Broken Ceramic Tiles

Material Breakdown

Yi Sun


Yi Sun

Material Artist


Hello, I'm Yi Sun (孙⼀), a Texture and Material Artist currently based in Shanghai, China. I studied 3D Animation at the University of Technology, Sydney.

I originally planned to work in the VFX industry but ended up in the video game industry. While I may be relatively new to this field, I've always been passionate about texture and materials since the beginning of my 3D journey.

I'm eager to continue learning new stuff and sharing everything I've gained along the way.


My planning process involves three steps: gathering reference images, analyzing each image, and breaking down essential elements. Just like we need a blueprint before starting to build a house, it’s absolutely necessary for us to come up with a plan for the material creation process. I would argue that spending enough time on the pre-production phase is more important than executing the plan.

White ceramic tiles are a very simple and straightforward, beginner-level material. You may think, “This is such an easy material, I can make it with my eyes closed without reference images.”

However, with a simple material like this, lots of details could be overlooked if we don’t pay enough attention. Simple material also means there is lots of room for artistic interpretation, so I want to challenge myself to transform this seemingly boring material into something interesting and beautiful.

I always gather my references from multiple sources.


Contrary to popular belief, I do use other people’s material work as reference from ArtStation and, but I’m not limited by that. I think they give me a good idea of how other artists interpret this material, what they choose to emphasize, what they choose to ignore, and how they handle certain details.

Additionally, I get scan data from Quixel Megascans, Substance Assets, and Poly Haven. I use scan data mainly for value reference, base color HSL, roughness value range, Normal intensity, AO intensity, and so on.

On top of that, I go to stock footage sites for some nice and clean close-up photos, such as Dreamstime or Alamy. I will also search photos on Flickr or Pinterest for more authenticity, to see what the material is like in the real world and the environment around it.

I also watch some YouTube videos to learn more about the subject I’m working on and how it’s made or formed. This knowledge might help me during the material creation in an unexpected way.

Sometimes I even watch ZBrush sculpting tutorials, oil painting, and miniature-making to help me understand the material better. So, in general, I use whatever I can get my hands on, as long as it can help me improve with the material creation.


After I get my reference images, it’s important to analyze them, breaking down each element and categorizing them. Not just leave them in the corner of the Pureref board.

Breaking down the material into small components is like making a to-do list for later, which has proven to be extremely helpful for me. By categorizing references, I will be able to observe more unique details on the same elements in various conditions.

The more details I can incorporate, the more believable the material will look.

Working with Height

Height is the most important map, and I need to make sure to give it enough love and care. I recommend spending around 2/3 of the time polishing and perfecting it. Having a solid Height map will help us later on when creating Base Color and Roughness.

Blocking is a concept I learned from drawing, animation, and 3D modeling. This means we start by building a low-resolution shape first and work our way up, adding more and more details as we go.

I highly recommend we do the same in the material creation. Especially if we have a node-based workflow, having a clear structure to work on will help a long way, otherwise, if we jump around on the node graph while we are working, we will get lost really quickly and overwhelmed by the amount of details.

I begin by setting up all the graph components for various material elements and label them in red.

Once all the components are prepared, I proceed to work on each one individually. Once I’m done with that component, I mark it as green and move on to the next.

The best thing about creating material based on modular graphs is that we can reuse different components in future projects. In theory, once we accumulate a large enough library of components, we can work on the same type of material much more quickly.

To be honest, I don’t have any groundbreaking techniques to share in this article, because I didn’t use any third-party tools. It’s all basic techniques you can learn from any tutorial, such as using slope blur for edge damage and using flood fill to gradient for surface undulation.

I want to emphasize in this article: The importance of having a solid plan before you start working on a material. The ability to observe details and extract the elements from reference images. The process of breaking down a complicated material into smaller components.

Base Color and Roughness

Although the Pick Color in the gradient map node can give me intricate color detail really fast, I am not a big fan of it because it is noisy and unpredictable.

Instead, I prefer to use multiple Uniform Color and HSL alongside with masks generated from curvature, height, and AO. This allows for better control and cleaner results.


When I’m working on Base Color, I don’t try to go for Quixel Megascan’s level of ultra-realism; I treat base color like a graphic design piece.

For example, I will try to avoid visual noise as much as possible to help the clarity of the material.
I pay a lot more attention to color palettes, visual hierarchy, and value contrast because they are important for readability.


After finishing the Base Color, the roughness should come naturally. I have my own way of doing roughness as well; I tend to not convert Base Color into grayscale or use curvature as the base of the roughness.

Because I have done that in the past, the problem is whenever I change the color value or intensity of the Normal later on, the roughness changes as well.

To avoid these issues, I prefer to use one grayscale value as the base of roughness, and simply add and subtract the same masks from the base color.
That way I can have better control over each element, creating contrast between sharp mask and soft mask.

Having a clear hierarchy of the grayscale values makes a reflection on the material look more dynamic and rich, I can also easily adjust the base roughness value without worrying about the need to change other parts of the graph.


Once I’ve finished the roughness map, I will switch the Base Color to pure black and check the reflection in the tomoco_studio environment. My aim is to ensure that there are interesting subtle details and contrasts between various elements.

I believe that when it comes to roughness, there’s room for artistic interpretation; we shouldn’t worry too much about strictly following the roughness value of the material’s physical properties. I think roughness is all about achieving a desired look and feel rather than mathematically correct values.

Lighting and Rendering

I think when it comes to lighting and rendering materials, there is no need for a complicated lighting setup or fancy HDRI lighting. Sometimes, one simple light source is enough.

A strong single light source, such as a directional light or spotlight, will create contrasted shadows to highlight the volume of the material and emphasize important details.

If there are too many lights and colors, not only is it distracting, but you will also lose the volume and details of the material.



Thanks for reading; I hope you all learned a thing or two from this article, then I will be really happy.

If you have any more questions or comments, please feel free to DM me on Artstation.