LogInEric Correia shares a detailed breakdown of the stages of his shotgun drum magazine project, providing insights into hard-surface modeling and texturing of realistic colored and refractive materials.
Introduction
Hello, my name is Eric Correia, and I am a Senior/Lead 3D Artist. I picked up 3D ten years ago and have been in the industry for about seven. I Initially started my career by working outsource freelance with contractors like Dekogon (Season 1+2). I discovered that it wasn't entirely for me, as I enjoy working closely with a team in production rather than collaboration. So, I found my way contributing to No More Room in Hell 2, which was initially a long-running passion project. Shortly after, this led to my first AAA gig working on Far Cry 6 at Ubisoft Toronto. More recently, I've been a Senior/Lead Artist at LuckyVR for the past four years, contributing to our game Vegas Infinite.
Throughout those years, I've always enjoyed mentoring, teaching, and especially discovering tips and tricks, so when I completed this mini-project and was offered the opportunity to write about my process, it was an easy decision. I've always been an enormous FPS genre enjoyer, along with shooting and handling actual firearms, which leaves little doubt about what led me to specialize in weapon art. It will almost invariably pose new challenges, and naturally, it is a very technical specialization, both things that I enjoy involving myself with. And, of course, guns are cool as hell. This platitude has always stuck with me since a young age.
Getting Started
I started this project because of my love of shotguns – their power, intimidation factor, many different platforms, the variety and flexibility of their rounds, and many usage scenarios. I even love the way ejected shells sound when they hit the floor. I wanted to make something that invokes the power and intimidation that a shotgun platform typically would. Due to the size of this entire project, this magazine would naturally serve as a sub-project, an element of a larger firearm in the future. Working closely with many similarly sized 12 gauge drum magazine references and polymer references helped me a lot, along with the Defenceport videos and image references.
Modeling
The modeling process was straightforward. I used 3ds Max and Plasticity. I prefer to block things out in polygons first since I find it much faster and less restrictive than it would be in CAD. The general challenge here was retaining accuracy while also getting more familiar and comfortable with direct modeling in CAD.
Blockout is just as important as reference gathering. Ask yourself if you have more information that isn't just visual reference, such as platform type, magazine size (single stack, double stack), caliber used, etc. These will all help you find dimensions and other useful information to get an accurate block out.
If you know your gun shoots 9mm, look up 9mm bullet dimensions. This will help you build your magazine, which will then assist you in developing your mag well or chamber and, lastly, your barrel. Once this is established, all other exterior components should fit in much more easily.
I tend to call this the “skeleton” of the blockout since it is objective and can be used to compare against gathered references. It certainly surpasses estimating everything and managing perspective distortion.
Manufacturing websites are dual-opportunity: they will often provide “known” information, such as the size and dimensions of the asset, as well as the high-resolution reference you were searching for.
Once the blockout was complete, I headed into Plasticity. Using the blockout, I overlaid new shapes and modeled the asset with a high-low detail mindset. I advise modeling only large shapes and forms first, then moving on to medium-sized, and finally, fine details. This will prevent issues and revisiting and make the workflow more fluid.
If you are having trouble identifying these shapes, take some time to colorize them in a few reference images to help break down and train your brain on proportions and shape language. It's natural and almost always expected to revisit a detail after realizing something doesn't fit or look right after modeling some other details.
I used to use SubD Modeling primarily, and I still follow it when I see an opportunity arise. SubD is a very useful fundamental that every artist should be familiar with. For example, it's much simpler to create the shells using SubD, especially since they are largely cylindrical.
The high poly stage was handled by importing and applying Dynamesh to the exported OBJ from Plasticity in ZBrush. It's crucial to have as much density as comfortably possible on your export to prevent any sort of faceting.
It's possible for some areas of your export to have some noticeable faceting still. I've come up with a solution in ZBrush. We can assign polygroups to all of our “hard edges”, then crease by polygroup and after that, subdivide them a few times. This will remove the faceting while nearly perfectly preserving our shapes, ready for Dynamesh and polish. The graphic below should explain and illustrate this better.
Using Morph states and Morph brushes to bring back different stages of polish is handy. I've found this to be faster and easier than masking edges out. For example, I could polish everything to a certain amount of sharpness, save that Morph state, polish some more to get a softer edge highlight, and then finally use a Morph brush to paint back the previous state of sharper edges.
After applying DynaMesh and polishing, you can add sculpted or embossing details for baking. If your asset isn't new, consider adding damage and wear pass and beating up the edges to get irregularity and worn edges to be baked down.
Details like text and finer details are typically best handled in Substance 3D Painter. However, I realized that the logo and text details were large enough to warrant being in high poly.
Topology & Unwrapping
The low poly process with CAD is its weakest area. Using MOI, or in my case, PiXYZ, to process the exported .stl/.step files into a much better starting base makes this stage much easier. Check out this graphic below demonstrating the topology and amount of clean-up work each exporter produces.
Here are the basic steps I run when creating my final LP after it is processed from PiXYZ. The first thing is analyzing and cleaning up errant vertices. There are not numerous, but they still occur. I constantly research or explore tools that make this easier in these cases. I found this setting in 3ds Max to select all vertices that are less than 3-edge valence, which is a vertice sitting on an edge with nothing else connected to it.
Next, I would remove topology that I know would either hinder future steps or be unnecessary. At this point, I suggest evening-out edge distances, especially for curved/organic areas like grips. Not only does this keep the shading nice and consistent, but it also helps keep your geometry density under control, not to mention it also offers a nice and tidy topology.
For larger forms and shapes, I analyze the silhouette and judge the fidelity of my model, piece by piece, frequently from many angles, to decide if I need to add or remove loops. Context is important here. If parts of the model are obscured or obstructed, then it can receive less geometry.
Despite the fact that this work is only for my portfolio, I still wanted to be sensible and demonstrate control of a budget. I classify this as “art muscle memory”: the rules and boundaries that I normally set in production scenarios. I wouldn't want to undo that, even periodically, and become comfortable with limitless production for the sake of my portfolio. I personally feel much more accomplished developing high-fidelity art with most of the typical limitations and boundaries in place – it feels much more rewarding and validating, making things look great with what you've got.
Anything can look great with several 4K texture sets slapped onto it, skyrocketing its texel density... But is that ever acceptable in production? Never.
The molded polymer rib details are a great example of stretching this boundary while still remaining practical. In most situations, these ribs would be baked down. However, this asset is large and could be close to the players perspective, so that warrants modeling them in. The first problem encountered with this is that there would be many hard edges and UV islands, which would also mean many texture seams that could be problematic for my masks.
We can remove all of the hard edges to have one big contiguous UV island, which is perfect. However, removing the hard edges would mean the shading would be completely unacceptable. Leveraging face-weighted normals along with some additional edges to control the shading brings this to an acceptable stage.
When the geometry placement and fidelity are at a place that I'm content with, I proceed to analyze the shading on all parts of the model. I try to do this as I go through all of the mentioned steps, but it's always good to do this as a final sweep at the end. As I'm analyzing for any strange shading, I'm also looking to see if I can remove some hard edges without introducing shading issues, as this would mean I could have fewer UV/texture seams, which is very beneficial to my masks.
It is also important to ensure solid shading and hard edge placement at this stage. Leveraging the hard edges can give you a jump start on your UVs immediately, abiding by the golden UV rule below.
Texturing
Off to Substance 3D Painter! Material definition is vital, especially since this asset is primarily polymer. I spent a good amount of time assessing and revisiting my values to ensure the polymer reads well from a good range. I tend to treat materials in the exact same manner as modeling – starting from large (macro) to medium and fine (micro) details.
Study your reference, and don't be afraid to include references to other weapons/components/subject matter – as long as the details read well and are interesting enough, include them.
Before starting, I always make sure to texture with ACES, a neutral HDR, not to influence any sort of colored lighting like Tomoco Studio and other viewport settings. This ensures that my visuals and renders will be a matter of WYSIWYG (what you see is what you get).
I advise building up layers similar to how they would in real life, starting from base material, tone, and color variation, gloss variation, proceeding to scuffs, scrapes, and colors rubbed off from other materials (it gives interesting micro-sized color accents), as well as dirt or dust build-up in cavities, grease, oil, smudges, leaks, and stains. The final step is to add some heavier damage if you consider it necessary, and global dust, specks, or fibers.
Try to be deliberate and intentional with your details and utilize stencils where possible. Distinct details like large wear should not be only one stage or one fill layer. Think of it as multiple layers built up. The scraped polymer likely had much lighter, fainter scrapes before it was deeply scratched – be sure to include those details to ground it and make it feel much more natural.
Keep things as procedural as possible, but intend to paint them in with multiple layers. This helps break up the natural uniformity of procedural masks while keeping them even further procedural by having a non-destructive paint layer set to multiply. Try to avoid tunnel vision – analyze your details from many distances and angles, take frequent breaks, and get second opinions.
Rendering, Lighting & Post-Processing
It has been a while since I've lit and rendered an asset, so it was a bit of a challenge requiring many revisits and rehauls. I've always understood basic lighting principles and set-ups, but getting your hands dirty, messing around with angles and compositions, and just trying things out was, while frustrating at times, quite rewarding. It absolutely helps to have proper settings first.
I completed rendering in ACES, the same color space used for texturing. I advise using Ray-Tracing along with Denoising. Be careful not to crank de-noising, as it can reduce and blur out the finer details in your renders. Also, try to keep the perspective lens to a minimum to capture the entirety of the asset without any depth or perspective distortions – about 200-300mm. Real photography would use these lenses anyway.
Some presentation styles could call for a lower mm/more perspective. It is entirely up to you, but presentations often shine with a high mm lens. I largely used Directional Lights to achieve my Key and Rim Lights. Sometimes I would use the HDR lighting at a very faint value if I needed to fill up some areas. Omni lights or spotlights could be used sparingly, but they certainly help accent certain areas that the Directional Lights are not hitting – be sure to try your best to blend in with the other lights as they stick out a bit and unintentionally light up other areas.
Use the Safe Frames feature under your camera settings, as this greatly helps set up proper composition and ensures nothing would be cut off from your render since it's quite misleading to eyeball it from a regular viewport.
At post-processing, try to keep things as natural as possible. It is easy to crank up some sliders, render, and post-process additionally in Photoshop, resulting in renders that look “fake” and over-processed. All of your hard work is essentially being tarnished. The values that were most impactful were exposure, contrast, sharpen, and clarity – all dialed in with careful control.
I exported the render as a PSD file since this would not affect quality. It allows me to further add adjustment layers non-destructively if I ever need to revisit and tweak a setting and to save out file formats properly for publishing.
Challenges & Goals In The Project
I had to make sure that I wasn't producing this asset in my comfortable and familiar pipeline. Instead, I was certain to present myself with some new goals and challenges to strongly ensure maximum gains that include the following: first of all, learning spec/gloss texturing. I've always used metal/rough workflow, and still do while at work, but I couldn't help but notice the benefits specular-gloss offers when in the right hands. Many weapon artists also adopted this workflow.
Secondly, a deeper understanding of CAD/Direct Modeling. Initially, I was comfortable with parametric and non-destructive CAD modeling such as Fusion 360, but Plasticity seemed really attractive to leverage and learn, especially with this project being on the simpler side. Plus, any excuse to get away from Autodesk's enterprise and cloud is a win for me. Plasticity is a direct modeler, not parametric at all, as one would initially think. If you are used to parametric modeling, you have to approach things a bit differently than what you're used to.
Next, I aimed at achieving an attractive texel density. This was especially challenging due to the unforgiving large surface areas of this magazine. It's easy to make things look pretty by splitting it up into several 4K texture sets, but part of the enjoyment for me is being conscious and sensible with portfolio work. Demonstrating that your high-fidelity visuals would translate perfectly well into a real-time production environment is a feat.
Also, I focused on texturing colored or tan polymer surfaces. This might sound silly, but I've found that colored materials, especially polymers, are much more difficult to balance than darker or black polymers.
Ever wipe your hands on your pants while eating on the go? Usually, one would if their pants are dark, but (hopefully) never on lighter-coloured ones. My point is that darker/black materials generally hide and subdue details quite well, so choosing a brighter, more prominent color means that I have to be delicate and deliberate with my applications.
Additionally, I aimed at texturing and rendering refractive materials, such as the lid. I always love seeing when weapon artists render transparent components, especially magazines. A lot of references had opaque lids, but I was certain to keep that transparent detail (and challenge).
Lastly, I would say that the main key to creating appealing art comes back down to the fundamentals. Regardless of the subject matter and how great the modeling and texturing are – lighting and composition will ruin the attractiveness of all your hard work. Learn the fundamentals of lighting, composition, and maybe a little bit of color theory/tone to understand contrast for more balanced and attractive renders.
Conclusion
If you've come this far, a sincere thanks for reading this! I hope you were able to take away something useful and informative from this write-up! If you found this helpful, please check out my blog on ArtStation, which continues to grow. I hope you will come across something informative and helpful there as well!
One more piece of advice: if you want to grow as an artist, be sure to involve yourself with like-minded communities. The Weapon Room is a fantastic example of a great community of artists with professional feedback for hard-surface art (though all artists are welcome, of course). I never thought I'd ever be a mod, but I've turned to the dark side. Hope to see you there!
Thank you, 80 Level, for this opportunity! I enjoyed putting this all together, hoping to offer some information that has yet to be touched on. Best!