It was autumn 2012 and I was making my Steam account. It turned out that my standard username at the time has already been taken by someone else, meaning I had to pick a new name. Fortunately for me, the final Guild Wars 2 beta weekend happened just a few months earlier. I’ve played out all five races the game has to offer, and I recalled the names I used actually weren’t lame as shit. I didn’t recall the actual names, so I went scouring for illegal screenshots in hopes some of the screenshots will show at least one of the names.
I found exactly one screenshot with my character’s name on it. And because it was less lame than my previous names, Tamius became my pseudonym on the internet, and charr my avatar.
Few years down the line, I stumbled Jason’s Tauren cosplay (Akulva), which features some animatronics.
The animatronics are pretty neat, and really made me want to build something like this. Now, I’m not into WoW, but that’s not the problem because guess what else is 8.5 feet tall (canonically)? Well the title of this blog kinda spoils the answer, doesn’t it.
There were several problems, though:
- lack of skill
- lack of time
- since I was just fresh out of the uni, lack of monies
so this went to the “when I’m grown up and rich” pile. While I’m still not really rich, I’m at a point where a project like that isn’t going to bankrupt me.
Before we even enter the R&D phase, we should probably set up some sort of outline of what we expect of the costume. We already know how we want it to look:
But here’s some additional requirements:
- Costume shouldn’t need a lot of space to store. This means that any hard skeleton inside the costume needs to be disassemblable.
- Costume can’t weigh a ton. Target is sub-15 kg. If it goes over 10 kg, I need to have some sort of exo-skeleton that bears part of the weight of the costume while standing still.
- Battery power needs to last me at least 2 hours at the con (if local), 5 if international
- Cooling is not optional
- I need to see out of the costume
- Animatronics for tail, neck movement, head, eyes, mouth (the more, the merrier)
- Fingers need to be able to move and grab stuff
- Eligible for contests, because if I go ahead and make a costume, I want to flex it at at least one event.
It seems that this list is self-contradictory on all points.
Yes, even the last point. On the internet, ‘cosplay’ often gets used for ‘if vaguely related to a video game, movie, comics, et. al. then it’s cosplay’ — but that’s a common misconception. It’s only ‘cosplay’ if it’s a costume
from the cosplay region of France of an established character within a franchise — so if you’re cosplaying as your wow main, then that’s just an ordinary LARP. And if you’re cosplaying as an original character expecting to compete, find costume contests and their rules before making your cosplay.
In my case, NMN (Na Meji Nevidnega) appears to have allowed original characters in their 2023 costume contest. However, the winner of NMN cosplay contest is then sent to International Cosplay Competition, which doesn’t. Which means that there’s also a danger that NMN cosplay contest rules will drift closer to the ICC rules, meaning OCs will become haram here as well.
This is less than ideal, especially since my costume will take at least two years to make (judging by Akulva). However, if that happens, I may get a lucky break: Rytlock Brimstone looks close enough for me to pivot if I get enough advance warning.
Current Options, Ideas (& Problems)
Determining the exact solutions is what the R&D phase is for, but that doesn’t mean I don’t have any ideas as a starting point. Let’s jot those ideas down, so we’ll see what changed by the time we build.
The Costume Size
There’s two options for how big I want the costume to be. Option A is “about as big as the game allows,” which seems to be about 2.4 m01Depends on who you ask. According to measuring game models, 2.4m seems reasonable — more reasonable than some resources which say charr only go up to 6 ft 10 (just a bit over 2 m). Pros of this options are:
- since stilts are required, some of leg bend can probably be kept
- compensates for small penis
But cons are:
- stilts required
- preliminary research says I’ll have to go slightly beyond 2.4 meters if I want to keep my feet inside the costume knees, which is optional placement
- much wider than human body, meaning costume shoulders will probably be where my elbows are.
And then there’s option B: “about as small as the game allows,” which is about 180-190 cm. I’m 180 as well, so 180-190 cm tall costume would be a reasonable fit for me as well. The pros of this option are:
- can fit through doors
- probably won’t be as heavy
- no stilts
And the cons are:
- nobody will know the true size of my penis
- leg less bendier, though this isn’t that much of an issue if you’re cosplaying an engi
- still wider than human body, meaning there’s gonna be arm compromises
We’ll decide on that later.
But if memory serves me right, if we’ll need to pivot — Rytlock is on the bigger side.
This one is pretty much a solved issue for me (almost). The solution is water cooling. I have done water cooling before, and it goes something like this. You make a shirt that has tubes sewn on in. You want to cover as much of the surface with tubes in order for the cooling to be as effective as possible. According to my prototype (for a different thing) a year ago, a few rows of tubes across my chest offered an acceptable cooling performance.
The cooling tube was part vinyl tube, part bits of hollow aluminum shower curtain rod. Turns out that materials that can conduct heat well usually aren’t particularly bendy, and materials that can bend usually aren’t very good at conducting heat. At least if your budget only allows for cheap and accessible things.
Water was pumped from a 1L bottle, which required a change of water about once an hour at the nearest sink. This is going to be outsourced to a helper. The biggest problem is that water cooling system adds up to 1.5-2kg to the weight of the costume, depending on the size of the water tank.
Akulva cosplay uses aluminum wire for the skeleton. The problem with that is that it’s not very disassemblable. My plan is to model and 3D print out a 3D skeleton that can be assembled for cosplay or disassembled for storage.
The skeleton should also feature moving joints at least in the neck section, so it’s able to turn. Segments of the skeleton would resemble a real skeleton to at least some degree:
Springs — or way more likely, rubber bands (we have to be conscious about our weight here) would be holding the skeleton in a default position of looking straight ahead.
Having neck turn seems to be reasonably easy as well. The way to go seems to be having a fishing line run through all the rib bones in the section that we want to turn. The fishing line is tied to the last rib bone of the section, and a motor-powered winch is tied to the first rib bone of the section. As the winch pulls one of the wires and relaxes the other, the imbalance of the line lengths will cause the neck to turn. Rubber bands should help distribute the turning across the entire section.
The unsolved problems in this category are:
- Getting in and out of costume
Neck Turning system
The principle behind this is simple: attach one compass to the costume head. Attach a second compass to my head. Have an arduino or something compare the headings of both compasses and turn the neck until the measurements match. This way, I can control where I want the costume head to look by turning my head.
Though while the concept is easy, I expect that actual implementation — between calibrating all the sensors and the motors — is still going to take a fair bit of time.
Eye and mouth movements
If I’m doing animatronics, then the costume should be able to handle eye movements, as well as moving mouths while speaking. For mouth movements, I can — worst case scenario — try to copy the homework from the guy who did the Akulva cosplay featured earlier in this post. For eye movements, open source eye tracking solutions appear to exist, meaning I can probably get away without reinventing the wheel.
Seeing the world + finger movements
The problem of not seeing out of the costume can be solved through the magic of a small camera that’s being displayed on a screen inside of the costume. Fingers can be controlled mechanically — or I can build a solution that’s powered by electro motors, using some kind of sensors to sense whether and how open my fingers are.
On that front, Valve Index is basically a screen that you mount to your face — and it’s relatively light in weight. Valve index controllers have finger tracking that does NOT rely on base stations to work (according to quick research).
Valve Index has a problem, though. With Index, tracking works through the magic of base stations that the headset AND the controller have to see — otherwise tracking doesn’t work. Lack of tracking won’t be a problem for me, as I only require:
- the headset to act as a dumb display that shows stuff from my cameras
- finger tracking and button inputs from the controller
I don’t care about tracking — but SteamVR probably does. Thus, I’ll have to do some research into either finding a way to make steamVR work without tracking, or by making a fake base station that blasts the headset with fake pulses. Because I don’t want to spend €200+ on another VR/AR glassess from China to get a dumb display for the costume.
This way of doing things would probably require quite a bit of computer power. Fortunately for us, there are Raspberry Pi-sized boards with DisplayPort (required for Index) and an AMD socket, which is convenient (even though expensive). Failing that, Frameworks will also happily sell you a laptop motherboard with RAM and CPU, with a 3D print for the case that allows the motherboard to be used as a standalone PC.
And that’s it for now.
Akulva took 1.5 years from start to finish. Let’s see if I can get that finished in 3 years.