Jose Gomez showed how he worked on animating the weapons for Neon Fury – VR first-person cyberpunk tower defense shooter.
I was born and raised in Caracas, Venezuela and I’ve been doing 3d modeling since early high school. 3d models were kind of magic to me, I was mesmerized every time I watched a CG character on screen. So, having access to making that magic was irresistible. I had a few jobs as a kid and spent every penny just pimping out my PC so I could mess around with 3d modeling software.
Finishing high school, I was already landing a few design jobs. For some time, my plan was to go to the US to study 3d animation as in Venezuela there was no place to do it at the time. However, I quickly found my first job in the advertising industry so I stayed in Venezuela. That industry was a school to me, I learned every aspect of 3d production. I got a handle of so many techniques and tools because I worked on a new project every couple of weeks, and every project was a completely different challenge. A few years later, me and Alejandro Martinez, a friend who had worked with me since the very beginning, decided to open our own studio.
We kept doing mostly TV spots stuff until around 2008 when we saw a chance to make a jump to video games. Our first prototype was Battle Tennis, a couch multiplayer arcade ninja tennis game that we got into a Microsoft XNA contest and came out with 20 grand as a cash prize. We used it to start Gasp Games, our games studio. We had made a few other games but by then it was around 2010 and we could start to feel the grip of the political crisis in Venezuela. Hardware and software were hard to come by due to currency controls, a law was passed that would send to jail anyone who made a video game that the authoritarian regime considered violent. The nascent indie video game industry there was forced underground. We barely survived while making Power Ping Pong, a mobile game based on Battle Tennis that we launched with EA/Chillingo. It was a really tough development given everything that was happening around us but with a passionate team it still ended up featured as Apple’s Editor’s Choice in a few regions.
Eventually, the Venezuelan crisis had me decide to shut down the studio in Venezuela and I moved to Colombia. There I started working with Teravision Games as Art Director on this project, Neon Fury, a VR first-person cyberpunk tower defense shooter.
At Teravision as Art Director I work closely with Luis Daniel Zambrano, our Creative Director there, to lay down the requirements for all the production we make. One thing we have learned working on VR is to try to unlearn everything we know about UX and design and just prototype and try things out as the interaction medium is so different. Even so, our first try was to work with different weapons that you could grab and switch from different parts of your body. That was something we learned from Robo Recall that worked really well there. However, after prototyping we found we wanted to lower the cognitive load on the weapon switching action as much as possible because Neon Fury was already pretty action loaded. And as such it would have a lot from players think through their tower defense strategy while actually being in middle of horde attacks. That’s why we wanted to find a way to make weapon switching as automatic as possible so it wouldn’t distract players from the game’s focus.
So, while our 80’s cyberpunk theme allowed us to pretty much “holomaterialize” different things in your hands and call it a day, I find a certain joy in watching complex mechanisms work. Let’s take the memorable Terran Siege Tank from Starcraft as a source of inspiration, or more recent and closely related the amazing Dominator gun from Psycho Pass. Also, our main idea was to make you feel like an 80s movie hero badass, and having this amazing mechanical device could surely help there. That meant we had to figure out a way to turn a pistol, a teleporter and a neon bow to a single device that could also fold up and retract away to leave your hands free to interact with levers and buttons for the tower defense interfaces.
A pretty fun challenge.
Luckily Paul Pereda, our concept artist is such a great guy to work with and stepped up to the task. We started working closely together, him painting out ideas while I prototyped with boxes and checked that all the pieces could work together in animation for all the different movements and transformations.
Early on we knew we wanted to make it as big and bulky looking as possible. This directed us towards having these metal parts and covers that gave us visual volume while still leaving enough empty space in between to contract the parts together.
It took a few days to get a box model that kind of worked, a title screen with all the animation timing set right, and to have all the major volumes create the overall shapes we wanted for different stages.
However, this means nothing until you try it in VR to see if it actually feels any good. So, for that I quickly kitbashed a super rough prototype in Zbrush and just linked the parts together to get the feel of it in Unreal within our game world. And boy, it felt so incredibly badass that we realized that all the hours we had spent working on it were not in vain.
From there we just had to refine and design genuinely looking machinery that could move around all the pieces. A process that involved referencing and studying all kinds of mechanical objects to design the different hinges, joints and motors. It’s not that it all had to physically work, and it doesn’t, but we definitely wanted to give the impression that this could possibly exist in the real world so we maintained a sense of realism in VR. This was a hard but fun process, on some days we would just get to the office excited that we had discovered a new kind of hinge on some weird lamp or something that would allow us to fold a piece the way we needed. Or we just had fun taking stuff apart to get ideas for making unusual things like the hammer of a gun that works like the inner mechanism of a retractable pen.
Since our devices had so many parts doing lots of things and working with each other, at first, we thought we wouldn’t be able to actually design whole devices and make them work. Usually, we would just leave them on paper until later stages, that was our traditional workflow. But that time we had to have each of those small mechanical parts conceptualized, prototyped, animated and tested specifically in VR. Only after making sure the parts really worked we could move on to detailing of each of them.
For that reason, we kept most of the mechanical modeling inside 3ds Max, as it was incredibly useful to model right on the animated prototype to make sure everything worked after every change. In fact, most parts never ended up going to ZBrush. Even the high-poly mesh was mostly just a smoothed-out version of the base mesh using OpenSubDiv to define smoothness levels of individual edges. We left most of the detailing for Substance Painter. Of course, before that we had to clean up the UVs and there Jorge Pineda, one of our 3d Artists, lent a hand defining the UV layout and painstakingly unwrapping every little piece.
The hands and gloves, however, had a more standard workflow, these were modeled and retopologized in Zbrush, imported to 3dsmax for UV mapping and animation and then given some Substance Painter texturing before in-game tests. I was pretty concerned at first about having visible skin as I knew our VR performance constraints would probably not allow for any complex skin shading. To my surprise though, just using a regular PBR shader gave us a working result for our lighting scenario. The skin was all hand painted and for that, I would usually just lay down a base color and then work with layers of red, blue and green to give the different fleshy undertones.
Once the time came to texture the gun, Paul laid down a few color patterns on concept. We wanted the gun to be kind of colorful to go with our theme and help us bring out some parts of the model. So, we got the gun baked in SP and started to lay down the general patterns and testing, always testing in VR to see if it worked. When working on VR, you have to make sure your silhouette is correctly readable from just about every angle. This part of the process was incredibly flexible as we were detailing in SP and making use of the new anchor points to have all the surface details drive our masks on top. That way we could lay down colors, get a sense of the general reading, and only leave small details for the late stage.
We also delayed having any kind of animation rig up until very late in the process. Since we were still making too many changes, we didn’t want to feel constrained by any specific hierarchy or pivot placement. So, for the entire prototype process we just had the mesh parts linked together being moved around. Only then we knew we were set on a final design that we moved to the final optimization of the model and the 90 bones big animation rig.
As the game is still in development, some things aren’t finalized. We still have to polish the lighting rig in Unreal and probably do some tweaks in the hands animations, bow deployment and UI display. It was also designed with add-on support, so it will have other modes. The thing is your hands and weapons are pretty much your main vehicle to move your character in a first-person game, especially in VR. So, you must design them as carefully as possible. This kind of gameplay brings back the kind of magic feeling from my childhood – being able to design these weird contraptions and immediately use them like real objects in VR is just nothing short of magic.