ZBrush Teacher, Igor Puškarić, talks about his career path, his favorite tools, and creating Hybrid Runner character.
Intro & Career
My name is Igor Puškarić, I am a 2D/3D generalist, and an occasional Zbrush teacher at Machina Academy in Zagreb, Croatia. I create artistic outsource development, mainly oriented towards game industry in Sci-Fi and Fantasy, covering multiple areas, involving concept art, storyboarding, 3D modeling, texturing, basic rigging, animation, and illustration. I also design and sell 3D models for general purpose, and I am active on CGtrader, Sketchfab, Unity Asset Store, and Cubebrush.
I’ve been professionally engaged in the gaming industry for over 9 years. Currently, I run my own business, Iggy-design, which I imagine to be an outsourcing studio, where I am the only employee.
I have been a creator since birth, but the actual road was slowly unveiling in front of me over the years. In a sense, I always knew, I would be doing something creative, but the actual game industry opened up for me only many years later, when I got access to the Internet, heavily expressing myself, sharing it with the world and communicating world-wide. When I read a newspaper, telling about some Zbrush programme, I knew instantly, I needed to try that out, even though I had no idea, what 3D really was, and what it could mean for me and the world as a whole.
At the very beginning, I had huge insecurities because digital art as a career was unheard of.
Especially, the idea of working for yourself wasn’t on high demand.
With no arguments and proofs that it will be successful, and without any professional experience or good practical knowledge about it, I only had my vision and desire to create.
My social environment wasn’t understanding at all, with only a few exceptions, but somehow I managed to get my foot in the door with the unreleased indie game, called Novus Aeterno, and, eventually, I experienced some tangible progress, one mini-win at the time.
During this journey, I occasionally won some contests, got some features, gave some interviews and did several public speaking events on visual game design, so I am very grateful for each of those because they all played an important part in my professional growth.
The second significant problem was finding good people to work with, which you can trust. This was a very curvy road, but I met some really epic people along the path. I am very happy to say that now I am working with some of the best partners.
Once you have these foundations, everything else is significantly easier and mostly gets reduced to an easy-to-solve technicality. I have worked on numerous diverse projects over the years. Most have been canceled, but it gave me a rich learning experience.
I work well when I am involved in the majority of visual aspects on a single project, and I love working with small, multi-talented teams. I believe, it provides the best level of communication and understanding, which is mandatory.
In terms of successfully released projects, as a concept and production artist, I have worked on “Fleets of Heroes” mobile game by Inner Hero LLC, released in 2018, and now I am heavily engaged in “Tanks of the Galaxy” by SpaceJar, which is going public as we speak, preparing early access.
Combining 2D and 3D together
For me, I would say, it makes no difference in general, but a certain situation or a type of demand can affect my pipeline and approaches. At first, I construct a final vision, and, then, I think through about the most optimal steps that can get me there in the most efficient way, using less time and providing the best quality possible. It is the “right tool for the right job” philosophy. If I encounter a new specific style, I try to reverse-engineer it, study it and apply it to the projects’ needs.
I can say for sure, 3D has greatly helped me to understand 2D illustration better. Paying attention to volumes and levels is also a way of showcasing things and technical vs artistic appearances.
For me, 2D and 3D are inseparable, and they inevitably compliment each other. Also if I include animation, it does wonder by getting a message across, as they all equally contribute to the character of an asset/concept.
For example, if I need to do an icon/button, or I need precise shapes in multiple button variations, I might do a solid 3D base render, which helps a lot in the long run with iteration processes and can also be more easily animated for Sprite-sheet approaches.
Texture on a model can significantly affect the model’s character, and this is where 2D understanding is incredibly powerful. Also, depending on the case, I tend to leave super-high frequency detail empty on the high poly model, because I expect that to be better and more flexible with texture only.
Especially, when I can experiment more and derive normal maps using tools like nDo, instead of being a slave to a fixed standard burned-in information from the high poly. This also helps do wonders with texture skin in the long run, like this. Also, my inclusion in the entire project gives me a more clear direction in creating, which means, I can avoid design overlaps, and can more easily anticipate polycount expectations, not overuse the same effect on multiple assets and other technical aspects.
When I get to design and illustrate, for example, over 20 icons, which must be unique, this is hugely important.
I strive to view things as holistically as possible, so I merge 2D and 3D a lot.
There was a long-lasting desire to do something personal, in my own style, without any restrictions that otherwise exist in any official project.
It has been a while since I touched Zbrush in a committed way just for myself, and its approach to concepting organics is still irreplaceable for me.
So I guess I wanted to transfuse this personal feeling of freedom into a powerful, running, aggressive beast that would combine biological and mechanical worlds in one existential perspective.
In general, I love nature and all the diversity of it, so my initial global inspirations are hugely originated in this general direction all the time.
Later, during the process, I play, test and apply various ideas that may seem interesting at the moment.
Sometimes these mini-ideas fail, but they inspire me into something better, so I consider mistakes as brainstorming jumping boards, actually.
For me, personal projects are something that any artist should do from time to time, as they are most fulfilling and for me in general, they get to be most successful pieces of art I do.
I have experienced this weird success effect with my Mechanical whale and Worker 12 models, which have been both made up because of the same desire to freely create, just like the Hybrid runner concept.
For me, everything begins with a very rough, general but strong vision, including what kind of feelings a person can get from looking at it. This is the main and most important aspect I hold on to, through the entire process.
So, any used program or method or technical aspect becomes a tool in realizing the vision, regardless of their 2D or 3D nature. Of course, strictly technical tips & tricks are something you learn forever, and the more you know, the more your vision is empowered.
Also, I do not use any reference, because I like to play freely with it. When it comes to personal work, I feel reference would limit me creatively and experimentally too much. And it destroys any fun, because, at that point, I don’t create anything.
This is also a reason why I don’t do realism and humans, and why some of my work feels unusual.
I do, however, take general inspiration from things I experience in life through games, books, movies and my immediate surroundings. I just don’t copy the exact forms from there.
In the beginning, I started putting in the shape by inserting basic spheres and just roughly stretching them, using basic move and cutting tools, playing literally like a child, until I get that little “yes, that’s it” feeling.
It is important that this is done, while the subtools are still very low poly because their general shapes are the most easily definable in this stage.
During this time, I imagine it moving, so I am careful with the way parts could be animated and pay attention to them later.
It is a highly iterative process, but once I come to a certain base shape, I feel more confident to take it to another level. The confidence comes from knowing: once the main silhouette is set, the hardest design part is over.
In terms of shape definition, at first, I look at the main silhouette and the desired stance of the creature, and, then, I look at the way the parts visually cooperate in a larger perspective.
Meaning, as your eye slides across the object you can feel where it’s fine, and where more effort is needed.
After the coolness part, you need more space to rest again, and if it’s done right, it creates a psychological back and forward attention loop. In a nutshell, I try to create a visual fun-park in a single model, so to speak.
When the main design points are anchored, it’s time to make the model richer in detail. I am careful here, so a new detail compliments the original vision and fits as well onto the existing subtool, as it does globally when viewed on the entire model.
So I would put boxes, nuts, and bolts on machine parts, and scales, skin, and gills on the biological parts, which also helps to define the theme and makes the experience richer. What also feels good is some logical connection in the detail.
Before detailing, it is time to add more resolution to the basic blockout from the first stage. This resolution will allow higher detail placement.
With symmetrical parts, I often work on a maximum resolution for the details at hand, but then I duplicate, heavily decimate and mirror them.
This gives me an authentic overview of the concept while keeping my workspace fluent without frame rate being in danger since in Zbrush it is very easy to climb into millions of polygons if you are not careful.
Then, I start treating each part as a separate model to add detail into. In terms of design, sometimes it helps to imagine this subtool is the only part you have, so you can consider it as a new 3D canvas, which can help with inspiration for new shapes.
In terms of the final stage, I tend to sculpt my own high-frequency details as much as possible and try to use Zbrush default alphas if they are generally useful enough.
I will often go into photoshop and create my own samples, which can provide a great authenticity, or will sculpt a detail inside Zbrush and screen capture it using Grab Dock.
This approach, however, can be very time-costly, and my time was coming short, so I used a SciFi alpha pack by Jonas Ronnegard for high-tech mini-detail.
And I can say for sure, it was a really significant timesaver, and I highly recommend them!
In perspective of what kind of details I look for, the best ones are often those that imply some sort of logical and believable function. This greatly helps with immersion, especially if similar motives are already often seen in popular culture or everyday life. These known elements, when combined into a new cognitive form, can be very interesting to look at, simply because they are assembled in a unique way.
This way you can keep the familiarity and originality in a cool and fresh mashup. ?
Sketchfab for me is an incredible tool and a sharing platform that has helped me many times over the years. In regard to Hybrid Runner, I had no time to retopologize, UV, texture, rig, animate a model, so I decided to use the power of the Sketchfab’s 3D editor to quickly enhance its’ appearance.
Its’ editing options are amazing, and, some time ago, they gave me an idea that has helped me a lot with quick visualizations and rough concepting purposes.
First of all, I knew that a posed model has a far better implication of energy than a technical T-posed model, and as a concept will probably work better this way, so I played with poses in Zbrush, until I got the energetic running pose that it has now. This knowing has its roots from conceptual 2D illustration a lot.
Before uploading to Sketchfab, I decimated the model and divided it physically into several logical material groups. This is all done in Zbrush, now, with folders, it is a real game-changer, and I can organize my scenes incredibly better than ever before.
After partial segmenting, I merge all of them again, chromes into “Chrome”, rubbers into “Rubber”, skin into “skin”, and so on. Then, I export them separately, and I use Blender to import them back into one scene, but still divide into groups by selecting them all and export them again, but as one FBX.
And now, when the model is prepared this way, Sketchfab knows which parts belong to their material, which comes in as a blank slate for you to edit.
After satisfactory posing the model and the technical parts were ready, it’s time to have some fun. ?
I use Sketchfab’s PBR materials configurations to apply a different material to each logical part of the model. This involved a lot of trial and error, as the challenge was to have them work together.
I really wanted the skin to have Subsurface scattering to better signify the bio tissue right next to the mechanical elements, so I did what I liked.
Soon, I saw a more colorful HDR environment, which accented the model better, but the SSS setup wasn’t working as good, was overblown and too transparent.
So I re-configured it until I was satisfied.
Similar balancing process was a thing for every material group on the model, until I felt confident about all of them working together as a whole, with regards to color, saturation, reflectivity, etc.
For PBR it was important to set a moderate reflection value, so it didn’t soak up too much information from the active environment.
When it was good, I went into lighting and camera FOV areas, where I manually set up the light probes to further contrast/shadow ups and downs, wanting it more to be realistically conditioned than to have an equally lit studio setting. I also toned the lights in blue, which proved to put out a great balance to the model’s defined texture tones. This also has no strict rule, it is mostly an intuitive and experiential thing, and it also greatly depends on the model, and what feeling I want it to carry over.
Then, that I was happy to this point, the time has come to configure Post Processing filters. These can also mean a great deal for the entire presentation when edited moderately. The most important one for me is Sharpness, which I normally set to 20%, and the other one is SSAO (Screen Space Ambient Occlusion), which does magic for all the little indents in your model.
They work best when combined, and here are the settings I use for that:
Also, for cinematic purposes, I added a Depth of field filter which can enhance the realness and special depths when the model is viewed from up close. This is also very cool when portraying small objects like toys and figurines. Next, I went into VR settings – this is where you can set an initial size and point of view for VR perspective. Here you as the viewer can get a much better feel of size and mass of the model. I decided to make it quite large so that its supposed agility is still believable. For that reason, I also lifted it off the ground a bit.
Finally, I decided to create a making-of video and a good overall presentation. Later on, I edited, shortened and narrated the process, took a short video clip directly from the Sketchfab scene, imported it into Camtasia, added some on-screen info, and the showcase was done. I also learned how to animate cameras in Marmoset and produced an experimental “studio edit” for the default T-posed raw sculpt. My reasoning for that is to save people from actually clicking on the Sketchfab scene and waiting for loading, especially on mobile devices. Also, sometimes, Facebook can simply swallow the embedded Sketchfab scene or show it without a thumbnail.
All in all, I had tons of fun and learned a lot working on this project. The greatest real challenge was having enough free time to finish everything.
Igor Puškarić, 3D Artist
Interview conducted by Kirill Tokarev