Marius Pörsel has shared the working process behind the Chris Hemsworth project, detailed how the character's head, face, and hair were made, and explained how rendering in Arnold was done.
My name is Marius Pörsel, and I am an Asset Artist which includes Character/Creature, Hardsurface Assets as well as Environments.
Over the course of my career, I've had the opportunity to work on some incredible projects, including Ant-Man 3: Quantumania, Thor: Love and Thunder, and Spiderman: No Way Home, just to name a few.
Ever since I was young, I have been deeply curious about the natural world and everything around me. I spent a lot of time observing my surroundings and trying to understand how things worked. This deep curiosity and passion for understanding the world around me led me to the world of 3D, where I found a perfect combination of my interests with also a passion for technology, problem-solving, and mathematics.
To me, 3D is a medium with infinite possibilities, where real-life physics and nature can be brought to life in breathtaking detail. I have always been particularly fascinated by the art of creating realistic and accurate Characters and Creatures.
It's this passion for my work that motivates me every day, driving me to push the boundaries of what's possible in the world of 3D.
The Chris Hemsworth Project
Always with the aspiration to constantly improve myself and with the desire to learn new things, I attended the highly recommended Character Creation Course of Kris Costa. The course includes a wealth of knowledge from his experience on how forms and colors can be used and combined to replicate a human character – in his course with a focus on the human head/face.
To complete this project, I decided to create Chris Hemsworth, as I really like his roles in movies and how he brings his own personality to his performances. There are a lot of references/photos of him on the internet, so I collected several ones of the same age and preferably shot with the same camera, as the focal length can vary a lot.
Due to the focal length issue, I tried to reference general anatomy as best as possible and sculpt significant features of his face. I decided to sculpt him with a facial expression, which made things a lot harder for me, as there were not many different photos with the same facial expression.
However, I persisted with this choice, as I felt that it made the sculpture more engaging and allowed me to gain a better understanding of how muscles and fat in the face move to convey an expression.
Head and Face
To begin, I started with a basic cube shape and used Dynamesh to sculpt the rough anatomy of a human face, with a focus on getting the proportions and anatomy correct.
Next, I added the eyes and other characteristic features of the face, without worrying too much about the likeness just yet. To ensure that the sculpture looked accurate from different angles, I brought it into Maya and tested it with various camera settings, as the camera settings in ZBrush can sometimes be confusing.
Once I was satisfied with the overall proportions and sculpted features, such as the hooded eyes and mouth, I reprojected my Dynamesh onto a retopologized mesh with UVs. From there, I focused more on the likeness of the person based on reference photos and added skin details using HD-Geometry to achieve a higher resolution for specific parts without the need for excessive subdivision.
As all the details were sculpted by hand without any scans, I created my own human pore and saved it as an alpha to be able to place them manually. For the finer details, I used a combination of tweaked ZBrush-alphas with several brushes. The eyes were handled separately with one geometry for the iris with sculpted detail on it and one for the eyeball.
Hairstyle, Eyebrows, and Facial Hair
For the hair, I used XGen Core in Maya which is perfect to create specific hairstyles. One of the most significant advantages of XGen Core is the ability to use clump modifiers based on guides, which replicate the direction and flow of hair.
So based on references, I placed guides for the hairstyle, eyebrows, and lashes, and worked through the XGen workflow, using masks and modifiers to achieve a convincing result.
There are many helpful tutorials available that provide in-depth explanations of the XGen workflow for further explanations.
The texturing process was done in ZBrush with Polypaint.
I primarily used the Standard brush with several alphas to spray in significant colors for the human face. It helped a lot to have a look at highly detailed closeup photography of human faces as references.
In addition to the Albedo, I also painted a Spec-Weight map and a Bump-Height map using a black-and-white method. Finally, I painted a mask specifically for the lips to ensure that I could color-correct them accurately and achieve a more refined look.
So, importing everything to Maya, I used Arnold to set up the materials for the skin, eyes, and hair. While lookdeving the skin, I used a few color-correct nodes to get more contrast or saturation.
The shader of the hair was separated into two, as his hair is slightly more blonde at the tips of his front head, so I needed a mask for that. For my final rendering, I kept my light setup pretty simple – I used one key light, one fill light, and a kick light from the side.
In conclusion, I can say the biggest challenge was the facial expression as it has so much potential to look weird very quickly in case something with the deformation of the muscles is not accurate.
Maybe next time I would not start with a facial expression to prevent a bit of frustration but I am happy that I continued with it because I learned a lot about facial muscles, fat placement, and deformations.
You may find these articles interesting