Creating Sci-Fi-Inspired Japanese Knife With Plasticity

Introduction

I'm Dylan Mellott, I've been working as a professional 3D Artist for 10 years, with 9 of those being solely dedicated to 3D hard-surface art specifically as of the time of this article. I've been lucky to have been able to focus nearly entirely on first-person weapons during that time. That said, a quick scroll through my portfolio will reveal that I've worked across a variety of styles in my career. I similarly consider myself very fortunate for these opportunities, as they really helped me dial in the kind of artist I wanted to be and the things I like to work on as the years went on.

For starters, I cut my teeth on a free-to-play hero shooter called Paladins.

After which I worked at DICE LA, where I helped ship Battlefield 1 DLC and Battlefield V:

I eventually set out on my own as a freelancer, where I was fortunate enough to work on several projects, including weapons for a stylized VR game:

I would later return to DICE LA to help ship Battlefield 2042:

Not long after shipping Battlefield 2042, I joined Bungie. Destiny 2: The Witch Queen had just launched, and the very same morning I was due to get my employment paperwork, it was announced that Sony had bought Bungie. Needless to say, that was an exciting time. I've been at Bungie ever since then, where I've helped ship multiple expansions and seasons for Destiny 2.

The Sci-Fi Tanto Project

OK! Now with that boring stuff out of the way, let's get to what brought you to this article. 

My friends, the paradigm has changed in the hard-surface 3D world. I know, I know, I'm sure you’ve heard that before when a flashy new software or technique drops, but this time? It's the real deal. 

Plasticity and Marmoset Toolbag 5 have shaken up the formula to such a degree that, in my opinion, it cannot be ignored. The CAD-like functionality of Plasticity, combined with the Bevel Shader of Marmoset Toolbag 5, means that not only can we model at incredible speed and detail, but we can take that mesh from Plasticity, toss it into Marmoset Toolbag 5, and instantly get smoothed edges and surfaces. Some artists have been using similar workflows for a long time, but never before has it been so accessible and easy. 

For Lookdev, Concept, and especially 3D Game Artists, this is a huge deal. Because it means we can reach those final results much faster and without having to fight against polygons along the way. For 3D Game Artists, we can even bake that bevel-shaded mesh down to our optimized low-poly, and the power of that cannot be understated.

I started the Sci-fi Tanto project between some other larger ones. Honestly, I was just perusing ArtStation when I saw a concept by Adam Lee and decided to take a short break from my other projects and do something smaller. I wanted something that I could take from start to finish in just a few hours, so this was the perfect digression.

For the Tanto, I wanted to try a few things. First, I wanted to further reinforce a workflow I've been practicing more on the lookdev/conceptual side of things, where I focus on getting that bevel-shaded, textured, and rendered result as quickly as possible. Basically, getting to the final, complete, and presentable idea very quickly. I've spent my entire career making optimized, performant, engine-ready 3D artwork, so for my personal pieces, I sometimes like to throw that all to the wayside and just focus on the fun artsy stuff. I first tried this workflow on my Cyberpunk Shotty piece and was fairly happy with the results and process.

I wanted to try it again here with the Tanto. Additionally, I wanted to dig a little deeper into texturing in Marmoset Toolbag than I had on the Shotty.

Depending on the goal, I may gather a bunch of reference images and concepts and merge various elements of them together, but for this instance, I just plucked different things from that single concept sheet that Adam Lee created. Namely, combining various elements of C and E together.

Modeling

The model was done entirely in Plasticity. It's basically where I'm doing the vast majority of my shape-making these days, unless the project or goal would require using something else. What makes Plasticity so powerful is how flexible it is. With that flexibility comes speed and the freedom to make changes without baggage. 

Being able to change, add, remove details, and even major forms with little effort cannot be overstated. Trying to do things like this via a polygon modeling workflow would be a massive, massive headache. Plasticity gives us a ton of freedom and saves us a ton of time by removing polygons from the modeling and shape-making equation. 

The blade and handle are, for sure, the most complicated parts of what is an otherwise fairly simple model. Let's take a look at the approximate approach for them individually.

Blade

The vast majority of what will eventually become the mesh starts out as Lines, which are kind of like Edges in other software. In the example below, I sketch out the rough shape that I need. It will automatically become a Sheet when I've connected the Lines, as indicated by the blue plane that fills in once all the lines connect. This indicates that I can extrude from this shape, creating a Solid. When you see a purple line appear, I'm hitting Shift on my keyboard, which draws some construction lines that will help you accurately place your clicks.

After I Extrude and get my rough volume, from there it's just a few simple Chamfers and a Fillet to make the blade edge.

Adding details from there is also quick and easy. I use a tool in Plasticity called Imprint Curve. It lets you take any Line or Edge and project it onto a Solid. Here I use the Line, Regular Polygon, and Center Circle tools. Watch how quick and easy it is to add not only larger changes, but details as well. To me, as someone who's spent so long polygon modeling, this is pure magic.

Here are some examples of how I did various details on the Blade:

To get the blade groove, I use Center Circle, then I connect two Lines to the outer edge of the circle. This automatically gets our Sheet for the groove. I then use Trim to remove the inner edges of those Center Circles. Lastly, I select all the edges and use Join to combine them into one object. From there, it's the same Imprint Curve workflow that we used above. Same idea for the plate that I extrude in the second GIF. You can see how quick it is to both build and detail shapes.

Handle

The handle is definitely the most complicated element of the Tanto. That said, I approach it in the same fundamental way, by getting the silhouette and rough dimensions using Lines and Edges. You can see approximately how I approached the handle in the GIFs below:

This next feature is a favorite of mine: the Offset tool. I use it for all sorts of scenarios, but the main idea is that it linearly scales your selection inward or outward, ensuring perfectly spaced distance. In this case, we're going to use it to get the thickness of our handle. After which we're going to use a tool called Loft, which attempts to maintain a contiguous, congruent surface between two Lines and/or Edges. 

It's important to note that you can delete control points on your edges or move them around at any point here. You can also Chamfer or Fillet them!

And from there, we take what is currently our Sheet (AKA geometry that is not currently watertight or Solid) and we make it a Solid by using the Patch Hole tool. From there, we bevel some of our edges to get more interesting shapes. Super cool!

Topology

Topology and the unwrap are complete afterthoughts for this workflow, which is why I like it so much for personal projects.

But OK, jokes aside, here's how I approach it for real. But we're going to look at two different methods: The Lookdev/Concept-y way that I used for the Tanto and most personal projects, and the Real-time Production way that I use professionally in my 3D Game Art career.

What's the difference between the two?

I use the Lookdev Method for:

• Rapid prototyping and concepting;
• When the goal is to reach a polished, presentable idea as quickly as possible;
• When final optimization isn't critical.

I use the Real-Time Method for:

• When assets require specific polygon and memory budgets;
• Professional 3D Game Art production pipelines;
• When final optimization is critical.

Regardless of which method, both use an essential part of this workflow: Blender and the Blender Bridge Add-on.

Blender Bridge does what the name implies, really, it gives us a bridge from Plasticity to Blender. But more than that, it comes with some pretty fantastic tools to get to your low poly very quickly. We'll cover that in the Real-Time Production Method section.

The Lookdev Method

Follow along with the Blender Bridge installation instructions, have Plasticity and Blender both open, and then import your model using the add-on.

Honestly, if all you care about is getting to the Bevel Shading, Materials, Texturing, and rendering from here, then you can just keep your Blender Bridge settings default and start slapping on UVs. I'll cover UVs more in depth in the Real-Time Production section, but if lookdev, concepting, and speed are the goal, you can pretty much UV Smart Project and move on.

Now, you could take this and export it straight to Marmoset Toolbag 5 and immediately apply your Bevel Shader and start slapping on some materials, and you'll be good to go. But, I do like to squeeze a little extra juice out where possible, so I used a tool called the UVPackmaster 3 add-on to help me optimize the usage of the UV space. I tick Normalize so that all of my UVs are the same texel density (basically a measure of resolution) and Heuristic Search to let the add-on have a few chances to solve for the best layout of the space. I select all but the left-most UVs and hit Pack To Others for a specific reason. It's because I know that the left-most UVs take up about as much UV space as they are ultimately going to, but all the other little parts have plenty of room to move around. By pushing Pack To Others, my selection will respect existing objects in the UV panel that are not selected. Very handy.

From here, you can simply export your FBX to Marmoset Toolbag 5. I'll cover Shading and Materials in a later section.

The Real-Time Production Method

If your goal is an optimized deliverable that you could use in a professional production environment, then, of course, things are a bit more involved. Thankfully, though, Blender Bridge gives us a fantastic and speedy head start on optimizing and low-polying our Plasticity mesh. This is where we'll work on converting our Plasticity mesh to usable geometry. But first, we need to do some preparation for Baking later. You can do this step in either Blender or Plasticity, but since this model is so simple with so few parts, I'm just going to do it here in Plasticity.

All I've done here is duped all my parts and grouped them, one group for the High and one for the Low. Each subpart within the folder needs to be named the same as its corresponding part, with the suffix denoting which group it belongs to. If you've ever Baked By Mesh Name before, you're already familiar with this process.

Then, I begin stripping off elements I intend to bake down from the High to the objects in my Low folder.

After stripping off elements I intend to bake down, The Low might look something like this on the right. Now, we need to head over to Blender.

Follow along with the Blender Bridge installation instructions, and have Plasticity and Blender both open.

Import using the Refresh button in our Blender Bridge. Plasticity graciously retains our naming and folder structure. We can see that reflected here when we use the Blender Bridge to import our content from Plasticity. I then hide the High folder, as that is the mesh we're going to apply the Bevel Shader onto in Marmoset Toolbag, so we don't need it for a while from here.

So when we import to Blender using the Blender Bridge, our topology is going to look like this. It comes in already triangulated, which is definitely not ideal at this stage. We will eventually want our mesh triangulated when we export for Baking so that we maintain tangency between our mesh and our game engine (a whole other entire topic), but at this point in time, it's more of a hindrance. If only there were a convenient way to get from triangles to workable geometry from here...

Let's change from Tris to Ngons, click a given piece, and hit Refacet. Not too bad for two buttons, eh? But we can see an obvious problem: it's reduced the edges of our curved handle too much, causing it to lose that nice smooth transition. Let's look at fixing that.

Tolerance and Angle are our two major means to adjust things easily. The general idea is: the lower the number, the more Refacet will closely mirror the shape of the original mesh. The tradeoff, however, is that we retain more edges and vertices, but that's OK in this case. By moving these sliders around, we can get somewhere more ideal for our needs. Watch as I flip through various settings and see how the mesh changes.

Now we're talking! I cannot understate how powerful this is. We can literally forgo polygons during the modeling and shape-making process, freeing us up from the trappings of that workflow, but still get a very quick and completely usable low-poly mesh in almost no time using Refacet. While it can struggle with more organic shapes or complex shapes and may require manual connecting of edges or rebuilding of faces, this is absolutely incredible.

Let's look at an example where it doesn't give us such straightforward results:

You can see here that it’s not done as good of a job solving this blended surface, but that's no problem. Here, I'm selecting all the faces I want to remove the edges from using a combination of drag select and manual selection. Then, I use something called the F2 Add-on. It should be native in most Blender versions. It's a bit of a Swiss Army Knife and does many helpful things (too many to cover here), but what I'm using it for in this example is to delete all the Edges within a Selection of Faces. Fast and helpful, right? Well, there's a catch... It leaves behind all the vertices, which is far from ideal.

However, you may have guessed, we have a great way to deal with this in Blender. I use an add-on called MACHIN3: Clean Up that I've bound to a hotkey. This add-on is incredible and honestly might be dark sorcery. It can be used to remove all sorts of elements, but I mostly use it for vertex cleanup as shown. Awesome.

Depending on your mesh target resolution and polygon budget, your low-poly may look something like this. Now, we have just a few more checkboxes before moving on to UVs.

Plasticity can sometimes apply its own Custom Split Normals to objects during the bridge process. This is a problem because it will interfere with the normals we intend to bake to our game mesh later on, among other potential issues. We can solve this easily by clicking the Data tab, selecting our objects, and hitting Clear Custom Split Normals Data.

Next, because our model is symmetrical, we can safely delete the half that we intend to mirror. I'm using a MESHMachine add-on tool here called MACHIN3: Symmetrize. It both deletes half of the mesh and can symmetrize it back later. Very handy.

Lastly, you may have noticed that when I'm in Edit mode, some of my edges appear blue or red. Blue indicates Creases (which will come from hard edges in Plasticity during the import process) while Red indicates Hard Edges (which I have been setting myself via hotkey in Blender as I did my low-poly prep). We're just going to clear them all to give us a fresh starting point. Our mesh normals are going to have broken shading, but we're going to start manually assigning our hard edges and UV splits, which will correct the shading and normals.

OK, phew... now that we've done all of this stuff, we can talk about the UV process for this method. We're going to tick a box for a tool called Live Unwrap. What this tool does is, as we begin to mark hard edges, it's going to automatically start laying out and placing our UV islands into the UV panel. Watch below:

You may be wondering how I'm selecting multiple faces at once here without moving my mouse or performing additional clicks. I'm using a tool called Select Similar. It's set to Coplanar, Equal, with a low Threshold. What this does when pressed is it scans your entire active selection for planes that are on basically the exact same angles. This allows us to scoop up tons of geo in one button press for a very fast UV hard edge markup phase.

After you perform your Hard Edge Markup with Live UV phase, your results may look something like this.

You can turn off Live Unwrap now because we are going to manually edit our UVs from here. If you accidentally mark a hard edge with it still ticked on, it'll undo all of your upcoming manual work. 

Note: This is also a good time to delete any faces that you won't see during a standard use case of your asset.

Next, we want to try and straighten our UVs where reasonable. The "where reasonable" part is there because we don't want to straighten our UVs or edges so much that it grossly warps our materials and textures in the end. Straightening is important for all types of Real-Time 3D art, but especially so for hard-surface 3D art. Because hard surface content usually contains many boxes, rectangles, and square forms, it means that our UVs similarly retain those shapes. However, depending on where, how, and why we put our Hard Edges and UV seams, we can end up with irregular shapes. Not all edges need to be straight, and they don't need to be perfect, but it's important to make the effort so that we can optimize our UV space usage. After all, if most UV islands have straightened edges, it means we can pack islands into our sheet and give them more resolution because it's easier to pack squares and rectangles into a square UV panel than irregular or awkward UV shapes.

The method I am using is hotkeying the Straighten UVs tool found within the UV Toolkit add-on. It makes this process go very fast. I also use the Align tools within here that I also bind to hotkeys.

Next, I use an add-on called UVPackmaster 3 to help me optimize the usage of the UV space and to UV very quickly. I start with a Normalized UV pack to make all our UVs the same resolution. It'll give us a good starting point for manually scaling various islands up or down depending on their visual priority. I also turn on Pre-Rotation Disable and set the unit to 90 so that it only ever rotates my UVs at 90-degree angles. Straightening our UVs and keeping them locked at 90 degrees helps reduce artifacts and pixelation when our textures get affected by compression in the game engine.

The more important a UV island is, the bigger it should be on this sheet. The inverse also applies; if something is very small or unimportant, it can occupy a much smaller footprint.

I like to apply a tile checker material so I can see how the scale of various islands compares to one another. Then I simply hit the Pack button in UVPackmaster 3, and for the most part, we've got our UVs. From here, I will manually select some less important islands and scale them up or down before repacking or adding additional cuts where it makes sense to reduce awkward or irregular UV shapes. UVPackmaster 3 will respect our UV scale, so it's easy to change its size and then simply repack.

Lastly, we need to set our smoothing and triangulation before we can export to Marmoset Toolbag 5 for baking. Unfortunately, my friends, this next add-on is not one I can share, but at the very least, I can get you pointed in the right direction on how to get there. Watch carefully:

Let's talk about what’s happening here. With one hotkey, I am selecting a mesh and choosing which axis or axes to mirror it across, but not just that. This tool not only mirrors the geometry, but it also applies the Triangulation modifier and offsets our UVs automatically for the mirrored geometry. Typically, symmetry, mirroring UVs, and applying triangulation would all be manual steps. Within the Mirror modifier in Blender, at the very bottom is a menu that allows you to offset your UVs by a specific amount. Again, I cannot stress the time and effort saved with tools like this. With Blender and some Python code wizardry, you can create your own tools to apply multiple, pre-set modifiers at once. Credit for this tool goes to one such wizard himself, Elliot Sharp.

Ok, now the last step before we can export and get to the bevel shading and baking: Setting the Smoothing.

Select your entire mesh, right-click, and select Shade Auto Smooth, and set the Angle to 140. Why 140? Because it's somewhat of a sweet spot for most hard surface game engine content. It's aggressive enough to smooth most surfaces that need to appear continuous (like curved elements or beveled edges) while still preserving intentionally sharp edges. 

Bit of a tangent here (no pun intended) but in the 3D Game Art world, there is very often a fundamental misunderstanding and much mysticism around how hard edges, UV splits, tangency, normals, and game engines all talk to one another to ensure fantastic and consistent-looking game art. This is a topic that could take up an entire article in and of itself, so for now, I will direct you to this fantastic tutorial and breakdown from Joe Wilson of Marmoset. 

We're safe at this stage to go ahead and export our High and Low meshes. Now, let's head to Marmoset!

Texturing

I textured this completely in Marmoset Toolbag 5 using a variety of Materials, Fill Layers, and Smart Materials. Substance 3D Painter is my usual go-to when it comes to production work, but for my personal pieces and workflow exploration I'm using Marmoset Toolbag 5 more often because I can just immediately start applying materials and getting an sense of the read at the same time that I'm doing my lighting, bevel shading, and presentation. It’s just extremely convenient and all-in-one software.

Let's look at the Bevel Shader in Marmoset and how it works. It works the same regardless of whether you're here for either the Lookdev or Real-Time production method.

You can, of course, start up a default Marmoset scene, but if you'd like to follow along or speed things up, I'd like to provide you with my Marmoset scene template. Now, let's set up a basic Bevel Shader. If you've never used this before, you're in for a treat. 

Import your mesh and create a new shader by clicking the + button at the top. I like to do a mid-gray Albedo, medium Metalness, and default Roughness should be fine for now. The most important thing to do is to go to the Surface panel and click the subtab icon. This will roll out a few options for normals, the one we want, of course, is the Bevel option.

When you click it, you'll notice that nothing changes. That's because we need to activate Ray-Tracing Mode. After you activate Ray-Tracing mode, your model may sometimes look like this:

Don't worry, though, if this happens, it's just because the bevel shader is far too wide. We can easily and quickly adjust this by moving the Bevel Width down. I also crank Bevel Samples to max, because why not? This makes it so that the quality of your bevel shading effect is much higher.

With Bevel Shader and without Bevel Shader:

Nice! You can see that all of our edges now have a nice bevel to them. Historically, to get this look, you'd have to do highly manual steps, either via ZBrush Dynamesh + Polish, Turbosmoothing with control edges, etc. Again, I cannot stress this enough, the ability for us to get smoothed edges with a shader and slider is incredible.

OK, before we can move on to Materials and Texturing for real, we need to cover Baking. In essence, all Baking is projecting shading, details, and other elements from one mesh to another.

Modern day texturing typically uses objects called Smart Materials, which are basically base Materials with the added bonus of coming pre-packaged with ambient wear, edge wear, dirt, grime, etc. Basically, they are a good launch pad and way to get you a head start.

A breakdown of a given Smart Material might look something like this:

• Base material: Steel, Plastic, etc.;
• Wear patterns: Edge wear, dirt accumulation;
• Procedural effects: Scratches, oxidation;
• Custom hand-edited effects.

But, to draw upon the strengths of Smart Materials, we need to bake several Maps, which are textures that contain various and different types of information that tell our Smart Material how to apply what effects and where. 

Let's cover Baking for both the Lookdev and Real-Time methods. They are nearly identical with only minor differences. Once we get to Texturing, everything is the same for both methods from then on.

Lookdev Method

Very simple here. First, click the little bread icon to create a Bake Project. Second, drag your High mesh into the High folder like in the example here. Third, duplicate that exact same set of High objects, drag them into the Low folder, and change their suffixes from _high to _low. And we're done. Objects need exactly the same name with just the suffix changed, for this to work. For example: Handle_high and Handle_low. Easy.

Note: I've gone ahead and named my objects specifically, but you can literally leave them as whatever Plasticity randomly names your objects. Just change the suffixes and you’re good. You can easily do this by Batch Renaming in Blender.

Next, click on your Bake Project here in the scene inspector, which will roll out a variety of options. The Bake Mode will be Interactive by default, but I prefer to set it to Offline because otherwise, as you make changes, Marmoset will automatically update your content. It's convenient, but the cost is that our performance can chug quite a bit. Offline mode makes it so that we only see changes when we manually execute them instead of being constantly updated.

Create an output for the Maps we're about to create. I set my Samples to 32, my Format to 32 Bits, and my Padding to Extreme. Max those suckers out, we're trying to make our work as good looking as possible after all. Next, a 4096 texture is fine here. Overkill for our real-time method, but that doesn't matter here. Lastly, tick Normals and Ambient Occlusion. These are the two most important maps to validate and get correct. You can basically be sure that if these look good, all the other maps we bake after this will look good too.

When you're all set, drag the Default material onto your Low folder, then click the Preview Material button.

You'll know you're set when you see the Normal and AO texture slots have been filled, and you can now start to see the AO appear on your mesh.

Let's talk about what we’ve done so far. We've taken a fully triangulated, Smart UV'd, but otherwise untouched model from Plasticity, put it in Marmoset and applied a Bevel Shader, created a Bake Group, dragged the High content into the High folder, duplicated that mesh, put that duplicate in the Low folder and changed the suffix to _low and baking a Normal and AO from that. If you're a real-time artist for games, reading this section, many alarm bells may be firing right now. But relax, breathe, it'll be OK. Remember, the goal is not to bake a perfect Normal or AO here, it's to give ourselves a launchpad for our Smart Materials to latch on to.

If you're not a real-time artist for games, well then, sometimes ignorance is indeed bliss!

You may encounter errors like this that are too big to disregard, but we can fix those easily by adjusting our baking cage by clicking on our Low folder and reducing the Max Offset.

Cage pushed too far out by default, causing errors:

Much better. Hit Start under Bake and we're good!

Lastly, let's bake a few more Maps that help drive our Smart Materials. These are my usuals. Now, we can move on to Materials!

Real-Time Method

Boot up Marmoset Toolbag 5. Click the bread icon, click the Bake Project it creates, click Load, and select your High and Low-Poly models.

Here we can see our High and Low meshes are imported and have been placed within their respective folders in our Bake Project. We can hide the High or Low via the toggle near where they are loaded, and they will import with whatever materials are assigned in Blender. I've taken my high and applied a simple color so I can keep track of how the shading looks as I work here. But the most important thing is that for our High material, we need to switch the Surface mode from Normals to Bevel. Nothing will happen just yet because we need to activate Ray Tracing mode to see the Bevel Shader work.

Click this button to activate Ray Tracing mode.

If your shading comes in looking very wide and soft like this, don't worry. We can easily adjust. Tune down your Bevel Width and crank up your Bevel Samples while you're at it. Bevel Width will determine the softness or sharpness of your edges, and Bevel Samples effectively increase or decrease the quality of that bevel.

With Bevel Shading:

Check it out! All of our edges now have a nice bevel to them. Without Bevel Shading:

Ok, we can toggle Hide High from here and hide our high-poly. Let's toggle Show Low and make our Low poly appear now. Keep Ray Tracing on, though.

Now with our low-poly toggled on, I change the Default material to be roughly similar to how the high-poly looked. Now we hit the Preview Material button on the left.

And there we go. Kind of crazy! We have some skewing issues we need to fix, but thankfully, Marmoset Toolbag makes that easy. But otherwise, we've got our real-time mesh bakes! It's pretty hard to tell at a glance that this isn't our Bevel Shaded high poly mesh, and that’s exactly what we want.

Let's fix the parts that are skewed, but otherwise, we're about ready to start looking at texturing.

You can see that this bolt is being pulled towards the corner. This isn't ideal as we want this detail to face outwards like it does on the high-poly bevel-shaded mesh. To fix this, we click on the Low for the element that needs fixing, we click Paint Skew, make sure the Value is black, and max out Flow. Then we paint directly on top of that area that is skewed.

And there we go! We can see the black mask affected that area of our UVs, and the part that was skewed is now straightened out. Easy! Go ahead and do this for any parts that are skewed like this, all across the mesh.

Lastly, we want to bake a few more Maps to help drive our Smart Materials. Here are my usuals:

Alright, let's move on to textures!

Texturing

From here on out, our Lookdev and Real-Time workflows no longer diverge. I'll be showing you the file breakdown I worked on for the final Sci-Fi Tanto project.

To get started, we need to make a Texture Project. We do so by clicking this little paint palette, which generates a Texture Project folder. Click on it. If your Input maps aren’t automatically plugged in, we can plug them in manually. These are the maps that we had baked in the prior steps, regardless of which workflow you used.

Take your main material that you’ve been using, and plug it into the Linked Materials slot within our Texture Project. What this basically does is connect any future materials or smart materials we drag and drop to any object within this Tanto material, in this case. Let's look at that in action.

Switch to the Texture tab above, which should spawn a viewport similar to this. You should see the Library appear below, if you don't go to the Window at the top of the software and click the Library in the dropdown. We can then begin scrolling through the library, where we can see all the tools at our disposal. Let’s drop a random Smart Material on our mesh and see the effect.

Here, I plugged in the Plastic Faded Painted Smart Material. While this needs much adjustment, we're immediately seeing the benefits of what Baking and Smart Materials do when they work together. Thanks to the Ambient Occlusion we baked, we're getting that yellow grime in crevices, pockets, and where surfaces connect. Thanks to the Curvature map, we're also getting a bit of edge tarnishing along the sharper-edged surfaces. Now, let's look at refining this material a bit.

Click on the Layers tab next to Materials. Inside here, we can see the folder for our Smart Materials. If we click the folder, we can see the individual layers and masks that make up our Smart Materials. This is where we can begin to fine-tune. I'll show you an example with this material before I break down what I used for the final project.

If we click on any given Layer here, we get incredible control of each element that builds up that specific layer. So not only can we fine-tune the Layers themselves, we can fine-tune the elements that define the layers. In this example, I want to get rid of all the scuffing, change the color of the green-yellow grime, and increase its amount to show you how the effect works.

I click on the Plastic Base layer and tick off the Roughness since I don’t desire the effect it’s producing. I can always add a different one of my own later. Now I want to tune that grime, so I click the layer and change the color to more of a dirt brown. I then double-click the mask of this layer, which unhides the masks that control how this grunge appears on the mesh.

Lastly, I tune the effect by tweaking the Intensity and Contrast, but as you can see, there are many, many effects you can tune.

Ok, now that you've seen a basic Smart Material and how it works, and the various Layers and Masks we can adjust to change different things about it. I'll break down the materials I used on the Tanto.

I apply the Steel Bare smart material here, but it's applied to the entire object when we probably only want it in certain areas. We want to isolate the material to certain parts, so we use Masks to do that. Select the Mesh Select tool. Click on what part of your mesh you want to isolate. Right-click on the folder of your Smart Material. Mouse over Add Mask and click Paint Mask (Black).

There we go! You'll do this for everything you want to apply only to specific elements or objects.

I approach wear and tear in similar ways from project to project. Eventually, you come to understand that materials become worn in certain ways and objects collect detritus in similar ways. You can observe this just by looking at any household object. Anywhere two surfaces have been touching for a while probably has at least a small layer of discoloration, dust, or wear. Objects that have been left to sit for a while will usually collect grime and dust or rust in the same locations. It's the nature of weathering.

For example, take this stock image of some heavily rusted chains. Notice how where the chains are physically touching, it’s especially corroded and discolored? Similarly, some of the extremities have kept their metal sheen? While this example is extreme, we can still borrow the principles of how objects are affected by grime, rust, dirt, dust, edge wear, and etc.

Let's take a look at what I consider the absolute requirements: Ambient Occlusion and Edge Wear. These are the baseline wear types that will instantly give more believability to your work, no matter what the style is.

I like setting up what are basically AO dirt layers and masking them off. It's easy to dial in the look by using the sliders like we covered above. I do a similar thing, but with Curvature and edge wear. You can get a head start by using some of the existing Smart Masks within Marmoset. Simply drag and drop them over your Layers.

This pipeline makes it really easy to start viewing your work early on. This means you can start getting ideas for what is working or not, or how various parts of your model are going to read when lighting hits them. Basically, after I get a solid blockout going, I set up my Marmoset Toolbag 5 scene and just pass the model between the software as I update it at various stages. Additionally, I update my lighting as I go to best showcase the model. 

I usually have at least two cameras going at once. My Main Camera, which will be the lens through which I capture my final renders, and then a generic camera (Camera 1 in my scene). My generic camera is what I use to look and move around my viewport as I'm looking for interesting angles or capture spots. When I've found what I think is a good angle, I switch to my Main Camera and approximate the location I discovered with my generic camera.

It’s important to remember that every single piece of art you do is different, which means they will all need at least a bit of special treatment in terms of how you capture their renders. Some art pieces are square and easy to frame, others could be things like spears or rifles, which are longer and may require a different Field of View. Aside from capture angle and lighting, certainly the next most important thing to tweak is your Field of View.

Here's how this looks with nothing else changed except the Field of View. Notice how squashed the handle becomes as the perspective of the camera distorts? For some projects or artists, effects like this may be desirable. It’s important to experiment and find out what your specific style is!

Lighting

I'm a firm believer in using as many lights as it takes to make your stuff look awesome. I also do my lighting in a way that I'm not sure if it’s considered unusual or not, but I light almost entirely using Directional Lights. I kind of look at it like "sculpting" with light, and directional lights help me achieve that.

I don't have any predefined limits or rules for myself other than I always make sure to have a Rim Light, because Rim Lights are sick.

At this point, I have a template scene that I like that I use kind of as my baseline for rendering. The most essential lights you need are your Key Light, which is generally responsible for providing the major light source for your work. Your Fill Light, which is generally positioned opposite your Key Light and fills in a bit more of the object that isn't being lit as much from your Key Light. And lastly is the Rim/Back Light.

Key Light off and on:

Fill Light on (note the handle area versus the above image) and my personal favorite, as I mentioned, the Rim Light:

This is one of those cases where I add multiple Rim Lights to capture different things. One for the top and one for the bottom. Rim Lights are especially fun because you can use color to heavily stylize the "vibe" of your presentation. I personally prefer my Rim Lights to be subtle. It's kind of like the Bass of a good song track: your mind and eye might not highlight it when it's there, but you can definitely tell when something is missing. 

Really it takes an objective look at your work. What shapes are working? What is unique about your model? What do you want to highlight about it? Your lighting pass can be used as a tool to accentuate and show off specific things about your work. Admittedly, this is a skill and not something that just happens on its own; you'll have to look at other work you admire and try to mimic what they've done and then filter that through your own artistic lens. That way, you can start to develop your own style of presentation.

Post-Production

The vast, vast majority of my post-production is done all within Marmoset Toolbag 5. I mess with the Camera Settings, and I usually toss in a Shadow Catcher, which helps ground our object in the scene and gives a strong shadow. Not required, but it's a nice addition.

You can see within the camera that there is a whole slew of settings to tweak. My favorite tab here is the Color Tab. By changing the Tone Mapping and setting below, we can get some really interesting and different renders. Have a look at some examples:

I highly recommend playing with the Highlights, Midtones, and Shadows sliders to see how they affect your work. Sometimes I'll take something into a photo-editing app and tweak the Levels, Contrast, and Saturation to fine-tune. 

And while I’m on this topic, I want to take a minute to express something here. There's a perspective that pops up sometimes that proposes that post-processing or post-effects or photo-editing your work is somehow "cheating". I'm going to let you in on a little secret here if you weren't aware. All the major players do it. There's a reason why post-effects and editing is its own entire skillset and vocation, it's because you're using artistic discretion and skills to put the proverbial "cherry on top" of the work. It's complimentary. It's that extra layer of juice that can take something from pretty cool to extra cool. Don't be afraid to leverage these tools to make your work look better, after all, in the end, that's the goal, right? You're supposed to want to make your stuff look as cool as possible.

Creating an Appealing Pose

Two initial thoughts come to mind. The first is working off of good material. That could be photo reference, concept art, or what you have. Design is an entire skillset and vocation in and of itself, so if it's a weak point in your skillset, then being able to choose an interesting design to work from is how you shore up that gap. I think when you’re starting out as a 3D artist, there is enough to learn that if you’re also concepting your own stuff alongside learning even just some of the things we’ve covered today, it's possible that it might make your journey more difficult. In that same breath, if you feel in your heart that that is the right path for you, then lean all the way in. 

The second thought that comes to mind is just digging deeper into whatever it is you're trying to accomplish and establishing your goals with yourself up front. Be thorough in your efforts. Look around at the landscape of content related to what you're attempting to make. Find work from other artists that really inspires you and try to make something that looks as cool as what they are making. Your skills will grow naturally by making these efforts.

Conclusion

The main challenges for this piece were modeling the handle and still being in the process of getting more proficient with Marmoset Toolbag 5 texturing workflows. 

The handle was the most challenging aspect of this project, just due to the nature of the obtuse curves layered on top of compounding shapes, but that’s also what made this interesting to work on. I find that the projects I have the most fun with are ones where I’m not totally sure how I’m going to get to the finish line.

Generally, I do full "hero" assets that are much more intricate than this one, but in making this, I was reminded that projects like this one are great injections of fun. They're small but interesting, simple with a small handful of complicated parts, and that gives you the ability to complete them quickly while still having some challenges and unknowns to solve along the way. Sometimes you can work on a giant project, and it can feel like a slog because it can take so long. Other projects are so straightforward that you’re just going through the motions. I think it's important to find something that lands somewhere that you can't take it entirely for granted, so that was a nice reminder.

On the technical side, I learned some better ways to loft between edges while maintaining tangency, which allowed me to get some of the more interesting curves on this weapon. I’m looking forward to leveraging more lofting and soft shapes for future projects.

Advice for Beginners

Where to even start. I have quite a bit to say here. 

This question is especially relevant. Recently, some newer artists and students have reached out asking this very question. It's no secret that the last few years have been pretty much unprecedented for our industry, and to say it's demoralizing is a complete understatement, and that's coming from someone with nearly a decade in this field now. 

I remember being a student when THQ collapsed in the early 2010s, and not even 10 years before that, Acclaim Entertainment went bankrupt. These were major publishers with a lot of studios and IPs under their umbrellas. Those shakeups affected and displaced a lot of people, some of them leaving the industry for good. While I can't compare firsthand to those earlier periods, the scale of recent layoffs and studio closures has been staggering and feels multiple times worse. But I don't think it will be like this forever.

The Shortcut Trap

An operative word that comes to mind in this current timeline is "shortcuts", and honestly, the abuse of them. I see far, far too many aspiring creatives running to buzzword, flavor-of-the-year options, hoping to find themselves a workflow that really activates them. Finally, this new thing is what they have always needed to be productive and achieve their dreams and goals! Finally, they found the magic bullet that unlocks their potential! 

But here's my personal experience and belief: whatever the latest-greatest thing is, the people who will come out on top are the ones who spend a long, hard time investing in actually understanding their craft. The people who are really dedicated to learning everything they possibly can about their craft and all the inner workings of it are always going to come out on top of those who expect to push a button and have a machine spit out their magnum opus in the long run. Learning the why behind the how, that's how you become someone who can create a lifetime of creative success for yourself. 

Additionally, there is an aspect of this artistic lifestyle and career choice that I feel isn't spoken about much. And at least in my case, it only arrived to me after nearly a decade of doing this, it's that when you make the choice to pursue this industry and line of work, a creative life, you are effectively choosing a lifetime of committing to investing in yourself and growing your skills.

There is no off-ramp unless you decide to quit for good. There is no point you can reach and where you can say, "I've done it, I've learned all there is to learn! I am complete as an artist!". To stay competitive and sharp, it requires constant effort and investment in your craft; otherwise the stagnation sets in. While that can sound intimidating, I wouldn't trade what this career and artistic lifestyle have given me for the entire world. It fundamentally changed my perspectives and life for the better.

My Personal Journey

I share this because I want you to know what's possible. I grew up as a lower-middle-class American country kid with three siblings, with virtually no artistic support or innate prospects. My dad worked in a factory, and my mom cut hair for a living. They were, thankfully, good people and good parents, but we were not artistic people. Where I'm from, saying you wanted to be an artist was something people made fun of. It wasn't something people seriously considered making a living off of, let alone becoming. I had to learn how to do this whole art thing the long way, and were it not for the support of other people who were also trying to learn this whole art thing too, it’s hard to say if I'd have pursued this lifestyle and career. If someone like me, who did not grow up in an art-conducive environment and had no innate talent, could make it, you absolutely can too.

Invest In Yourself

Ultimately, in my heart of hearts, I do believe things will recover one day. The thing I really latch on to, and what I recommend, is to invest heavily in yourself. Your skills, your abilities, your perspectives, the quality of your character, all of it. Invest in becoming someone who is capable, flexible, adaptable, talented, dependable, reliable, kind, empathetic, and remarkable. I genuinely believe that molding yourself into such an individual, while really spending the time and effort to refine your craft and abilities, will make you an exceptional individual. Even if times are tough now, when the pendulum eventually swings back and opportunities begin to arise again, you are going to be so much better off for having made that investment in yourself. That you can absolutely bet on.

Dylan Mellott, Senior 3D Hard-Surface Artist

Interview conducted by Gloria Levine