Ali Jalali discussed his career path in the CG-industry and did a breakdown of his mythical character, Zahhak, shared his favorite software and mentioned a few useful tips on hair production.
In case you missed it
you might find this interesting
I'm Ali Jalali from Iran. Currently living in Singapore for about a year working as a character artist at Ubisoft company Skull and Bones project. I have been in the CG-industry for about 14 years by now, started by working in small studios doing modeling and animating for TV commercials and TV series in my hometown. After that, I moved to Turkey and worked for about 4 years in small studios. Working on my portfolio after work till midnight, and I started to get some freelance sculpting projects, which was so much fun, made me decide to come back to my hometown expand my freelance jobs and continue doing sculpts. After a while, I decided to move again, and here I am, in Singapore working full-time at Ubisoft as a character artist.
I have been interested in characters from childhood, after watching my favorite tv shows and cartoons, I used to draw their characters and add extra stories and characters to them.
After entering the CG industry, I knew already I want to model and animate characters. Starting off from small teams forced me to learn a little bit from everything so I could decide better which field would be my favorite and also helped me to do my job better in contributing to other production fields.
So I spent a few years making character models, rig them and then animate them, which made me realize that I’m not improving as I wanted, so I decided to choose one to put all my focus on it, and that was modeling. I liked animating as well but since I needed to do the rig as well it was less attractive for me at that time. Although my animating knowledge has helped me a lot in my character modeling, especially, when I’m doing a posed figure.
I like how you can tell stories in a single shot, frame or pose of a character and, even on a T-pose, you have the potential to tell stories, for example, how old is the character's cloth, which parts are torn, what caused the torn, you can tell what he/she has been through by many small details here and there.
Inspiration and Concept for the Project
Zahhak is a villain character in a famous ancient Persian mythology book called Shah Nameh (Book of Kings) written by Persian poet Ferdowsi between c. 977 and 1010 CE. This book has lots of stories and sections and has been completed for many years as the poet says he has spent 30 years writing and completing this book. It has a key role in keeping our ancient language alive today.
Here is a link for those who like to know details about this book.
About the Zahhak in short, he was the son of a ruler, he doesn’t have character stability, and, because of that, he has been used by Ahriman (devil) to grow chaos in the world. Ahriman appears to him in different stages and types, seduces him to kill his father, on another section appears as his cook and at the end asks to let him kiss his shoulders as an honor. After kissing, Ahriman disappears, and two snakes grow on Zahhak's shoulders that cannot be killed or cut as they grow again. Ahriman appears to him as a counselor later on advising him in order to prevent snakes killing him they should prepare them a soup made of two young men's brains.
Some times for my personal projects, I have ideas in my mind for a long time, I keep them there even for a few years, get inspired by many things during that period, and then there is a time that I feel I cannot do anything else except executing that idea being in my mind for a long time. For Zahhak, I have been inspired consciously by Lord of the Rings, Dark Souls 3 game, and renaissance paintings and, of course, unconsciously by many beautiful arts I tend to watch every day on the internet.
I gather tons of images regarding the details, poses, expressions, lighting, etc.
Everything you see in the final product has been changed a lot during production, as it goes on, I let it evolve and apply changes until I’m happy enough.
Asking opinions from my artist friends is another big thing for me. They are kind enough to give me ideas and help me shift the quality up as much as possible.
As for my personal artworks mostly, I like to make the character in a pose, so I start loose and do a rough blockout using basic shapes like spheres and cylinders. Paying attention from the beginning to major artistic values in posing such as a line of action, flow, gesture, and readability of the pose. Then I start merging them using dynamesh and add some bone and muscle landmarks trying to find my proportions and refine the posture.
Next, I imported a base mesh body and posed it like my sculpted model and projected all the forms to the base mesh.
Of course, it’s difficult sometimes to find the most accurate proportion in the pose, so I used a human skeleton and try to match its pose to my character, this helped me to revision my landmarks and proportions.
When I feel happy enough with my pose, then refining muscles and smaller shapes come into the process.
Even after studying anatomy for years, I still keep lots of muscles and anatomical references by my side when doing a project. I’m learning new stuff every time I do a body sculpture.
While sculpting, I might end up doing some parts a few times in order to achieve the result I’m looking for. For example, I did the old skin wrinkles about four times over and over again till I was happy.
For accessories and surrounding environment, I like to use basic objects as well and then replace them with more refined and detailed objects as the project goes on.
The Throne was a bit tricky to design. I didn’t want it to remind any region or period of time historically, so I tried my hard to come up with something a bit different that has a strong visual element. I used Maya for modeling and UV, Substance Painter to do the texturing.
As for skulls, I used a 3D model I have downloaded years ago from ten24 website. It’s a scanned skull, and I use it in my different projects and also for studying shapes and forms.
I used ZBrush boolean to make the broken skulls hollow.
Working on the Hair
For hair, I tend to do a rough sculpt in ZBrush first to have the main idea about the silhouette, length, and shapes. Then after decimating it, I bring that to Maya and keep it in my scene as guidance. In Xgen, I use guides to control my hair and try to get as close as possible to my decimated rough sculpt. Here are a few tips I always try to use in order to achieve a more realistic look for my xgen hair:
- Using maps for controlling density and hair width is mandatory.
- Adding Noise modifier and make it affect the hair in a random scatter can add a realistic feeling to the hair.
- While I use a clump modifier directed by guides, I also add an extra clump modifier to have secondary grouping strands in hair.
- Making long hair is challenging, and one of the difficult areas to handle are ears intersecting with hair strands. In order to avoid it, I recommend using Region Maps. By painting different colors on your base mesh, you can easily control the effect of guides on strands. Each region will have its own direction, and you won’t have hairs following a few guides at the same time. And, of course, you will need a clump modifier to control the hair directions. Just don’t forget to set the Region Mask number to 1.
For this piece, my main focus was on the final look, so I skipped many production necessary steps such as remeshing and UV mapping. For character, I used my already remeshed body, and, for environment and assets, I just sculpted in ZBrush and then remeshed them using Zremesher.
Regarding the UVs, I used the awesome UVMaster plugin in ZBrush, applied some poly groups on my object, and turned on Polygroups in UV master properties, so it does an auto UV according to my polygroups.
I also tried to avoid UDIMs to make it more straightforward.
For organic models, I usually do a first pass polypaint in ZBrush and then use it in SP to add more details and set the roughness as well. Exporting a Cavity map from ZBrush and adding it to the Substance file helps a lot as well. You can add it as multiply over the skin and, by tuning some levels and Hues, you can pop up your skin details, you might need to add a blur filter on it to reduce its hard-edged effect.
The most challenging part of texturing is always the skin because the process to create a realistic skin in CGI is not a straightforward process, you can’t just paint what you see in reference and call it done. It’s a combination of a few maps working together to make the final shader look like natural skin. You need to control Diffuse, roughness, SSS weight, SSS color, and scale. The process includes lots of back and forth to texturing software and testing the render in order to achieve a good result.
Here are my skin layers:
I tried to keep the skin shader as simple as possible, you can create more maps and masks to have better control over skin quality.
I love how Substance Painter has changed the texturing process in the industry for the last few years, coming from the age that Photoshop was used to make all textures and masks, I enjoy texturing more than ever nowadays.
Here is a GIF showing my throne layers in Substance Painter:
Well from the beginning, I gathered a lot of references to get an idea about my final mood, vibe, and lighting. Along the way to the final render, many things were edited and revised until I felt satisfied enough. Lighting became challenging when I added some candles around, it took me quite some time to figure out a good composition and light value for candles to not only distract the eye from the main subject but also to add the appeal of illustration. Controlling color values by light helps a lot to direct the viewer's eyes to a specific area of the image.
Here is my lighting breakdown for this scene:
I’m not familiar with compositing software such as Nuke so I came up with using a free Photoshop plugin called EXR-IO, just add your desired render layers in Arnold AOVs, turn on Merge AOVs, and you will have an EXR image containing all layers in it, which is extractable only in comp software and, with EXR-IO, you will open the layers in Photoshop and do the final comp.