My name is Yuditya Afandi. I’m a character artist and am passionate about modelling and sculpting. I also love drawing. I’m from Central Java, Indonesia. Currently I’m a student in Yogyakarta Multimedia High School.
I started to learn 3D when I was fifteen and now I’m nineteen. I didn’t go to any school to learn 3D; instead I mostly learned from YouTube, online courses, and books. But now I’m in a school that has animation degrees available. But so far I still feel like I’m learning more just from online stuff and talking to other people that have the same interest. But I still have to finish the school, because if I don’t, my mom will be disappointed.
I created this model to learn more about fantasy characters and tried to do this from absolute scratch to really feel the process and understand the workflow. I also wanted to learn to translate a 2D painting into 3D model, not only recreating, but I wanted to add my own touch to the character as well. Fortunately the 2D artist (Terin Kim) allowed me to do that for one of his pieces.
In this project I was really inspired by the look of Warcraft, Final Fantasy, and Dishonored, as well as art from artists like Adam Skutt, Blair Armitage, Carlos Ortega Elizalde, Maria Panfilova, Rafael Grasseti, Seungmin Kim and many more. It looks like a mix of inspiration but I think that’s what got me into the final render.
I used a PC with core Intel i5 CPU, 4gb of RAM, Nvidia 730 gt as the graphics card, and Wacom ctl 472 as my pen, the rest is just usual PC components. For the software side I use Blender for all the main tasks and Photoshop to help me paint the textures. Up until now dynamic topology is still one of my favorite features in Blender. Other things like Texture Atlas and Looptools were really helpful for this project. I’ll be honest - this is my first project working with Looptools within Blender. I didn’t know about that add-on before...shame on me.
I started this project by creating a basemesh using skin modifier, which is very handy. And then I sculpted the body straight using dynamic topology. I don’t want to show to body sculpt here because that might be uncomfortable for some people. And after that I just blocked out the costume. No detail at this point, I just wanted to see the overall look and where I was going to go.
And once all the blocking was finished I started to retopologize the sculpted model using basic tools like surface snapping, fill faces, etc. After the model was retopologized I transferred all the sculpting detail to the new mesh and then added more detail. I had to really be careful with the amount of polygons, because if it was too high my machine would stop responding and just crash all the time. It’s quite an old machine. For the final model it’s around 2,700,000 vertices if I’m not wrong.
Still blocking, anything could happen in this process. For example in the final result the eyes are actually smaller and the leg is actually shorter. Revisiting every angle and going back to readjust is just part of the process.After retopology, I did the UV mapping using texture atlas add-on to combine all meshes with the same types of textures so I could apply one texture to multiple objects and not worry about the scale of the UVs because all UVs had been “average scaled”.
After that I did some baking. A lot of it is just the AO to help me for texturing, so I could put that on top of the base color and set it to multiply and then shift the color to get something that I wanted.
This is one of the textures from the project. This is just a base. I did add surface imperfections like dirt, scratches, and smudges in the final material.And then I just painted the textures inside Photoshop. I did a basic rig with Rigify to get the basic pose and then I sculpted on top of it to fix some errors. After that I did the hair with particle hair. I added the parent hair and then combed it. I think the key for believable hair is the layering system with different settings for the child hair.
And generally I did the lighting first before I went to materials just to make sure that the lighting worked properly in basic 50 percent grey material. I used three point lighting (with additional lights) because I think it’s one of the best lighting setups for presenting my model and it showcases shapes really well.
After that I used basic PBR materials from Blender Guru, just plugged all the maps that I needed like roughness, diffuse, normal, bump, etc. I used to use a really complicated node setup but moving forward I tried to keep it simple and effective. For the skin specifically I used Blender Internal for the SSS while skin gloss and the rest rendered in cycles.
After all that I started to set up the camera with a 50-80mm lens. I didn’t want it too low or too high because that could destroy all the work that I’ve done. I rendered this project with only 300 samples with denoising on.
And then the final step was compositing the final image in Photoshop. Combining all the layers such as the character, shadow, background, etc. to create the final image. And doing some fixing if there were any small artifacts. I think it’s not a problem to get the rendered image and then add the final tweak inside image editor. Because it would take a lot longer to go back and re-render the image.
So that’s how I did this project. Obviously there are a lot more things that were involved but I can’t show you guys here because that would be too much for an article :D I’m sorry if there is any mistakes. Thank you very much.
About the Artist