logo80lv
Articlesclick_arrow
Research
Talentsclick_arrow
Events
Workshops
Aboutclick_arrow
profile_loginLogIn

How to Create a Visually Compelling Post-Apocalyptic Walkie-Talkie

Lets follow Senne Bovée to explore the intricate journey of crafting a detailed, post-apocalyptic walkie-talkie. The realistic textures and details serve as tactile storytellers, echoing the items history and usage.

Introduction

Hello everyone, my name is Senne Bovée. I am a 21-year-old, aspiring prop artist, and student of Digital Arts and Entertainment at Howest University of Applied Sciences. I am studying their Game Graphics Production major with a minor in technical art.

I have loved working in 3D ever since I attended a Fusion360 workshop in high school about 6 or 7 years ago. After the workshop, I taught myself how to use Blender, and soon after, I also learned to work with Maya, Substance 3D Painter, and other tools during my studies. 

I created this project for a final assignment for the Game Asset Pipeline course at DAE. As the name implies, this course focuses on the entire pipeline involved in creating game assets, with a particular emphasis on AAA quality. This workflow and many of the techniques used were quite new to me. Despite having some experience with all the software and tools I used, I learned a lot during this project and am excited to share some of that knowledge with you, and perhaps share a few tricks I picked up along the way.

About the Walkie-Talkie Project

As I mentioned earlier, this project was created as a school assignment. We were given a few rough guidelines to follow, but other than that, we had the freedom to create almost anything we wanted.

The primary guideline was the theme. We were instructed to create a prop, vehicle, or character that would fit in a post-apocalyptic game set on an archipelago of islands.

Another major requirement for this assignment was the use of AI image generators during the idea concepting stage. While this remains a highly controversial topic, the teachers wanted us to at least be familiar with what is possible and how it could be integrated into a larger pipeline. Although I am on the fence about these tools myself, I do see the importance of being able to work with them, should the industry move in that direction.

I started my project without a definite plan for what I wanted to make. However, I did have a few ideas that I was excited to try out. I really wanted to try texturing duct tape, and duct tape fits right into the post-apocalyptic theme. With this in mind, I started brainstorming logical uses for duct tape and quickly decided I wanted to create some kind of scavenger tool or weapon. 

My first step was to make a mind map to quickly explore various ideas in an organized manner. I considered the ‘problems people in this world might face and their possible solutions. I also thought about items a scavenger might find and how they could repurpose those in a world with limited resources.

My first idea was a multitool that could serve a similar purpose to Fallouts Pip-Boy, where the player could check their quests, inventory, and so on, but also function like a Swiss Army knife, where different pieces could be detached or unfolded to be used as scavenging tools. While I loved this concept, I couldnt quite figure out how to translate it into a visually readable or interesting design. The AI tools were also of no use; they just generated a visual mess that would have been difficult to take anywhere.

Learning from previous projects not to fall in love with an idea too quickly, I decided to try different things and perhaps keep things a little simpler. A good design should be quickly understood by the viewer, so why not recreate something pre-existing but make it look handmade? My next few ideas were simple: a diving helmet, a harpoon gun, a mine/grenade, or even a radio of some kind.

I figured the radio could still include many of the multitool’s functions; the player can still keep track of their inventory and quests, but it would be much easier to design and convey to the viewer what they were looking at.

So, I decided: I would design a radio or a walkie-talkie. Long-range communication would be quite useful in the worlds setting. Now, what would it look like? I had previously mentioned wanting it to be some kind of handmade scavenger tool, so, I decided to go scavenging myself. I walked around my dorm and gathered everyday items that could be repurposed to create something new.

I collected a roll of duct tape, a TV remote, a 9-volt battery, the circuit board from a Raspberry Pi, and a bunch of other knick-knacks. I fed these parts alongside my idea to OpenAIs DALL-E 3 and let it generate several images. I combined my favorite generated images with my personally collected references to create my final concept. 

Modeling

It was time to get started on the model. My software of choice at this stage was Blender. I really like how customizable the viewport display is and I much prefer high poly modeling in Blender than in Maya.

I began working with different proportions and sizes and eventually settled on something similar in size to a modern smartphone. It should be large enough to create an interesting model and nice textures, but small enough to be easily carried around. I took all my favorite parts from my references and the generated images and made a rough block out.

This initial model just needed to convey the idea and be identifiable as a walkie-talkie. Once my block out was completed, I asked my peers and family members what it looked like to them. If the majority told me it looked like a walkie-talkie even without textures or detail, then, I knew I was ready to start working on the high poly version.

While I love modeling and bringing my ideas to life, I am somewhat casual when it comes to proper topology and modeling techniques. If the workflow allows it, I will try to go about it in a faster and more efficient way. For example, there are metal rods on the right and left sides of the screen. Instead of making the entire mesh around the screen one part, I used floaters or floating geometry. These parts have the details I need, but are flush with the rest of the mesh. This makes it so that when I do bake this down later, you wouldnt be able to notice it and it would appear as one solid piece:

I didnt model the indents for the buttons and other similarly small details. Instead, I kept it a flat plane and used height maps in Substance 3D Painter later to get those details.

I also changed quite a few things from the block out. I added a sort of ‘rim’ or shell around the sides of the walkie-talkie; this way, I wouldn’t have to worry about modeling the inside since it would be closed off. 

Most of the high poly model was standard stuff. I would model the shapes, subdivide them, and add control edges and bevels where needed until it looked nice, and I could move on to the next piece. While I was doing this, I noticed a few parts in the block out that were just there for the sake of detail and didn’t really serve a function. I made the decision to remove them or rework them into simpler versions, as I wanted to make sure the viewers’ attention was focused on the screen and circuit board. I removed the spring and screws from the antenna for example, and simplified the way the keychain ring was attached.

Now when it comes to the duct tape, I might’ve gone a little overboard and added more detail than I should’ve at this stage in the process.

I had just purchased a pack of wrinkle and fold textures that I knew I wanted to use for this project, but rather than adding those in the texturing phase, I added them here already. I subdivided the duct tape strips several times and added displacement maps with my newly acquired folds. In retrospect, I wouldn’t have done this since this added over 13 million triangles to the scene. But I thought it looked quite nice and gave me a boost of motivation.

Topology

Now that I have my high-poly done and each piece of the mesh looks like it serves a purpose, I am ready to move on to the low-poly mesh. For this, I decided to move to Maya. I prefer Maya’s unwrapping tools and, since I like to unwrap each piece as soon as I’ve modeled it, it only made sense to move to this software. I saved and imported my 300+ megabyte mesh into Maya and got to work.

Normally for remeshing, I like using Maya’s quad-draw tool, but since this model was so simple and straightforward, I didn’t see the need for this. So, I got to work and, piece by piece, recreated the high poly. The way I like to work is by keeping the silhouette of the low-poly and high-poly as close as possible, then going back in with the target-weld tool and merging any vertex that doesn’t contribute to the silhouette into one of its neighbors.

I then add seams over all my hard edges and anywhere else where it is needed, and start unwrapping. I would add my seams, unwrap, straighten, and pin some edges, unwrap again, and again and again. I did this until everything was as rectangular as possible, that way, I could pack them tightly together and use as much of my UV space as I could.

As for UV packing, I always start with Maya’s auto-packing. It usually gives me pretty good results and helps me make sure all my UV islands have the same texel density. I then manually go back in and move islands around where needed. If I have space left over which, in this case, I did, I will scale up the most important parts of the model so that I can make the textures even higher quality. In this case, I made the LCD screen and the circuit board ever so slightly bigger to make the entire UV map more compact.

The last step was to assign all the different pieces of my high-poly and low-poly to baking groups and export them as FBXs.

Texturing

Texturing is what I was most excited about. I also spent the most time on it. I imported my low-poly into Adobe’s Substance 3D Painter and baked down my high-poly. I then went back and forth between Substance and Maya to fix most, if not all, the baking issues and ended up with all my different maps.

Once all the baking looked correct, I got to work, starting with the duct tape. Since this was the catalyst of this entire project, I wanted to make sure I got it right. Throughout all my work so far, I have been researching how duct tape was created and how to accurately represent it in my textures. I learned that the outer layer of the tape is made from polythene, a type of plastic, which is commonly dyed using aluminum powder, giving it that recognizable metallic look. Since this is plastic mixed with aluminum, it is partially metal and non-metal, and one of the rare cases where a metallic map can be a value between 0 and 1.

The nice thing about having physical reference, rather than just pictures, is that I can view the material from different angles and lighting intensities and really analyze how it behaves.

Here you can see how I built the duct tape material layer by layer:

  • This is the base layer, here I define the main color, metallic value, and roughness value.
  • I use my AO and Curvature Maps that I got from my bakes to add highlights and shadows to the base color. This accentuates the wrinkles even more.
  • I then added a gradient, where it is darker near the bottom, and has a light yellowish discoloration on top.
  • Now the Height Map. This probably sells the look of the duct tape the most. I used Substance 3D Painter’s ‘Anisotropic noise’ texture stacked horizontally and vertically and blurred it slightly.
  • Now, I add roughness variation using a grunge texture as a mask. This does a lot for realism, since nothing is ever perfectly clean.
  • Finally, I added a dirt layer. I use my previously made height map as micro details in a generator. This way I can add dirt in all the little crevices.

The duct tape also needed an inside, even though there aren’t many places on the model that you can properly see it. Working on the underside of the duct tape was much more straightforward, since it doesn’t have those unique mid-range metallic values. The material is basically built up in the same way where I stack gradients, roughness details, and height details.

I ended up turning both into smart materials so that I could easily re-use these in later projects. The next big material to tackle was the circuit board, since this arguably takes up most of the visual space. So far, I have been working in Substance 3D Painter. However, I decided to take a different approach and work in Photoshop.

I started by exporting the ambient occlusion map to Photoshop. This way, I knew exactly where on the texture each capacitor was located. I then meticulously drew out all the connections and random lines, making sure it was visually busy, yet convincing. I also decided to add some text and symbols commonly printed on circuit boards for additional realism.

Now, things weren’t as simple as just drawing out a texture. A circuit board isn’t perfectly flat and needs height variation between the two different shades of green and the text. Therefore, I had to also create a black and white mask for both of those. Finally, the only thing that I still needed was a mask for all the metallic bits seen at the end of some of the connections.

With these 4 maps, I could go through a similar process as I did with the duct tape. I added the green texture to my base color, used my masks for variations in the height and metallic maps, and finally added a bunch of dirt and grime overtop.

The next textures I’d like to talk about are for the battery and the screen since I used another piece of software yet again. I was keen on applying my knowledge of graphic design from high school. I figured it would be fun to design a simple battery brand. I didn’t really have much of a plan for this going into it, but I knew I wanted something colorful to add a point of interest to the back of the model. I then again applied my usual layers of wear and tear, this time also including some scratches.

I also used a very similar approach for the screen. I wanted to make it look like a segmented LCD display, and I like using Illustrator to quickly iterate on that idea. I tried making sure what was displayed on the screen also looked functional. I also made sure there was a bunch of information for the player to see that could potentially be integrated into some kind of game mechanic. 

You’ll notice the current frequency on the screen is set to 144 MHz. This was a conscious choice. After doing some research, I learned that 144 MHz is the amateur radio band for emergencies like natural disasters. It made sense that in a post-apocalyptic setting, it would be important to communicate any dangers or incoming threats efficiently and quickly.

This leads me to the final technique I used for texturing, which is photography. I knew I wanted to add a note glued to the walkie talkie, that had something important written on it. So, what better to remember than the emergency line? I tried making this in Illustrator and Photoshop first, looking for a nice font to use, or to even hand draw it. But then I figured if I wanted it to look real, why not do it for real? So, I dug out an old newspaper and a pencil and wrote it down, before taking a picture and importing that into Photoshop.

Yet again, I did this in several layers. First came the base color, followed by the mask so I could offset the height of the paper, so that the edges would catch the light.

And finally, I decided to make a mask of the graphite. That way, I could push the height map inwards ever so slightly and change the roughness of the pencil marks to make it as accurate as I possibly could. 

The last thing I wanted to talk about when it comes to texturing is the anisotropic effect on the dial and the rods next to the screen. I had never done anisotropy before, and figuring out how to use it and exporting it correctly was challenging.

Anisotropic reflections happen when you have a material with many small parallel grooves. Each of these grooves reflects the light individually, creating a unique effect, especially when these grooves are circular. You can see this effect on things like pans and pots, where you have a reflection on either side converging in the middle.

To create this as a material in Substance 3D Painter, you need to set the shader template to “PBR-metal-rough-anisotropy-angle”. This will allow you to create 2 new layers, an ‘anisotropy angle’ layer and an ‘anisotropy level layer. This is very useful since you get to control the intensity and the angle of the effect individually.

But the biggest issue is that many other software like Unreal Engine or SketchFab want a single texture RGB texture that has all the anisotropy information baked in. So, Substance’s two black-and-white textures are useless. Luckily, I found this wonderful free converter made by Dongkoon Yoon. Using this converter, I was able to use my anisotropy outside of Substance 3D Painter.

After finishing this project, I learned of a really simple way to create an anisotropic effect using only a normal map, with no need for converting specific layers at all. You can achieve this effect by simply adding a cone shape to your normals. It is important to note that it wont show the reflection mirrored around the center, like you normally would. However, it is a very easy and cheap way of replicating the effect.

Rendering

Near the end of texturing, I ran into some technical difficulties which cost me a lot of time. Because of that, I was left with just a few hours before I had to go present my final assignment, and I still didnt have any renders. I knew I would render in Unreal Engine, since it only made sense for a game asset to also be rendered in a game engine.

I imported my model and used a basic 3-point lighting setup as a base, and added a few extra lights here and there where needed. Unfortunately, due to the lack of time, these renders ended up being rushed, and it showed in the final quality. I also got a lot of feedback on what I could’ve done better from my professors. To keep things brief, my framing and focus were off. I had a very shallow depth of field without a clear focus point and, because of the framing, sometimes parts of the model, like the antenna, were cut off. 

Several months later, I revisited the project. I had scored quite well on this assignment, and my post on SketchFab had also gotten staff picked. I knew I didnt do the model justice in my original renders, so, I decided to start over.

First things first, I changed from landscape renders to portrait renders. Since the model is quite tall, this only made sense. This way, it can fill up more of the frame and show off more details.

The second change I made was the background color. I noticed the original dark blue didn’t create enough contrast, leading to the renders looking quite forgettable. So, I decided to go to ArtStation to look for inspiration. I found this post by Aleksandr Kulykovskyi and I really liked the soft beige background and the lighting used. I used this as a starting point and tried my best to recreate what I liked in their work before taking it in my own direction. Once I was happy with how everything looked, I had close to 10 lights in my scene.

The first was a very dim skylight that contained an HDRI. Next was another subtle light, a directional light. These two were mainly there to make sure even the darkest shadows wouldn’t be pure black.

I then added my core 3-point lighting, which turned into 4 lights, two of which I used for rim lighting. I had a spotlight shining down from above to cast a gradient of light onto the model, brightening up the top and drawing your attention to the screen first.

Finally, I added a few more super subtle lights. These weren’t meant to light up the model, but to catch all the height detail.

As you see in these screenshots, the tiny bit of light spill on the circuitry really brings out and accentuates the height detail and effort put into these materials. I finally ended up with these renders, of which I am very proud. 

Summary

Creating appealing props requires storytelling through every detail. Consider the user, usage, and utility of each detail or element of your design. The better you understand the purpose and context of your prop, the better you will be at judging whether it looks good and relevant. I tried to give each piece a purpose and a reason for it to be there.

You can also do a ton of storytelling in texture work. You can make things look old and used, really make them feel like they have history. You’re making a gun? Add scratches on the magazine to show it has repeatedly been ejected and re-inserted. Are you making a ladder? Add some footprint or marks to show it has been used.

The more of your prop’s history you can convey in just the model and textures, the more real it will feel, regardless of getting each material physically accurate, or each piece perfectly modeled.

Just be careful not to overdo it. Simplicity is key. Avoid overcrowding your design with excessive details, make sure you still have contrast and nice proportions.

But most of all, just experiment. From this project, I have really learned that each texture or material is like its own little challenge. I had a ton of fun figuring out and learning about how certain materials are made in the real world. And being able to translate that newly acquired knowledge into an art form is really rewarding.

Just a few months ago, I didnt think I was capable of creating something like this, yet here we are. Before I wrap things up, I would like to thank Alexander Delagrange and Dries Deryckere for their great feedback during and after the completion of this project.

Senne Bovée, Prop Artist

Interview conducted by Gloria Levine

Join discussion

Comments 0

    You might also like

    We need your consent

    We use cookies on this website to make your browsing experience better. By using the site you agree to our use of cookies.Learn more