LogInDmitrii Kondratev joined us to talk about the legendary Nagant revolver recreated in 3D with 3ds Max and ZBrush, detailing his approach to the texturing process and rendering setup with Substance 3D Painter and Marmoset Toolbag.
Introduction
Hello everyone! My name is Dmitrii Kondratev. I am a 3D Artist, but for the past year, I have been completely immersed in weapons and work at Digital Forms as a Senior Weapon Artist on a realistic first-person shooter. I always wanted to develop games, but for a while, I didn't understand what that really meant. I enrolled in a university for information systems and technology. Despite having a good understanding of programming languages, I realized it wasn't quite my thing. Maybe I was lucky because I dropped out, and a day later, a friend called me and suggested an internship at a 3D studio. The day after the interview, I opened 3ds Max. From that moment, it became my main program. I didn't understand anything and want to thank my team lead, Alexey "Jips" Klimov. I don't know how he tolerated what I brought him.
I think I managed to do everything, starting with characters, then equipment, ships, tanks, weapons, and stylized props for mobile games. Realism, semi-realism, sci-fi, and the like. I was drawn to weapons because of their forms; modern weapons are quite aesthetic, but over time, I realized that WW2 weapons attract me more, probably due to their imperfections: many prototypes, inaccuracies, some weapons were just welded from pipes, many defects, and the like.
Main software:
- Autodesk 3ds Max
- ZBrush
- RizomUV
- Marmoset Toolbag
- Adobe Substance 3D Painter
- Adobe Photoshop
Revolver Nagant Project
It so happened that I got a bit tired of weapons, or rather hard surface, and decided to try myself as an environment artist. And with this task, I said goodbye to weapons. I searched for a weapon I would like for a long time because I make modern weapons at work, so I chose something old but intact. I wanted to make something small and compact, obviously a pistol. I looked through a bunch of pistols and read their histories, and my choice fell on the Nagant revolver. A truly legendary revolver. It has been in hundreds of movies. The simplicity and reliability of its design, the high penetrating power of its cartridge. It was said it "should stop a horse". Everything was implemented, and hopefully, no horses were harmed in the experiments.
I started by collecting references. The standard was the 1912 version. My advice is to gather as many close-up references as possible, especially worn ones with defects, chips, dust, and the like; this will help not only in detailed modeling but especially in textures.
Modeling
As I mentioned earlier, my main software is 3ds Max. It was decided to do everything in it. Almost everything was done using Boolean operations, which is very important because you can adjust almost anything until the very end without starting the detail from zero. I also have a mini-pack of probably all existing bolts because modeling them repeatedly is boring.
First, I make a simple blockout, where I create the main forms without small and medium details, and then I move on to medium details. At these two stages, for weapons, Boolean is not mandatory but highly recommended. To understand this, look at how firearms are made. Ordinary machines, blanks, drills, milling cutters, and the like are used. All this is replaced by Boolean in a 3D package.
In general, when you don't understand how to make a certain detail, just imagine how it would be made at the factory, you can look it up, of course, it comes with experience, but it's better to start with the basics and understanding right away. Without this, nowhere. Next, I move on to small details, at this stage, the LP is finished.
For high poly, I use the subD method. I'm just used to it, and I can add some defects in ZBrush on top of it. This is the end of the HP stage:
Topology
In this particular work, my attention was more focused on textures. Therefore, I didn't think much about topology, leaving it mid-poly. It was important for me to spend as little time as possible on all stages except textures and rendering, because I have less experience in them, which means it will take more time.
I approached the UV unwrapping a bit more carefully, probably out of habit. I divided the UV into 3 sets of 4k to make it more convenient and faster to texture. The main part, the inner part with the magazine and the pistol grip, the pistol grip is wooden, so it was easier to separate it. The main program is RizomUV.
One of the UV sets:
Texturing
The main program is Substance 3D Painter. Tone mapping – ACES (I consider this a mandatory point. Each software and engine works in its color space, and ACES helps average this out. You can read more about it here). Usually, I start with a height map, texture, inscriptions, defects, and the like, using anchor points. In the future, I can adjust the parameters for color/roughness/metallic together. After this stage, I set up the base, the main color, the main variations on all channels.
A problem arose at the texturing stage. The problem of choice. The revolver is old, and there are too many variations of wear and tear. There are different types of metal, just painting over random metal processing. Blued, Heat Bluing, Browning, Rust Bluing. Too much information. Therefore, I chose references with different wear levels but the same processing.
First, I made the height map, texture, and chips according to my classic scheme. Then I proceeded to the base. This time, I decided not to bother too much and took the main color variation from the metal albedo with Megascans. I mixed it with the main color through an overlay. Then, I added roughness variations using standard grunge maps from Substance 3D Painter.
A big problem was with the wooden handle. It turned out that I had almost never made wood, or rather gun wood, if you can call it that. There were many options, but all were unsuccessful.
In the end, it was decided to start with a base without texture. In the future, it will give the correct color variation and roughness work. The large facture, or rather Checkering – a method of applying notches to stocks and grips to increase reliability in hands (non-slip), will add depth and not spoil the picture. And it just looks cool. When the base was ready, everything played a bit differently.
The result satisfied me, and I decided to finish there. I recommend looking at this SKS, as it helped me a lot to understand the layer hierarchy and some tricks, such as using stripes for the main fibers. Also, speaking of textures, I recommend choosing references not for the sake of references but for interesting references. The main thing is to think when choosing the main reference for texturing, you should understand where and when this item was used, who used it, how expensive production was, this will help understand the required level of wear of the object.
Rendering
Render, render, render... This is my main problem: I have almost no experience. And what I had was unsuccessful. But I didn't give up. This was the final push, so I had to finish.
Specifically, for myself, I decided that I would almost not use post-effects during rendering. The main thing is to choose the angle and set the light, the rest I will pull out at the post-effects stage, thanks to the person who invented Adobe Photoshop. I decided to make two types of renders: first – on the table, trying to authentically fit it into another texture, second – abstract, practically without unnecessary details. The HDRI is standard, I just chose a random one that I liked and gave a little blue tint. Then I made two renders, separately the revolver and the table with shadows.
I decided to highlight individual parts so that the highlights were a bit uneven. It was decided to make abstract renders just on a black background with subsequent additions to the background.
There, I also added a spot light to emphasize the shape. After this, I used Photoshop, where I applied the Camera Raw filter, which allows for finer post-processing using curves. To understand the work of curves, I recommend watching this video. Unfortunately, I'm not a guru of these tools, but it helped a lot.
Conclusion
In conclusion, I want to say that you just need to not give up, no matter how banal it sounds. This work took a lot of time, primarily because of the frequent desire to abandon it and start something new. This is a typical story for many artists. It seems like you're doing something wrong or didn't get what you wanted, although from the outside, it's clear that the work has great potential. Therefore, I want to thank my girlfriend for pushing me when it was needed.
When you want to create quality work, first and foremost, focus on references: could there be a chip there? Why is it here? How did it appear here? I often ask myself these questions; it helps me stick to reality and not invent reasons for certain characteristic defects in the model. And one of the most important things is to ask for feedback; this is one of the key points. First, you don't know everything, second, you don't see everything. Also, you just get used to seeing it. In general, it's a must-have.
Strangely enough, I encountered problems at every stage of my work, the reason being ironic: I got used to working in production and following the tech task, which is deeply ingrained, and it's very difficult to deviate from this program. Therefore, I forcibly did many things differently, surprisingly! I had time to think, and now, looking at the work, I clearly see things I would like to improve at all stages. Probably, excluding low poly, the task was not about that. I think I managed to convey the physical properties of the metal quite well, but I would like more variations in the base color. I am satisfied with the wooden grip, although I would like to make it a bit more accurate. To refine the masks a bit. In the end, I am also satisfied with the render, although it also has its flaws.
The final beautiful result is made not only by your work but also by the environment. I get a lot of pleasure when I see how a particular artist creates a composition. Adds other objects to the scene.
Anyway, I finished the work, glad it resonated with people. I want to thank 80 Level and especially Gloria Levine for the opportunity provided. It inspired me to grow further, and I hope it helps others too!