3d artist Kit Grande talked about the way he sculpts 3d guns and creates skins for them.
3d artist Kit Grande talked about the way he sculpts 3d guns and creates skins for them.
Introduction
I’m currently working at Dambuster Studios as a Junior Environment Artist in Nottingham, UK. I took Art and Design Technology all the way through school and spent pretty much all of the time I could either in the studio and workshop or playing CS 1.6/CSS, Quake III. After a pretty tough (but correct!) decision to pursue Art instead of Classical Civilization/Anthropology I went to Arts University College Bournemouth to do my Art Foundation year. From there I went onto De Montfort University to study Game Art with an incredible bunch of artists who are still constantly inspiring and pushing me.
For a few years I tried to make things for Quake, Morrowind, CS 1.6 but I didn’t get very far and then got distracted. When I returned to 3D at DMU I quickly found my love for the combination of art and technology. I was very fortunate to spend 6 months during my studies working as a Marketing Art Intern at Ubisoft Montpellier (taking marketing screenshots for Assassin’s Creed Unity: Dead Kings). While it wasn’t exactly the work I wanted to be doing at the time, I learnt a huge amount during that time and I was driven to focus on 3D in the evenings. After graduating I started working at Dambuster Studios on the last stage of development for Homefront: The Revolution.
Modeling Weapons
When I joined Dambuster I was asked to make a baseball bat for a cinematic. From the work I did on that I was then asked to work with Pete Cassell, our Lead Weapon and Vehicle artist, to help polish, bug fix and finish the weapons and vehicles for Homefront. I was really lucky as a recent hire to get this opportunity but during that experience I understood that this was an area that I wanted to push myself towards.
There are many different reasons that I really like working on weapons. On a visual level I’m very excited by technical, machined, hard surface and functional objects; there’s something very satisfying about precisely manufactured moving parts. On a game level I have always played FPS games and in an FPS the weapon is the tool that allows you to interact with the game so it holds a great deal of importance in the gameplay. On an artistic level I love the amount of character, personality and story you can put into an object, this is something that I’m eager to improve and develop to the point that you’d be able to tell the personality of the owner just by looking at the object. I can also get a bit too fixated on details sometimes and that’s a trait that can really pay off when making weapons as every last millimeter of it is scrutinized.
Peculiarities of Weapon Production
Every field has its own peculiarities and everyone approaches things differently but one of the big things I’ve found about weapons is the focus on modeling your materials and production methods. As an example machined metal has a very tight edge to it and stamped metal tends to have a softer edge on the stamped details.With this knowledge you can communicate the production methods to the viewer just with your modeling. This is a small detail but lots of people pick up on it (consciously or not) and it will really help sell your materials.
It’s important to know your brief or requirements when starting something like a gun because there is definitely stuff you can get away without modeling and requirements vary. Breaking down what you need out of the model at the end and breaking down how the gun is constructed really helps to give you an understanding of which corners you might be able to cut. For Homefront we had a game mechanic that allowed you to swap a section of the gun out to create different homebrew weapons which meant that certain areas which may not normally be needed had to be modeled. For my Beretta 92 I basically just wanted to do a study, I wanted to improve my understanding of the functions and components. I set myself the challenge of having a torn down render with all the parts as separate pieces.
Production
Primary step as with all art is research and reference (it’s a skill in itself). For real weapons this can be especially important as there will always be someone out there that will point out the details you missed… Oops.
The next step I go to is block out, it’s such an important stage as it’s the time that you get your scale, proportion and large forms locked down. It’s the same with all artwork you should work large shapes -> small shapes. After that it’s just a case of using the right tools for the job. I used 4 different modelling techniques for the high poly of this project; Boolean-Dynamesh, standard SubD, double smooth and sculpting. They all had their uses and reasons.
Boolean-Dynamesh is very good for creating complex machined parts as it mimics the subtractive nature of machining processes. The edges will end up consistent, the workflow is quick, somewhat non-destructive as the operands remain editable and allows you to avoid tweaking topology endlessly. Both Ubisoft Red Storm’s presentation at Zbrush summit here and Ben Bolton’s more recent breakdown of the process on Polycount here are really valuable resources for understanding the method.
Tip: Make sure any curved forms have a high segment count on the curve at this stage because later on the dynamesh will maintain the form of your underlying geometry and if this isn’t smoothed enough you’ll end up with some nasty artifacts. E.g. I didn’t manage to catch it on this area of the high poly unfortunately so there are visible lines of tighter geometry.
Standard SubD by comparison can sometimes be a bit quicker and gives you the flexibility of control over each edge width.
Double smooth can sometimes be a bit problematic but is a quick way to model simpler parts. The theory being that you use the “Separate by: Smoothing Groups” to control the first turbosmooth modifier with a couple of iterations and then add a second on top to round off the edges. In reality I should probably have used QuadChamfer for this.
Sculpting gives you access to much softer surfaces and high resolution detail, the custom grips being an example of sculpted surfaces and the stamped writing on the slide being an example of higher resolution detail.
My aim with this project was to have all the details in the high poly without the need to add anything to the normal map information later.
Texturing
I personally like Substance tools for texturing but there are many alternatives. The main thing I like to think about when texturing (and modeling) is, how is this actually constructed in the real world? Make sure your layer stack makes sense; bare metal shouldn’t be on top of paint, you should be scraping paint away to reveal the bare metal. It’s a small thing but it quickly has an effect on your mindset. I personally think it’s really important to understand the actual methods you would use if you were creating this model as a physical object.
Substance Painter is an environment that I find really allows you to play and have fun. I’m known to end up making stupid, tasteless versions of my work all the time just because I have a mad 5 minutes and want to see what will happen when I do this or that. The results are normally quite garish, colorful and silly but they do result in me learning techniques to achieve different effects so if in doubt have a silly 5 minutes every now and then.
In Substance Designer one of the big things I’ve found is if in doubt return to your ref, it can be easy to get stuck looking at your result and not see major differences between what you have and what it should be. Another big thing is knowing what scale you’re creating the material for, is it 2meters x 2meters or is it 25cm x 25cm? This has a huge impact on the end result looking correct. When creating materials to use in Painter it’s tempting to EXPOSE ALL THE PARAMETERS but you want your material to look good and be easy to use. Limiting what the user can do can really pay off in making a solid customizable material that consistently looks good and gives the desired result.
Experiments with Materials
Weapons are great for experimentation but can often be quite a stale or dry subject matter as the focus is often on accuracy to the stock version of a weapon. I really enjoy playing with customizing guns and trying to give them a story, personality and character. Ultimately I would love the viewer to be able to tell who the owner is from the weapon. I really like skins that don’t break the believability of the rest of the game. The variant should make sense, be physically/realistically viable and memorable. Think of different materials, different patterns, completely changing certain elements can give you a great effect.
If you do go for a paint or vinyl type of skin, believable placement and wear is really important but can be easy to overlook because it’s “a skin”. Having a reason or motive for making a skin can be really helpful in giving you a direction other than just looking cool or different. For example, does the owner particularly like one colour?…
In respect to experimentation, have a look online see what actual gunsmiths are doing, have a look at how people in the US are customising/personalising their non-tactical weapons. Woot/Damascus steel is something I find fascinating and have wanted to work with in real life and in games. I started creating some variations of Damascus in Substance Designer so that I could then use them to texture in Substance Painter. Initially it was just a fun lunch time project but I saw an opportunity to use some of them to do a skin of my Beretta. I’m not particularly pleased with how it came out and there is lots that I would do differently next time. For example; if this were used in a game the blued Damascus steel grip would barely ever be seen and so the amount of work that went into it wouldn’t have been a great use of time.
Using Toolbag
I really enjoy working in Toolbag 3 as I find it’s a really versatile and quick tool to create pretty things. Possibly the biggest improvement for me is the baker, nothing else out there has quite got to this level of hands-on baking. As baking is often a slightly black-box-y thing where you don’t know why things aren’t baking properly or what’s actually going on I really like the control, hands-on feeling and realtime update Toolbag 3 gives you. Painting your skew map right onto the model and seeing your flat details just pop to how you want them will never get old. Other than that Toolbag has always been really nice for rendering micro-environments (which I seem to do a lot) as the shadow resolution is by default higher than most game engines because it’s focused on rendering smaller things rather than entire worlds. Also now in Toolbag 3 the global illumination makes a big difference.
Lighting and Composition
As always, reference is key. I troll through Instagram, Pinterest etc finding images of workspaces, small objects that I thought were lit really nicely. I have also spent a lot of time lighting and photographing objects and illustrations that I’ve made over the years and this practice really helps as the theory is the same in the real world and digital world.
One of the things I like to do is make sure that I have as simple a lighting setup as possible. In this scene I have one key light which is emulating the lamp on my desk that I use when working non-digitally. The blue lights are to emulate the LED strips I have under my monitors and give a subtle glow from off camera. And a fill light to just subtly create a reflection to show up some of the edges on the asset, by using an omni light you can easily find a position relative to the camera so that you can get those highlights where you want them. Lastly just using the Sky from Marmoset with a HDRI in which makes sense for the setting at a low brightness. This is all simple product photography for which there are many tutorials online.
Compositionally it’s really important to have a solid focal point which holds the viewer’s attention and lighting that enhances this composition. When making a cluttered image that looks like a real person’s workspace, clutter or unrelated objects which flow around the photo leading to the focal point are a really nice thing to include (which I need to do more of). Again compositionally you should be working large forms to small forms.
The only thing I did differently in this render is to have something slightly blocking the lamp (I just used the handle from the case) it can be quite cool for getting the DIY style and add a less perfect quality to the light.
I’m only at the start of this journey though and only really have a few months production experience with weapons so I still have tonnes to learn.