Finally sitting down to write about this guy. Lot’s of technical issues, a pandemic, work getting super busy- made this process a tough one. My main goal for this project was to add Zbrush into my workflow. I learned from my shark model that trying to make high-detailed, organic models with Maya alone could be a big time suck. And part of increasing my skills also means increasing the proficiency of my workflow!

I did the concept for this owl and tried to go a little closer to the human end of the human-creature design spectrum. Eventually, I want to do a human sculpt, but I am playing it safe by sticking to stylized creatures as I practice my anatomy sculpting separately. I wanted to add some stylized, implied feathers to the body. I’m trying to learn one key thing at a time, so although it was tempting to try to do some PBR feathers in Substance, I knew better than to try to learn too much at once. I was going to limit myself to simple, flat colors for now.

Before I hopped into Zbrush I modeled a super rough, low poly version of a mesh for the body in Maya. I imported the OBJ into Zbrush, added geometry, and sculpted. I did not sculpt any of the clothing details since I was confident I would be able to make them low poly right in Maya. I think this was a smart move, as it allowed me to put all my focus in making a nice body mesh without having to worry about figuring out hard surface modeling in Zbrush for non organic bits such as the belt buckle. I did my best to sculpt in the feather details on the wings and chest.

Once the body was sculpted in Zbrush, I imported it back into Maya to retopologize. Now, this would have gone super easily (I learned so much from my excruciatingly drawn out retopology process of the shark) except that I thought I could model the feather details into the low poly mesh by hand instead of just using normal maps. I thought I was being smart by focusing just on adding Zbrush to my workflow, when I could have saved a lot of time, effort, and geometry by baking the normals from the high poly mesh onto the low poly! So yes, in my last project I baked normals when I didn’t have enough detail for it to work. And then for THIS project I didn’t bake because I was afraid it wouldn’t be worth it- even though I DID have the details for baking! Oh well, there’s always next time.

My too extra gene definitely took over when it came to the feather details. I kept adding in more and more geometry to try to make the edges of the feathers stand out. In addition to adding a bunch of geo that wasn’t necessary, I also made the transitions from the higher poly feathered areas into the simpler areas pretty rough. In the end, the feathers were barely noticeable and definitely would have looked better had they just been normal maps. So, for my next project, you sure bet I’m gonna bake some normals!

The other details such as the clothing and beak were straightforward to model. I remembered from the shark project that making a crater in the body model helps avoid clipping. So I made sure to keep a nice hole in the face to set the beak into, so that was good! These details did end up being much higher poly than they had to be. Later I realized I could have fixed this by retopoing everything together, which I’ll mention further on.

UV unwrapping went pretty easily since he is stylized and had a cloak covering a lot of the tougher seams. My next hangup however happened when my student version of Maya expired. I had luckily exported an OBJ of what I had so far, but when I opened it in the newer version of Maya, a lot of my geometry was messed up. The cloak in particular had to be redone a few times (I tried to speed through the redo because I wanted to get this project done) since I kept making yucky geometry when I extruded it for some thickness.

Finally though, I had a UV unwrapped, low poly owl boy that I could chuck into Substance Painter! Yo, Substance is FUN! I stuck to just adding base colors to the different materials. I assigned a different material to each separate piece in Maya, which allowed me to have them as separate UV islands in Substance. This was incredibly helpful and felt like I was working with layers in Photoshop. As I painted, I realized I needed to do some kind of eye painting or sculpting tutorial. Because there was not a good way for me to add the iris and pupil onto the eye I had made (because of its shape and how it rested in the skull) without making it look super uncanny. Instead, I just plopped an iris and some shine, which was different from the concept art but looked much less strange than it would have with a pupil.

To make the feathers I had so painstakingly modeled pop even more, I outlined them with a darker line. The amount of time and geometry those feathers cost me was NOT worth it! And at this point I realized I could have just drawn on the feathers in substance and left the wings as low poly as possible. I kept the materials mostly the same and just changed the colors. The eye, brooch, and buckle were the only parts that I played with reflectiveness and metalness for.

When I was happy with how the owl looked, I rendered it right in Substance. I would like to render in Unreal 4, but once again I wanted to save adding another software into my workflow for another project when I’m not as overwhelmed.

After completing the owl, I realized I could improve my workflow by retopoing all the separate objects into one mesh. Since I kept the beak, eyes, and each clothing piece as its own model, I was wasting geometry and potentially making it difficult for someone to animate and rig this if we needed the model to move. If you don’t see all of the inside of the boot, or the back of the beak that is in the head, there doesn’t need to be geometry in there! If you’re weight painting the model for animation, you don’t want to have to worry about multiple layers of geo (body and the clothes on top) moving correctly together.

For my next project, I want to make something small and simple in order to (finally) learn how to bake the details for a high poly mesh to a low poly and apply PBR materials. So I’m going to get really creative and make a dagger. Some nice scratches, maybe some rust, it’s gonna be a normal mapped and PBR-y kinda guy that won’t take me weeks to sculpt/model.

Shit I learned:
-Keep your retopology SIMPLE! Do not try to model details by hand that are on your high poly! Just BAKE if you have a high poly mesh.

-Retopo as much as you can into a single mesh if possible! You can save a bunch of geo on those clothing pieces and details!

-Don’t let your Maya license expire mid-project, heh.

-Substance is fun and putting different materials on your OBJ in Maya to organize your UV islands in Substance is big brain.

When I dropped out of my game art program I was determined to get good as quickly as possible while filling as many holes in my portfolio as I could. So, of course, my first project was WAY too big in scope. My approach with coming up with this project was to pick a studio I would want to work at, and try to make assets that could fit into the games they make. So I chose Dauntless. A monster hunting game when you slay monsters and then craft weapons and armor from the parts of the monsters. I would model a monster, then model a character wearing armor inspired by that monster, and make a bunch of weapons also inspired by that monster. Yikes. Too much! Especially for someone still figuring out their workflow.

I did the concepts for a few weapons and the monster itself, but didn’t finish sketching all my ideas before I jumped right into modeling. I was impatient, okay? I did some quick brainstorming sketches before landing on a monster with four legs. I was pleased that I did little sketches about how the monster might move and attack, since being a concept artist doesn’t mean you just design the look. You might have to design the movement, personality, attacks, etc as well.

To “warm up” for the big, scary, quadruped shark monster model, I decided to first make a simple weapon. I did not have Zbrush at the time, so I made a chain blade model entirely in Maya. And since I didn’t have Substance Painter either, I painted the textures in Photoshop and then put them on the model.

This chain blade was a pretty clear indication of my skill level at the time. I was wasting geometry, was presenting/lighting my renders poorly, and I was spending an insane amount of time trying to paint the texture so the seams wouldn’t look janky. But hey, I managed to get through the steps of concept, modeling, UV unwrapping, and painting. Now, I was ready to push myself.

The big thing I was looking to learn during this project was retopology. Which was bizarre, because I didn’t have Zbrush to make a high poly sculpt. Since I didn’t fully understand retopology, I thought I could make a really detailed, high poly model through Maya and still get detail through baking. Unfortunately, I also was going to rely on my hand painted textures to give the illusion of detail on the surface- so my biggest downfall for this project was not understanding how normal maps work. I did, however, give myself a break by purposely designing a creature that didn’t have a complicated texture such as feathers or scales. So I guess I deserve a few points for being kind of realistic about my skills for that part.

The model took me months to do. MONTHS! I started with one face and extruded one edge at a time using my concept art as in image plane. But I was adding geometry for the sake of adding geometry, I thought it would give me more control over fine details when really, it just made it harder for me to play with the shape of the shark. Just like in painting, I really needed to work big and broad and add details as needed. I did start getting more comfortable attaching the geometry for more difficult shapes (such as arms and tails) to the main body. But I was definitely not time OR geometry efficient, I wasted a ton of time and energy.

In my head, I thought having too much geometry would be fine, because I could just retopologize and then bake the high poly mesh onto the low poly anyway. So I retopologized it. I retopologized it for WEEKS! And then, the moment of truth finally came. I was finally going to bake all this DETAIL of my high poly monstrosity onto my low poly. Just like a professional 3D modeler would do! And lo and behold, there was pretty much no detail to bake. This was because my high poly mesh was exactly like my low poly mesh, just with more geometry! At this point, I realized that I might as well just go with my high poly mesh for the final model anyway.

So lesson learned. If you’re gonna retopologize, make sure there’s actual fine details that you need. And if you’re gonna model something, use as little geometry as you can since you can always add more later. The wild part is that even with me adding all this geometry to get nice and detailed, I didn’t add craters for the toes where the claws would rest in. So as you can see, the claws just clip right into the body model. Yikes!

I did go back onto my higher poly mesh and try to condense vertices to clean it up as much as possible. A very kind professional modeler in my Twitch chat provided me with this tutorial image. This was a skill I was happy I learned, because in my future projects I felt very comfortable going from a high poly area (such as the face) and condensing the geometry down at the neck, so that the chest did not have extra, unnecessary geometry.

For this project, the UVs and texture painting were also very time consuming. Mostly because I once again was not confident on how to do it and was still figuring out my workflow. I spent a lot of time painting over the seams again and again. I relied a lot on my lighting to blow out my colors and give it a cartoony look after rendering. I kept my shadows and highlights pretty tame and my brush soft, mostly because I was not confident enough to make bold details.

I’m still happy with how this project turned out because I was able to survive going from concept to a textured model. And the hours I spent modeling and then cleaning up geometry will helpfully allow me to speed up the process for the next project.

I never ended up making the armor or other weapons for this project because I realized my skills were VASTLY improving after each model. And I knew the quality of the project’s different pieces would be incredibly inconsistent. So I’m going to hold off on this idea until I have a better workflow nailed down. :]

Next project I’m going to focus on using Zbrush and Substance Painter into my workflow!

Shit I learned:
-Less geometry at the beginning gives you more control over the shape

-Merge vertices when going from high poly to low poly areas to lower the polycount

-Make craters on the main body mesh for things such as teeth and claws to go into so they don’t clip

-Don’t retopologize and then bake your normals if there’s no detail to BAKE AAAAH!!!

I’m starting this blog as a way to track my progress, self critique, and share information as I improve my 3D art skills. I became interested in game art a few years after graduating college when I really had to take a seat and think about what I wanted to do for a career. So I went back to school to study art and game design and was super pumped to learn new stuff and kick some ass!

I was loving my classes the first few semesters and was really excited that I was learning how to make art for games. Or so I thought. I had a big wakeup call when it was time to look for an internship and was met with job description after job description of a bunch of shit I had no idea how to do. Yes, my classes were fun. But they were covering things like video editing, flahs mobs, and abstract art. I didn’t know what Zbrush or Substance Painter even was, I had no idea how to rig, how to UV unwrap, or optimize assets for a game engine. I realized I had been putting too much trust in my school’s curriculum and decided I needed to take a step back and really figure out what I needed to know.

So I did research. I found game artists on social media and followed them. I watched tutorials on Youtube. And most importantly, I looked at the different job descriptions and wrote down all the software and skills they were asking for. Honestly, it was super discouraging. Here I was thinking I was making a ton of progress and was close to having a portfolio and skillset to get me hired. And then I had the rug pulled out from under me when I realized how far I really was from my goal.

This Patreon post was what showed me how far I really had to go. It gives the pieces artists had in their portfolios when they broke into the industry. This was where the bar was set for ENTRY level positions. It feels much more achievable to me now, but back when I first saw this, it felt pretty close to impossible.

I tried to teach myself Maya on my own, focusing on 3D modeling. But the interface and learning curve was so rough that I continued being frustrated for a long time. I eventually got to take a 3D class, which did really help. But I felt so behind where I should have been (being so many semesters in), that I made the decision to drop out and pursue learning on my own.

During my final semester, I was able to make some sloppy-meshed, horribly UVed, uninspired creature models. But I was comfortable navigating Maya and could finally use the program without being terrified one wrong click would ruin everything.

Now, I’m focusing on one 3D project at a time. Every time I finish one, I’ll do a critique on how I think I can make it better, the problems I ran into while I was working, and what I think I improved on since the last project. Hopefully, this can serve as a way to keep my own confidence up when I go back and see that I am improving. Also, I hope this can be a way for others to learn from the beginner mistakes thatI made!