Flintlock Pistol
Introduction
My name is Jan Kasperczyk, and I’m a Senior 3D Hard-Surface and Weapon Artist based in Kraków, Poland. I’m excited to have the opportunity to feature my work here again. Currently, I’m working at Room 8 Studio as a senior artist on multiple high-profile projects, such as Call of Duty and Apex Legends.
Previously, I was a senior artist at Bloober Team, where I contributed to the Silent Hill 2 remake. I graduated from the Academy of Fine Arts in Katowice, specializing in graphic design.
Alongside my professional work, I’m passionate about teaching and currently run online courses and my own original mentorship program to help aspiring artists grow their skills.
References
I wanted to widen my portfolio with another weapon. I made a decision I’m going to do a vintage gun from before the first world war era. I started searching for good references.
During my search, I went on a trip to my local National Museum in Kraków with my wife and child and was lucky to visit the historical weapons and armory exhibition.
When I wanted to find the photos of the gun that was of interest to me I stumbled upon a website of the New York museum The Met.
There, from many guns that I found I chose the Flintlock Repeating Pistol with Lorenzoni Action, Bearing the Crests of Vice Admiral Horatio Nelson (1758–1805). It had the perfect angles of photos, and they were very well documented.
I used photos of other guns as additional references (Of course more than once I wanted actually change the target of my project but I stuck to the Flintlock Repeating Pistol with Lorenzoni Action telling myself I always can make another one).
I searched for some additional references but I have to admit the museum ones were more than enough.
The quality of the references is something unique. When doing a prop like this I highly advise browsing through museum sites. The angles and resolution are unmatched. The end result will benefit from that approach instead of a google image search.
Blockout & Mid/High modelling
For the entire modeling process, I used Blender 3.2. The blockout and basic shapes were honestly the hardest
part of the whole process. Without excellent references, it would be a very long task.
I started with a blockout of the three main components. The grip and the main body with the pipe. For setting it up I used the references as backdrop images from all four sides
For the modeling of the form and first-level details, my goal was to be around 80% accurate with the references. I didn’t want it to be identical because I felt some design decisions could benefit the model.
The back part of the grip would not look good in 3d and wouldn’t offer any interesting challenges. I decided to keep the form and change some of the details.
Shaping the base with simple techniques. Getting the proportions correct lets you model calmly and with pleasure in the later stages without coming back to change anything major.
Modeling of the high poly was just a process of layering the details and shapes. For most parts, I used simple Modifiers (Mirror, Bevel, Subdivision, Weight Normals)
For the places I marked with the letter A I cut out the holes with Booleans and then refined it by having on the grip part.
After having my high poly done I prepared it for ZBrush. The parts I knew I’m not going to dynamesh I subdivided with quad-looking geometry so i don’t get weird sculpting artifacts. I exported the model to Zbrush as and FBX with all the parts as separate subtools.
Sculpting
After the import to ZBrush, we are choosing to import as subtools and than we have to determine which parts of the gun will go through heavy topology changes. For example, the grip will have carvings sculpted into it, so it has to be dynameshed. All parts of this nature I dynameshed and all others just subdivided.
Every part has been treated with the polishing of the edges to make them look more natural.
After I was done with sculpting the imperfections I would copy the file and one version I would use for export of high poly to Blender. Before export, I would decimate the entire model to 25% with Decimate Master and get down to around 10 million polygons which were manageable for Blender and Substance Painter to import.
On the other file, I usually export the low poly parts of the model that I will not retopologize by hand.
From the entire model was just the grip that I was decimating to the point where it was low enough but still retained all the shapes. We have to remember that in the era of nanite we can leave a bit more topology as long as it’s not too heavy in MB and is possible to unwrap easily.
Low poly
Low poly was actually a super fast process. The parts I dynameshed in Zbrush I decimated previously so they were serving the purpose of my low poly (purple on the screenshot). The rest of the model was basically optimizing my model pre-ZBrush export (gray on screenshot).
Some parts that were dynameshed but didn’t go through heavy topological changes I recreated from an old version of the model with places where I welded the parts (Pink on screenshot). After that, I did smooth groups and went on to get cracking on the UVs.
UVs
For the UVs, I use all the basic principles that make a good UV. I’m checking the texel density and lack of errors with a checker. For packing I used UvPackmaster.
Baking
For baking maps, I’m using a few approaches from the easiest to the most complicated. For first one that actually worked for this project is simple settings like on the screenshot. I’m tweaking the values a bit and that is pretty much it.
If it gives me glitches I’m creating a custom cage in Blender. If that for some reason still doesnt work I’m moving to marmoset and hand-painting the rays of baking. For many flat surface projecton I will straight up use marmoset baking tools because they are so flexible
Texturing
The texturing for this model was an incredibly complex task. It’s impossible to show the entire process in a breakdown but I can offer a few tricks.
The Patterns
The first trick was to prepare the masks for the carvings in photoshop. For that, i used this tutorial and made my little changes and hand painting.
Metal Material
1. For the base material I used a blend of rough steels from the bridge and mixed them with masks to break them apart.
2. I wanted a lot of coloration so yellowish rust that didn’t have roughness changes one was off the steps.
3. I used nice rust combined from 4 different rusts from the bridge. I masked it out according to how the parts were used and somewhere I used just rust spots and AO-based masks.
4. On top of everything I used a ton of subtle colorations to break the colors apart.
5. Apart from that I used all the other tricks, like shiny edges and dust in corners.
Wood Material
1. Important for old woods is a blackened area near the place where it connects to metal. It gives it a well-grounded look of a used weapon.
2. For the carvings I masked them out by AO and hand masking. I’m adding a Voronoi texture that uses a height channel and gives a look like it was carved by hand.
3. The base wood is a layer of rough wood with 3 layers of colored wood mixed together. Where the woods on top are masked we have nice scratchings to rough wood. Apart from that, there are the usual suspects. Scratches, dots and color variations.
Rendering
For rendering, I used Marmoset to show that it looks good in PBR, Real-time and RTX workflows.
I’m not going to write much about it apart from key rules. I used the classical 3 light setups: Key, Fill and Rim. As for the rest of the key features, you will see them in the screenshot below.
Closing words
First for some general tips!
– Watch your references all the time
– Change a considerable amount to make your own spin on the asset that could improve it.
– Get lots of feedback from people that you would want to learn from if possible. Don’t post on general facebook groups because all you will get is likes and no reasonable critics.
If anybody that reads this article will want to make a personal project please find a reference of something that you really like visually so the time you spend on the work actually pays off.
Try to imagine if the asset will look pleasing at the end so your personal time is not wasted. I really do hope you got to learn at least one thing from this breakdown. Thanks for reading and keep on improving!
If you have any further questions feel free to dm me on artstation. Also, I would want to extend my personal thanks to all the good peeps at my workplace BlooberTeam in Kraków. You guys give me the chance to improve every single day. Thanks