Course overview Course overview
Stylize cartoon to comic-realism characters
The main objective of this class is to understand the process of creating a stylized character, and to have a finished stylized character at the end of the course. Students will start out by gathering their favorite concepts then laying down a foundation with fundamentals of the form. Students will go from primary forms to secondary forms which really boils down what it takes to make a stylized character. Furthermore, the course will go over features of the body and face to create that sense of appeal. What students will get out of this class is understanding how to get a stylized character from start to finish.
Stylized Characters in 3D WHAT YOU’LL LEARN
The more you know, the better.
Igniting your imagination
Hannah Kang works as a Digital Sculptor and Designer at Dreamworks Animation in Los Angeles, California, previously worked at Legacy Effects Studio. Hannah believes her best art makes you laugh! Your art is at its best when you know what you want it to do! Hannah Kang graduated the Gnomon School of Visual Effects in 2014. Hannah has been delivering characters, creatures, and assets on projects like Guardians of the Galaxy Vol 2, Avengers: Infinity War, Pacific Rim 2, Underwater and many more. Legacy provides one of the industry's most diverse portfolios requiring you to work on projects styles ranging from Call of Duty, Doom, and Crash Bandicoot to Avengers, Iron Man and the Georgia Lottery's Flying Pig!!!. While Legacy tends to focus on real-world translation of costumes and character builds that are created to be worn by actors. Hannah's personal style is often referred to as "bleedingly cute," showing a love for a feature animation style and a strong side of dorkiness--all with the aim of making viewers' first response to laugh out loud. Hannah has been teaching professional and student artists improve their portfolios for the past two years, and considers herself a student still--with a strong pursuit of continuous education. Hannah would like to someday work on children's books and create a universe of her own.
Stylized Characters in 3D Student gallery
Winter TERM Registration
Oct 26, 2020 - Feb 1, 2021
Hannah Kang is an exceptionally talented artist that brings a real world perspective on how to execute at the highest levels of the entertainment industry. Hannah gives excellent instruction, insight and direction on what principles to focus on when articulating aesthetics and design in 3D space. Being able to learn from Hannah in a live Q&A environment with weekly dialogue has really helped expedite my learning and progress. Hannah’s class is much appreciated and highly recommended.
Honestly, Hannah is the best instructor I've had since graduating university.
Companies that hire our students
environment design Benefits
What makes this learning experience unique?
Receive personal individual feedback on all submitted assignments from the industries best artist.
1+ Year Access
Enjoy over 365 days of full course access. This includes all lectures, feedback, and Live Q&A recordings.
Certificate of Completion
Earn a Certificate of Completion when you complete and turn in 80% of course assignments.
Learn anywhere, anytime, and at your own pace with our online courses.
Speak to an advisor
Need guidance or course recommendations? Let us help!
Show us your skills
Not sure if you have the skills, or are you proving you do? Show us.
The Warrior Sculpt
Interview with Christian Gonzalez
Christian Gonzalez took CGMA‘s class Stylized Characters in 3D and talked about the Warrior project created during it.
My name is Christian Gonzalez, I am a Chicano artist born and raised in Santa Ana, CA. I am a graduate from California State University Fullertonwhere I got a bachelor degree in Fine Arts in Entertainment Art/Animation. I currently work as a 3D Visualization artist at Norman International Inc. I have also interned and freelanced for Gadget-Bot Production where I created a couple of characters for their I.P. (Kaidro the Awakening) In the next upcoming months I hope to join my fellow titans and colleagues in the animation industry creating characters and worlds that bring every animation to life.
Ever since I was at CSUF I’ve wanted to expand my knowledge in 3D, especially character modeling and countless times I was suggested to take a course at CGMA. It was 1 year and a half after graduating school and many stubborn attempts into teaching myself how to properly create a stylized model that I decided to start my journey at CGMA. Honestly, it was the best decision I have made. I learned so much from Hannah Kangin 3 months than what I learned in a year and a half on my own. For this course, I wanted to understand a more cohesive approach to creating a character from a concept. I also wanted to gain speed and knowledge into how to fix any issues I could possibly have with any form of a model like this.
Going into this course I already had an understanding of a character pipeline in ZBrush. This allowed me to focus on Hannah’s main objective for this course which was making the concept come to life. Hannah broke down the process of creating a character to a T. During the first week, we picked out a concept. Out of the 3 characters I had gathered the Warrior by Anton Voronovich stood out the most to me for its distinct shape. I loved the proportions of this character and I knew this character would turn out great. Throughout the weeks Hannah sculpted alongside us so if we ever had any question of how something was done, we would know how to solve it. She is one of my favorite instructors alongside many from CSUF. She taught me how to read a concept and create a mesh that was studio-quality.
Start of the Character
When it comes to creating a unique and neat-looking stylized character I am the kind of person who isn’t afraid to push forms or start a model over if it is not working for the project. Luckily, I didn’t have to start this character over except for usual touch-ups.
After selecting my concept I mapped out my character with paintovers. Hannah pointed out that they would allow us to see the different parts we would need to complete our character. A simple but necessary step for a character like the one I selected with such dramatic proportions throughout his body.
Next step was blocking out the form into simplified shapes. In this case, I used a combination of spheres cut up and transformed to create the base mesh.
Detailing & Working in Maya
For detailing, I relied on two of my favorite tools which are Anatomy for 3D Artist by 3D Total and male anatomy figure also by 3d Total. After preparing the blockout mesh I dynameshed the pieces together and started to cut out planes and add secondary forms and muscles.
In the week that followed, I got pretty far in the model and decided to bring the character into Maya for retopology. After that, I started modeling the clothing.
My sword and other small props are all created within Maya. I blocked out the proportional shapes with primitives, then subdivided and manipulated the mesh in order to get the shape I wanted.
To make a character feel organic one must understand the body mechanics of a human really. Whether you are going stylized or more realistic, a general understanding of anatomy is key in order to make something look real and not bizarre. Now, to get a pose to look more organic, I usually study the human form and take reference that will allow me to create the subtlety in the pose that makes every overlapping mark work together to get a clean and organic result. Hannah said that silhouette is key and no matter what pose you go for, remember to balance your character.
When it comes to texturing a character whether it being stylized or not I always try to keep it as close to the concept as I can. For this character, in particular, I painted as I modeled in order to get a better feel of the character.
Once the model was complete it was time to render for presentation. Since this course focused mainly on the process of creating a stylized character in ZBrush I decided to use Keyshot. This external renderer allowed me to play with different HDRI lights and get great previsualization render as well as add different materials with no UVs.
To anyone who wants to improve their craft, I highly recommend checking out CGMA and their online courses taught by professional studio artist. It helped me develop a better eye for modeling. Thank you, CGMA and Hannah Kang for your awesome course!
Christian Gonzalez, 3D Artist
Interview conducted by Kirill Tokarev
Crafting a Stylized Character: ZBrush-Maya Workflow
Interview with Nabil Chequeiq
Hello, my name is Nabil Chequeiq, I’m originally from Morocco and I love 3D character art. For someone like me, it was hard to specialize in character art in Africa, especially since there’s no one to learn directly from, so I spent the last couple of years developing skills on my own. Recently, I had the chance to move to the USA to learn more and after that, I moved to Canada to work at ReelFX.
Working with someone from the industry and having the chance to talk to this person is a really great experience, especially if you enjoy what this person does. Before CGMA, I didn’t have an established workflow, but during the course Stylized Characters in 3D, Hannah Kang showed us the way she works and I enjoyed it. Now, every time I want to start a new project, I use the same workflow learned at CGMA. I recommend this class to everyone who wants to learn stylized character art.
Start of the Character
I chose to work on my own design, so Hannah told me to do work on shapes to develop a better idea of my character. Then, she chose the most appropriate shape that would serve my design best. When working out the idea, I started playing with cylinders and other forms – I found them to be easy to manipulate as you have more control over them. The biggest problem here was to envision the final result in my head and try to transfer it into 3D.
Most of the accessories were either created in ZBrush and retopologized in Maya or made from scratch directly in Maya. I’m a traditional box-modeler so if I want to start an asset in Maya, I choose a plain or a box with low resolution and then keep iterating as I go. If I want to do something in ZBrush, I use a simple curve brush, then export the result into Maya with a plugin called Styx to add holes and edges. Styx is a really a good export-import tool that makes my life so much easier.
For UVs, I really like to utilize the UV tool in Maya 2018. The biggest challenge with UVs is to know where to put your seams, plus don’t forget to add holding edges to your geometry so that if you smooth the object you don’t have stretching textures. It took me some time to figure out the right places for my seams because I wanted to use Substance Painter for texturing instead of Mari. I had to be careful with the cuts.
For detailing, I used Substance Painter and Arnold (for rendering) and for the small cloth details, I used XGen’s hair generation system. Then, I applied a hair material with the same texture I have on the object to make the overall colors uniform.
For the leather, I looked for a lot of references on the internet to learn how the leather is made. After the research, I jumped to Substance Painter and began exploring and testing different custom and smart materials to match the real-life references. For the stitches and seams, I used a MultiTool Pack for Substance I found here. It’s really a good painter brush that has all the seems and stitches I needed for the project.
I started using XGen only recently but I have found it really good and easy to work with. As a beginner, I don’t start directly in Maya and always use ZBrush to create my hair base. Then, I go to Maya to extract curves and convert them into guides. During the process, I faced a problem as I didn’t know how to give the hair multiple color gradients. What I did is create separate groups and then give each group its own gradient with some variety which worked perfectly as I had more freedom in terms of colors and control. For those who want to learn XGen, I recommend Tarkan Sarim, he is a really good teacher.
For me, every project has its own challenges from defining shapes to rendering or material and I learn new things every time. A good way to test your character is to ask people for feedback and always try to work outside of your comfort zone to learn more and more. During this particular project, I was fortunate enough to have the chance to talk to Hannah Kang and get feedback from her. And don’t be afraid of re-doing things!
Nabil Chequeiq, 3D Character Artist
Interview conducted by Kirill Tokarev
Yuzu: Stylized Character Production in ZBrush
Interview with Jason Chan
Hi, my name is Jason Chan, I am a 3D Character Artist from South Africa. At the moment, I am working as a Motion Graphic Designer at a small advertising agency. I have a bachelor's degree in graphic design, but my true passion still lies with 3D character design. I taught myself how to make 3D art in high school and kept frequently entering online competitions to improve my creativity. Recently, I have finished my contest entry for ArtWar4 at Cubebrush. It was a challenging process to create a unique character, but I managed to save a lot of time using the tips and tricks that I have acquired from Hannah Kang’s 3D Stylized Character course at CGMA.
I used to be a traditional 3D box modeler and all of my characters, up until Yuzu, were created using only a mouse and keyboard. Modeling organic objects was a painful process, and that is why I decided to join this course. For almost half a year, I struggled to teach myself ZBrush and didn’t find using a drawing tablet easy. My attempts weren’t appealing and they made me really doubt my artistic abilities. Hannah Kang made sculpting fun and easy, I managed to learn a lot in 9 weeks.
I enjoy creating my own 3D characters but now I want to focus more on digital sculpting, so I've been sculpting concepts made by other artists. When I do create artwork, I think about the content that I am presenting on my portfolio, and thus a lot of time is spent searching for ideas. Out of the blue, Manda Schank's Yuzu illustration struck a cord one day when I was browsing through Twitter. This character covered all the aspects that I wanted to learn in ZBrush. The theme also fits well with my baseball girl design BURN!!!.
Fortunately, Manda had followers who drew fan art of Yuzu for her. Those artworks became bonus visual references. I also took reference from character concept art for Disney’s recent movie. I liked how their facial expressions were illustrated and how simplistic yet impactful their characters were presented.
(Don’t forget to ask the concept artist for permission before you do anything. And always remember to credit them in your final work if it is not your idea).
In order to better understand the character's body shape, I like to do a quick paint-over of the concept. This traced sketch allows me to notice things that I might have ignored during the first glance.
The workflow is exactly as it is described on Hannah’s CGMA webpage. I try to keep my process clean and simple. Once I have done my high resolution sculpt in ZBrush, I bring it into Blender for Retopology and UV unwrapping. Then, I texture it in Substance Painter and render in Blender Cycles. I often go back and forth between different software because there are always changes.
During the course, we were made sure to have a solid anatomical foundation before we headed towards sculpting. The move, smooth, and pinch brushes were used at the beginning to build the body using simple shapes. After dynameshing and zremeshing it, the clay buildup and smooth brush came in handy. Hard-surface objects such as shoes, cap, baseball bat, and small accessories still had to be box-modeled inside Blender, because sculpting them wouldn't give me a clean result.
When I was posing Yuzu, I paid a lot of attention to the silhouette and focused on the parts where the body would turn. I used the transpose tool inside ZBrush, which is really convenient when it comes to posing quickly. However, using this tool sacrifices the chance for your character to be animated, leaving it as a digital statuette.
My biggest challenge in this course was sculpting the jacket. Provided with a limited amount of illustrations of Yuzu, it was difficult to imagine the whole posed jacket as a 3D shape in my mind, so I took photos of myself in a thick jacket, holding a baseball bat, as an extra reference. Hannah also suggested me to look into apparel wrinkles, as the thickness of the jacket is determined by their size.
The jacket was the only object that I didn't manually retopoligize, simply because Zremesher did a better job. As presented below, I have masked each section into poly groups so that ZBrush could cleanly Zremesh the whole jacket. In order to give the jacket that fabric impression, I polished it up with the alpha brushes from Poliigon’s fabric wrinkle brush library.
I intended to give Yuzu particle hair, however, for this specific character, poly hair worked out better and it used less time and power when it came to rendering the final image. A new material was given to the hair and tail so that each strand started off with a different gradient. You can follow the Blender tutorial by Pancake_Manicure on how to make poly hair (find it below).
I used Substance Painter to create different UV maps. This character wasn’t planned to be submitted for a game, so I gave her 4K texture maps. An additional subsurface scattering map was hand-painted which allowed me to control the skin’s translucency at different areas of the body. I also tried to keep the texture details minimal so that it doesn’t take the attention away from the stylized look and feel. Photoshop was used to enhance the UV maps so that they would come out more vibrant.
When it comes to lighting a character, I like to use a basic three-point lighting setup. An HDRI of an indoor environment was added to give the materials more reflection. A small point light is positioned in front of each eye so that the eyes have a white dot reflection. This is important because the reflections make your character look alive. Based on my observation in photography, I also added in a big soft light at the bottom pointing towards her face to act as bounce light.
The final result was color-graded inside Photoshop. A vignette was added to bring more attention to the character.
There will always be some form of challenge in art creation, but that’s what helps us grow as artists. I am glad that I had the chance to learn from Hannah, as I really enjoyed the course. It was worth waking up at 5 am every Friday morning to have a live Q&A session with everyone in the class (time zone issue). I recommend this class to anyone who wants to improve their stylized character art. Receiving advice from Hannah gave me a boost in my confidence and artist abilities.
Jason Chan, 3D Character Artist
Interview conducted by 80.LV
Stylized Character Art: ZBrush and Maya Workflow
Interview with Yon Lee
Yon Lee did a breakdown of a stylized character made during the CGMA Stylized Characters in 3D course and dived into the details of each production step in ZBrush and Maya.
Hello! My name is Yon Lee. I’m a senior animation student at the Art Center College of Design which is in Pasadena, CA. I love making 2D works in 3D. It is always exciting to turn 2D paintings into 3D models! I like to see how I change 2D designs into 3D. I’m going to graduate this fall, so I’m working on my portfolio now.
When I learned ZBrush for the first time, I thought it was pretty cool software because you can make things like real sculpting, I mean like real clay sculpting. Learning about 3D sculpting was a totally new world to me, so I just fell in love with 3D sculpting. It is the most exciting work for me, so I really enjoy what I do. While working in 3D art, I have worked as an intern for Real Art Daily Productions and a freelance 3D modeler for Epix Animation Studios and Ignite Animation Studios. After finishing my freelance work, I’m currently working on my personal project, a devil mermaid. This is also from the CGMA class.
The reason I decided to join the CGMA course was because of the instructor, Hannah Kang. She is a 3D modeler in Dreamworks Animation, and I have seen her wonderful artworks, so I thought it would be a great time for me to learn about 3D sculpting from Hannah Kang. My goal in the CGMA course was to be more professional. I thought: “Yeah, it is time to be more professional… Let’s do it!”. I wanted to know more about ZBrush sculpting and also the animation pipeline. Especially, I want to learn how to build details in ZBrush because I was not good at building details in ZBrush.
Working on the Idea of the Character
First, I decided to pick concept art which has big and small, nice stroke flow, and great proportion. David’s artwork has a big and small shape and nice stroke flow(look at his legs!), so it stood out among artworks. Also, David is a really great artist, so I always wanted to make his artwork into 3D. Thanks to the CGMA course, I could have a chance to turn David’s art into 3D. When I saw this artwork, I thought he is a nice guy who enjoys his life. Look at his face. You can feel how the character is relaxed and ready. So I tried to bring his nice feeling into 3D. Relaxed smile and eye-looking for a date, haha.
First, I made a body and face without clothes. It was important to keep an eye on the proportion because if you set the wrong proportion, it would be hard to fix when you posed the model. During this process, I mostly used ZBrush. By using clay build and dam standard, I make details on the body and face. Turning the camera view, so I make sure the model looks nice in every angle. While I worked on the body, it became heavy and heavy because I did dynamesh. When I thought my model was heavy, I always did Zremesh.
After making the body, I moved to the next part, clothes. For the suit, I used Maya because it was pretty hard to make the suit in ZBrush to me. Good thing was the original concept art doesn’t have much detail, so I could just keep focusing on the proportion of clothes. Comparing my work and the original art, I tried to follow the stroke flow of the original artwork. This was different from realistic modeling because when you work on the stylistic character, it is always important to keep the stroke flow. I know it sometimes doesn’t follow the anatomy, but it is for the stylistic world. It is fine, you don’t have to make everything realistic.
Working on Colors
I matched colors on the original art because I wanted to bring the mood that the original artwork has. I used the Arnold shader in Maya. I didn’t put detail for the texture because what I wanted was making people focus on the shape for the first.
Working on a cartoon-y character is tricky. It is like a wire-walking between stylistic and realistic. I wanted my model to look fine in the turntable, so I made the list for 1 to 10. The first thing was matching the front view of the concept art. It was the most important thing to me. Then, I turned the camera and checked if it looked okay in every frame. If there were any mistakes, I went back to ZBrush and fixed it.
For the detail on the suit, I sculpted all wrinkles in ZBrush. As the character is stylistic and cartoon-y, I avoided making clothes too realistic. My character has nice stroke flows, so I tried to keep these flows in wrinkles, too.
I also wanted to make a wireframe turntable video, so I did retopology for everything like the body, the suit, the hat, and etc. The wireframe should look clean, so always keep in mind to keep the clean and nice topology. The good thing about the wireframe is you can check overlapped polygons on wireframe mode. If props overlapped too much, I fixed one of them. It always helps to keep a clean and nice wireframe.
Lighting progress is fun, but also hard work for me. What we need to get from lighting is the nice shape language, but if you use too many lights on the scene, you will get a flat and boring render. Before doing lighting in Maya, I used ZBrush BPR render process. It is good to use ZBrush lights before rendering in Maya. You can get a fast test render by using ZBrush lights. If you think test lights look fine, then move into Maya and set lights.
Try to set lights as similar as possible like in ZBrush. When you test the render, keep the render setting as a low resolution for saving time. You can set the high resolution after the test render looks nice. As I said above, I didn’t want my character to look flat because of too many lights. Therefore, I only used three lights. Two lights are for showing the nice silhouette. Lights on the edge of the character help to show the shape. However, sometimes you want to make some parts darker. Especially when the light hits too bright on the shoe. When I tested the render, I found that the left shoe looked too bright than the right shoe. So, I put the light blocker to make the left shoe a little bit darker. Then, I added one more area light for the face. The face is the most important part for me. I always want people to focus on my character’s face for the first time, so I usually put one more light on the face. Making the face part brighter than the feet part is helpful to hold people’s eyes on the character’s face. When you put the focus light for face, don’t forget to make it smooth to avoid too many differences from the bottom part.
Posing! Posing is the biggest challenge for me. It is always tricky to pose the character even though you put the original concept on the side. In my case, I render in Maya, so it always has small gaps between the ZBrush camera and Maya camera. When I sent my first ZBrush pose to Hannah, she gave very detailed feedback. As you can see in the left pic, my first pose was in the wrong position and proportion. Also, his eyes were too big, and it didn’t have a relaxed smile on his face. It is helpful to ask for feedback from other people because when you are working on the project for a long time, your eyes cannot catch the wrong parts due to tiredness. In my case, Hannah’s feedback was correct to fix my posing.
Based on Hannah’s feedback, I fixed the proportion and rescaled eyes. After fixing the pose, I imported everything into Maya. When you set up the camera in Maya, you will realize that camera is not the same as in the Zbrush. So, there are some differences when you render in Maya. It happened to me, too, so I fixed again in the Maya. When I pose the character, I try to put the mood of the original art. As I said, I wanted this character to have a relaxed and ready smile on his face, and the body should have a nice shape flow. When you fix posing and it looks fine, then it is time to render!
During the CGMA course, I learned that every stage in sculpting is important. Blocking, sculpting the body and clothes, retopology, rendering. Every stage is important to make a nice 3D model. You can’t just say “Ah, I don’t like sculpting the body, I’m gonna fix it later when I do posing” like this. Keep eyes on from blocking to rendering. One small part that you missed can come up with a big flaw later. So put your focus and love on each stage!
I’m going to make some environment models for the next, and my goal is to make the character with the environment, so it looks like a shot of animation.
Yon Lee, Character Artist
Interview conducted by Ellie Harisova
Creating a Stylized Viking Warrior in ZBrush and Substance Painter
Interview with James Hyun
James Hyun did a breakdown of his stylized character Viking Warrior made during the course Stylized Characters in 3D at CGMA.
Hi! My name is James, and I’m a 3D Character Artist from South Korea. I focus on making stylized characters and props for games and animation. I started learning 3D at Rochester Institute of Technology, where I earned my BFA in Film and Animation (focusing on 3D Animation). I graduated in 2017, but it wasn’t until early 2019 when I transitioned from trying to make it in 3D animation to focus on 3D modeling.
Before I started modeling and sculpting, I spent a little bit of time freelancing in feature animation and advertising post-graduation. I was lucky enough to contribute to Blue Sky Studios’ 2017 film Ferdinand as a Stereoscopic Technical Assistant. I also got to get my hands on a whole bunch of projects ranging from TV and online ads for brands like Verizon and Cinnamon Toast Crunch, content for billboards in Time Square, etc. as an Assistant Editor at a post-production company called CHRLX.
I recently finished a remote freelance contract as a Junior Character Artist with Stalwart Games, a game studio in LA, so I’ve been focused on personal projects these days to grow my portfolio and skill sets. I recently finished my latest character piece Egyptian Mage, check it out here. I’m also open to new job opportunities!
Viking Warrior: Goals
I took Hannah Kang’s Stylized Characters in 3D course from July to September of 2019. I was going into the class with a generalist knowledge gained from university, and some extra modeling know-how learned from YouTube and Gumroad tutorials. I wanted to put my accumulated knowledge and skills to the test and make sure what I was doing was fit for studio production. With Hannah’s professional modeling experience at Legacy Effects and DreamWorks Animation, I thought the class was perfect for me.
Our first week’s assignment was to come up with a handful of concepts that we wanted to sculpt, and the Viking concept was not actually part of my initial selection. In my efforts to make the most out of each character I tackle, I have a small list of guidelines that the concept I choose should meet. These requirements include points such as “does the concept have hard surface elements?” and “is this concept going to challenge me in a way that I haven’t been pushed before?”, etc. To be honest, none of the initial concepts I brought to class hit the nail for me. I was paying a premium for this course and had this valuable opportunity to get one on one time with an industry professional, so I wanted to challenge myself and give it my all. So in week 2, when we were supposed to start blocking the concept of our choosing, I instead continued my search and stumbled upon Viking Warrior by Kim Sunhong on ArtStation. The moment I saw the design, I instantly knew I had to give it a shot. It was beautiful, it was complex, and it had a back view, which really helped out. The character was just detailed enough that it would pose a challenge, but still doable if I pushed myself and made the most of class.
Approach to the Character
I’d say my ZBrush workflow is nothing out of the ordinary: low res blockout - high res dynamesh sculpt - retopology (typically in Maya) - high res projection onto the retopologized mesh - finish high res sculpt on retopologized mesh.
Earlier in 2019 before taking the CGMA class, I had taken a full stylized character and hard surface tutorial by Matt Thorup (aka the RedBeard) and an anatomy tutorial by Rafael Grassetti. These really helped me learn a typical ZBrush pipeline and basic anatomy before getting into the CGMA course. It also gave me more opportunities to ask Hannah questions pertaining to my character specifically and get information only possible in a closed environment like a class, rather than spend time on generic knowledge you can get outside of class. You can learn typical workflows and anatomy on platforms like YouTube and Gumroad, but discussing project-specific techniques for a particular type of hair (as an example) can only be done when you have direct access to a mentor/instructor.
Because the Viking concept was a little more on the realistic side of the spectrum, Hannah and I did extra work on the character’s face to stylize it a little more and fit the purposes of the class. I also approached the fur on the character in a more stylized fashion (rather than using hair cards or Xgen) as I had never done stylized polygonal fur before and it was one of the challenges this character presented. A blobby base shape of the fur was placed appropriately and detailed with Simon Chapman’s Fur Brush set. To keep the character from looking overly noisy and maintain a stylized look, I made sure to vary the shape and sizes of the fur clumps, laying down the larger clumps first and filling the gaps with smaller tufts after.
Overall, I tried to find a balance in the detail level to be believable but not necessarily photoreal. I think my personal sweet spot is that “stylized but not overly cartoony, and realistically proportioned but not photoreal” look if that makes any sense to you. This meant my character might include more detail than your typical “stylized character”.
Smaller details would be added later through textures, but any larger surface detail that cannot be tiled (such as stitching, pores, etc.) was sculpted in ZBrush. While these details can be done in a texturing suite such as Substance Painter, I personally had more access to tools in ZBrush for such purposes at the time. Since then, I have slowly been amassing more tools for Substance Painter so that I have flexibility in what stage I add surface details. Whether you add details in the sculpting phase or texturing phase, they both have their own pros and cons.
UV mapping was done in Maya just before posing the character when she is still in her A pose. Nothing out of the ordinary, I try to pack the UV map as efficiently as possible without overlapping and keep separate materials grouped together. This wasn’t a game character so there were no worries for how many UV sheets were being used.
Back in ZBrush, once the character is posed with the Transpose Master tool, I pick a subdivision level for each subtool that is high enough to retain the shape and larger details of the mesh but would not bog down the texturing and rendering process. I export out this version of the character as a .fbx file to use in Substance Painter and UE4. I also baked the normal and height maps from ZBrush to use in Substance Painter. Any masks and color maps I painted in ZBrush will be exported for texturing as well. Bakes can be done within Substance Painter too, but at the time of making this character, I was more comfortable with baking in ZBrush. It’s all up to you in what pipeline you use!
Like I mentioned before, my characters aren’t super cartoony, and considering the detailed nature of the concept, I tried to pack in as much detail into the textures as possible without going overboard. Having “just enough detail” is a fine line to ride and is something I still struggle with today. I wanted the character to be beaten up with wear and tear from her previous battles. You’ll find fabric and leather patterning, dirt grunge, scratches, splatters, stains, blood, etc. all around which were done with a mixture of customized Substance Painter materials/smart materials, purchased alpha textures, and hand-painted methods. I’ve personally gotten a lot of mileage out of Travis Davids’ tileable texture packs for adding detail on almost any surface.
For the hard-surface elements, once the overall shape of the model was made, I gave it a generous amount of subdivision levels. Then, I masked the shape of the design and simply pushed it in or out with the 3D gizmo. Make sure the mesh has clean and even topology when doing this so that the pushed faces won't cause too much aliasing and require minimal cleanup. I often like to start hard surface elements in Maya rather than in ZBrush due to Maya’s more comprehensive topology manipulation tools, but ZBrush 2021 is supposed to have added new topology features so maybe I’ll be able to keep the process entirely in ZBrush with the new update. I didn’t retopologize afterward to conform to the design at the time, but in hindsight, it would have been a good idea for optimizing the mesh.
For her hair, I started with a rough block out to get the overall shape. Then, I smoothed and retopologized the block out to get the underlying shape that the hair strands would be placed on. For the hair strands, I used the RB Flat Hair Brush, provided in the course by Matt Thorup I mentioned earlier. For her shaved side, I simply placed geometry inside the character’s head and pushed out where the shave would be.
I like to render in real-time due to its time savings and near-instant feedback. Yes, traditional render engines can give better image quality, but my lack of patience, increased visual fidelity, and time savings keep me allured to working in real-time. I’m personally a firm believer that with the development of technology (especially real-time ray tracing!) there will be a convergence of traditional and real-time rendering sooner than you think!
Anyway, Viking Warrior was rendered in Unreal Engine 4.23, the latest version at the time. The lighting setup started with a base studio HDRI plugged into the “Sky Light”. I added two directional lights for the key and fill lights respectively, and two vertical “Rect Lights” on the rear left and right sides of the character for rim lighting. For supplemental lights, I added a small “Rect Light” in front of her face, and a larger and wider one on the floor looking up. The “Lightmass Importance Volume”, “Post Process Volume”, and the “Sphere Reflection Capture” objects are also critical in getting pristine renders.
One thing to note is that for all the textures used on this character, I forced the texture mipmap levels to be lower than the default values, thus producing higher quality textures in the viewport even when the camera was far away from the character. While this can create unwanted visual noise particularly in lower resolution renders and video, for the high-resolution static images that I would take for my portfolio, it was the perfect solution. While you can completely turn off mipmap levels per texture, I personally reduced the mipmap values in the master materials. You’ll want to decide for your particular project whether you want no mipmap levels at all or if you want to reduce the mipmap level visible based on your camera/viewport. This video by Elias Wick was one of the many that helped me figure out mipmap levels:
For the actual raw renders from the engine, I used the “High-Resolution Screenshot” function at roughly 6K resolution. Why so high? Mostly just because I can and to have more information when doing touchups in Photoshop. In hindsight, 6K was probably overkill and ate up more file size than necessary. I’d recommend sticking with around 4K resolution unless you have the processing and storage headroom and like super high-resolution images (I mean, who doesn’t like high-resolution images?).
For the turnaround video, I used the “Sequencer” feature and exported it at 2K resolution. A group was created for the character mesh and keyframed to rotate 360 degrees for the turnaround. To render in the “Lighting Only” view mode, you’ll have to open the Level Blueprint and enter the command “viewmode lightingonly” in the “Command” section of the “Execute Console Command” node. Then, plug the “Event BeginPlay” node into the “Execute Console Command” node, compile, and render from the Sequencer.
One of the challenges with this character was figuring out how to get the highest quality renders possible from UE4. Since I was only concerned with visual fidelity, I pushed the poly count, texture resolution, lighting quality, render resolution, etc. all to the max that my PC could handle. A lot of time was spent tweaking mipmap levels, antialiasing methods, light settings, etc. and fighting with glitches, engine quirks, and small mistakes. Also from my understanding, the default anti-aliasing method in UE4, Temporal AA, doesn’t even work when exporting videos from the sequencer or while in-game. Why is that the default AA method? I’d personally recommend you change that to FXAA or MSAA in the Project Settings section.
The lighting itself tends to pose challenges for me as well because lighting isn't my specialty. The end results usually work out as long as I put in enough time adjusting and testing but I often go through a process of self-doubt in the beginning when things still look ugly (which is similar to my sculpting process too). I start off with the typical three-point lighting setup and build up to a more complex system.
It’s been almost a year since I finished the Viking Warrior, and while I’ve learned so much since then and there are aspects that I would have done differently looking back, it is still one of my favorite projects. It helped get my first character artist job with Stalwart Games, and I am very grateful to Hannah for all her help during and outside of class. Further thanks to fellow artists and former CGMA classmates, Jason, Ahmed, and Joseph. We all still stay in touch and this connection and networking has been one of the valuable benefits of CGMA.