I used on all my dragons. It’s a very simple tiling with a vertical offset.

To make your sculpt more dynamic, you can blend your shapes with directional inputs (I mainly used anisotropic noise and stripes). Here, this adds a nice flowing look to the dragon’s mane.

Then you just stack shapes up. It takes some practice to think and plan things in grayscale values, but it ultimately becomes so natural and easy that you almost never get stuck. You instinctively know what needs to be leveled, darkened, masked, etc.

Here, again, each individual piece is simple – it’s the combination of these simple steps that provides a complex result.

Once you’ve got a satisfying height map the rest is pretty easy: you add a little bit of normal noise/damage; you build the albedo, roughness and metallic maps using the shapes you’ve already created as masks… and that’s it!

Tips and tricks

– Experiment with the vector input of the tile sampler to control the repartition and rotation of your pattern.
– Diversify your damage/wear techniques: slope blur is not the only answer! Several passes of multi-directional warp often give an interesting result. Your cracks/damage shouldn’t cover all your maps; don’t forget to use masks when adding them. I think part of Substance Designer’s magic resides in the art of proper masking. More than half of my blend nodes use masks.
– Don’t insist on making everything procedurally: sometimes it’s just not the smart approach. Focus on what’s efficient and on what saves time.
– Don’t overlook the other maps: you can achieve a huge range of different effects in your shader editor. Here (below), for example, I used a mix of two directional noises as a translucency map for the lanterns and the shoji panels. This allows the light to flow and to cast interesting shadows.

My Approach to Learning Designer

You may wonder how a thirty-year-old philosophy teacher ended up spending her evenings creating materials with Designer. The answer to that question is the same as for many other people: covid! Living through a pandemic and having to stay at home for weeks left me with an unexpected amount of free time. At that time, I didn’t have any classes to teach, so I was basically free to do whatever I wanted for almost six months.

I’ve always considered myself a creative person, and even though I didn’t take this path professionally I still tried to keep the flame burning, mainly through photography and drawing. Everything around lighting and composition fascinates me, so when I discovered the 3D world in March 2020, it was a revelation. Suddenly I could create light – multiple lights! – add fog to a scene, add in as many cameras as I wanted, and so on. It was exhilarating. Hours would fly past when I was in front of my computer (they still do). I was quickly drawn to environment art, I loved the idea of creating immersive scenes in 3 dimensions. But I also realized that those scenes were missing something, even with good lighting. That’s when I became interested in materials. I came to understand that the way objects look and feel, the way they interact with light, and ultimately with the eye of the viewer, is driven by their textures. I dabbled a bit with Blender’s shader editor but couldn’t really get the result I wanted. At the same time, I was awed by the stunning material orbs in Artstation… I was and still am inspired by so many talented artists: Pauline Boiteux, Daniel Robichon, Jan Trubač, Matthias Schmidt, and Elie Paquiet, just to name a few… I’m always amazed by the way they give life to their materials even if their approaches are very different.

Night Cape, a standalone project.

I quickly noticed that those artists all had one thing in common: they worked with Designer. And that’s when I fell into the rabbit hole!

Having zero background in the field, I’d be lying if I said that Designer’s procedural workflow didn’t intimidate me at first. I was daunted by those huge graphs, and I felt at first that this wasn’t for me, and that I would be terrible at it because my background is more literary in nature. I was wrong! Even if Designer’s nodes seem complicated to you, you just have to keep experimenting/playing with it until the process becomes more familiar.

I started like everyone else: I followed tutorials. The first two that I watched were the tutorials on muddy ground and old plank materials by Wes McDermott, which are on the Substance Academy channel. Wes is a great teacher; these tutorials really are responsible for my love story with Designer. I think that once you have followed those (that is, once you’ve actually created the materials step by step), you’re pretty much set to experiment on your own. It’s still best to keep watching other follow-along videos just to see how other artists approach their materials and enrich your own workflow with their ideas. But after a certain point, the best way to learn is to start your own projects.

The value of long-term projects

The Shire, a part of the Lord of the Rings material collection.

When I got past the tutorials phase, I understood that I needed a mid-term goal to practice and improve my skills. I also needed that because I had to reconcile my day job with this new objective of becoming a 3D artist. Clear project planning helps a lot juggling several tasks and staying focused without losing interest in what you’re doing. My first long-term project – the Lord of the Rings series – was what really kickstarted me as a material artist in the community. There are a few advantages to working on large-scale projects like this:

– It encourages you to have a more environmental and narrative approach: materials are never isolated pieces; they function together, echo one another and, in the ensemble, they create a coherent scene. Working on sets of materials helps to keep that in mind. It provides your work with context.
– It gives you a clear roadmap: once you’ve settled on a certain theme or common thread, you know what you have to do and you can plan the weeks ahead in a more efficient and engaging way. When I was working on the Lord of the Rings, it gave me a real boost to know exactly which pieces I’d be working on after I’d completed my current set.
– It keeps your audience engaged… and makes you accountable. Working on long-term projects with regular updates and publications creates an expectancy in your followers. They’ll be curious to see what you’ll do next, and you’ll feel more committed to it (and less lonely, if you’re working from home).

Rivendell, a part of the Lord of the Rings collection.

What’s next?

I have hundreds of project ideas in my head – which is great and overwhelming at the same time, especially as classes will soon resume and I’ll have to split my time and energy into two once more. Finding that balance is not easy!

My objective is to keep improving with Designer, but also to adopt a more professional approach to it: I still have a lot to learn in terms of optimization and fitting my work into a pipeline. There are so many new and exciting features that I look forward to playing with, notably procedural modeling. I also want to get started with Unreal Engine this fall to see how I can make my materials work in real time.

Hopefully I’ll keep growing as an artist while also becoming more efficient from a technical point of view.

The complete Joseon-era Korea material collection, in four parts, can be viewed on Louise’s Artstation page.

Meet Louise Melin

Louise currently divides her time between teaching philosophy and creating 3D materials. Her 3D work to date includes the creation of comprehensive material collections, such as her Lord of the Rings collection and her Joseon-era Korea collection.

Read more