Researcher Robot: Learning Hard-Surface Techniques

Evgeny Romanov did a detailed breakdown of his Researcher Robot project, shared his modeling workflow in Maya, discussed topology in RizomUV, and talked about his approach to learning different 3D software. 

In case you missed it

you might find these articles interesting

Introduction

Hello! My name is Evgeny Romanov. I live in Russia. I’m 19, and I am doing 3D art. Now I work freelance as a hard-surface artist. I am self-taught. In childhood, I had LEGO, with which I build up a lot of buildings, vehicles, etc. I think that period of life gave me 3d imagination.

I have randomly found a video with sculpting some organic stylized characters in ZBrush by YuriyFyr. And I just started learning ZBrush and to become a character artist. After a couple of weeks of learning ZBrush, watching a lot of streams about the 3D industry, I have given up the university. After some time, I started learning polygonal modeling, and I like it more and I started doing hard-surface. For me, hard-surface is more interesting because it’s connected with functional mechanical parts. I am interested in developing how everything works. And, of course, I like sci-fi style so much. I love to see how organic anatomy can be transformed into mechanics.

Researcher robot

About the Project

I just saw a concept that caught my attention with some originality.

In this concept, I see his storytelling. He looks like a researcher with light armor. He fully outfitted with interesting equipment that helps observe the environment. If you can see, he has an open skillet, it’s really rare. And you can not do much cheating. This is needed to make movement mechanics with the back and head to work all together in the aggregate. And he looks a bit flat from the side. I decided these problems with organic make the silhouette look like an S shape.

I have a lot of references to real anatomic human beings, how the back, muscles, bones look like. Of course, he is not like a human, and I can’t apply all organic rules.

I have references from other concepts to see how another concept artist decides on the structural problems. I think the reference is necessary to create anything.

Modeling

First of all, I build all blocked basic primitives in Maya, even don’t combine them, I just build a basic silhouette with cubes.

I often do blockout with creases that really helps to show how smooth the shape looks, it’s really important not to make a mistake.

And I tried to make an interesting silhouette. I just move lattice points and try not to follow the concept.

In general, I like Maya navigation. How tools work, how works pivot editor works. I use a lot of plugins for Maya like NitroPoly, script pack “Armored hotkeys”, and more 20 self-written shortcuts.

I don’t care about ngons or triangles in subd pipeline if shading looks good enough. I do bad shading if nobody sees it, this saves a lot of time. I am sure that it is no longer a hack, but a skill that is often used in production.

Subdivision pipeline isn’t everything. Sometimes, I create some shapes with Subd, then I have a bridge with Blender (it’s a script) to do additional details with HardOps.

Why HardOps if Maya has HardMesh? Because HardMesh works unstable, in 70% HardMesh can’t boolean for some reason, and you always need search “right” parameters that often spoil rounded parts. HarpOps works better. And you can use MeshMachine together. If your model has repeatable details, you can create a plug with it.

I tried MOI3D for CAD modeling. It’s a really interesting program, but some limitations slow down a workflow for me. And now, I don’t really want to use CAD software. But I saw Vitaly Bulagrov using a plugin for MOI that helps recalculate subd geometry to nurbs. But anyway, I like more poly modeling.

I tried ZBrush for hard-surface, It’s super powerful software. I wish I hadn't realized it until later. In Zbrush, you can easily do blockout because you have a lot of ways to control the shape in sculpting.

For example, I needed to do big inside blocks, I just did a lot of live booleans in Maya, imported to ZBrush to make rounded edges with polish features. And you can control the strength of other parts with masking.

You need to understand what way gives you a faster result. Analyze what materials you will use to render. How many shadows would be there? I have some parts with bad shading in shiny material, but I already decided not to use shiny material for these. In the end, I saved time.

In my opinion, Subd isn’t so long if you have experience with it.

Inside of the structure in a cut.

In the future, I will train to work fast in ZBrush, I think if you train your hand there, it will turn out very quickly.

Topology

RizomUV has a better performance than other software for UV, he can work really fast with high-poly objects. I just separate the robot in 7 parts. And I uploaded each part for automatic UV. I did UV by hand for the bigger part. Rizom can work with bad geometry that has a problem with lamina, double faces, etc. This UV is not for production, but it’s enough for the portfolio. I'd rather spend my time on art than pointlessly on the technical aspects.

Texturing 

I think texturing isn’t my strong side. I was really inspired by chappie looking. I like his gangster street style.

I decided to do the same and draw all graffiti specifically as sloppy as possible, which would make it realistic. I thought it would be cool to have Japanese branding everywhere since Japan is famous for its robots, and if in the future it is the Japanese who have developed a lot of details for this robot.

Rigging

My friend did all the biped setup that robot. When he was rigging, I realized that I had made a lot of mistakes. After that, I started to realize how to concept more functional mechanisms. It was necessary to put in quite extreme poses and watch how everything works. Anyway, he did rigging well with the biped setup for using motion capture.

We have already used ready idle mocap and animated equipment. I fixed hand animation and changed angle rotation some parts for more interesting looking.

The animation was really important for storytelling stuff, that robot has parent small spherical mini-robot which looking around the area. And in general, the slightly stylized clumsy animation looks very interesting on robots. An example would be the Pathfinder robot from Apex Legends. Thanks to the use of childish, naive, clumsy animation, it looks amazing. 

Rendering 

I used Redshift for rendering. I did a couple of lighting setup and compared the pros and cons. I also worked with textures in parallel.

The first time, I wanted to try composing and multipass rendering for animation. I just make color correction, fixed problem with anti-aliasing, shadows, etc.

Biggest Challenges 

It’s difficult to answer because I did it in my free time. According to my calculations, it took exactly more than 250 hours. It was very much time spent on studying Substance Painter. I think that I spent about 150 hours on the blockout because a lot of things were thought out, ref board was going, some solutions were thrown out, I had to come up with a back and all the mechanisms, in fact, there is a lot of effort and time. But I got from this project the most important thing is the experience and pleasure during the process.

Evgeny Romanov, Hard-Surface Artist

Interview conducted by Arti Sergeev

Join discussion

Comments 0

    You might also like

    We need your consent

    We use cookies on this website to make your browsing experience better. By using the site you agree to our use of cookies.Learn more

    Researcher Robot: Learning Hard-Surface Techniques