Somewhere in the Midlands
14 September 2017by Craig Pearson
“How did they do that?” isn’t something you often ask yourself when looking at a Garry’s Mod screenshot. It’s generally obvious how it's done when you can see the joins. But I asked myself that when I came across Vioxtar’s work, because he doesn't build typical Garry's Mod contraptions. He builds worlds.
Asking myself wouldn't get me any answers, so I took an extra step and asked him. His answers are as detailed and impressive as the work he produces, and gave me a whole new perspective on what it takes to turn Garry's Mod into art. He's broke the game more times than you can imagine, lost sleep, and worked for months on a single scene trying to bring the sandbox to life. I hope you enjoy his story.
I’m Vioxtar, or Michael Efraim in the real world. Garry’s Mod addict since 2007.
I first got into posing/scenebuilding around 2009. I was 16. Garry held a competition looking for a main menu background for the new Garry’s Mod 12 release, and asked the Facepunch posing community to make one for him. I never thought much of Gmod’s posing branch before, as I was only ever interested in Sandbox building, but I happened to stumble into the screenshots section and saw Garry’s thread. The idea of having my picture as Gmod’s official background lured me in. I posted three submissions, which were my first ever poses.
I didn’t expect much. The thread was packed with talented people posting hundreds of amazing works with spectacular effects, lighting and of a quality I had no idea how to achieve. I had a lot to learn, but I guess Garry found what he was looking for in my last submission and, to my surprise, chose it as the winning picture. The instant gratification of recognition was amazing, and I was immediately hooked. Ever since I’ve become increasingly involved in the posing community, making more pictures and trying to constantly improve. Today, posing in Gmod is my number one creative outlet.
I throw props around in Garry’s Mod to see what works. Sometimes I find a model that inspires me. Sometimes it’s someone else’s artwork. Sometimes an idea pops into my head and I list it down and come back to it later. Sometimes I find that some background I made in some picture makes for an interesting concept by itself, and I want to explore further. Sometimes I have a story I want to tell.
When I have an idea, I try to think what composition would show it best. I spend extra time trying to figure out which angle best captures it, what zoom level works with it, and what type of lighting would highlight it best. I care about perception in a picture, it’s my number one guideline. A good idea needs to come across to its audience as best as it can, so you want it to be percievable. I’m usually only equipped with a single frame, and I want to make it count.
I always find myself putting extra effort into worldbuilding. Over time I began gravitating towards building my own universe within the realm of Garry’s Mod: a reimagination of the Garry’s Mod Sandbox gamemode. I started asking myself questions: "What would it be like if there was a world out there populated by players with working Physguns, Gravguns, and Toolguns, and the ability to spawn whatever they pleased?" I began answering these questions with pictures, ‘captures’ from that world, belonging to my ‘Server Dust’ series.
This world started to take shape with every picture I put out, and I found myself submerged in a universe that excited me, that sparked my imagination. It was unpredictable, and chaotic, and expressive. I could babble on about it endlessly, but in short: it was the perfect canvas I could project myself onto. I try to describe the world and tell its story as much as every picture allows me to, and to provide insight into the mentality of it through its scenery. So I guess sometimes people might miss the bigger picture, the puzzle holding all these pieces together.
Definitely Physgod. Another scenebuilder (Crazy Knife) gave me the (epic) idea of building a giant head made out of props. We brainstormed a bit and I worked on it for two straight days over a weekend. Every prop I spawned fell into place, and the creativity flowed out of me. It was fast, simple work that turned out effective, and everyone loved it. Putting together an artwork like that is extremely fun but rare. I usually feel weighed down by the complexity I force myself to put into every picture, and the obstacles stack up fast and hurt workflow, and ultimately, make posing less fun. This wasn’t the case with Physgod.
The busiest picture I worked on was a picture I never published, because I never finished it. It mainly consisted of failures, but it was the seed of my progress, and launched me forward. I called it 'Spawn Point Zero', and it doesn’t come without a story. On August 2014 I published what was my biggest picture at the time, ‘Rudolf the Gmod Wizard’.
I worked on it for five months. It had huge amounts of detail, and was packed with all the effects I could fit in it. It received a lot of attention on the Facepunch forums. Garry saw it and tweeted about it, and you ended up printing several copies of it, hanging one in your studio, and sending me some of the prints as well. The amount of recognition I got for it was astounding. I wanted more of that recognition, so I attempted to go for something way bigger, and started working on Spawn Point Zero.
Spawn Point Zero was set in an abandoned Garry’s Mod sandbox server. It was supposed to portray what an average sandbox server would look like after years of not being reset. I opened up a document and started writing a list of ideas of what I wanted in the picture. I also drew a rough sketch of the “final” composition I wanted:
A wrecked contraption in the foreground, and a couple of more in the background. I started building, crossing off ideas from my to-do list with every new contraption I added into the scene. When I was done with the first layers, I realized there was still a lot of empty space left, and suddenly my ‘amazing’ composition felt insufficient. So I added more layers.
It was going to be a beautiful, gigantic render packed with awesome effects, top tier lighting and a story that tied every layer together with its neighbors in a scene you could lose yourself in. Every ‘layer’ would be an entire scenebuild of its own that you could zoom into, with each layer was filled with Human Players interacting with each other in crazy ways, causing complete and utter chaos.
With every idea I marked off my to-do list, I ended up adding five new ones. What started out as a relatively simple composition with ~500 props blew out of proportion into a composition that easily exceeded ~3000 props. I found myself a victim of my own idea. There were nights when I couldn’t sleep because my mind was still iterating ideas on what to add next. I was way too deep into it. It needed to be great, it needed to be mind-blowing… how else was I going to get the recognition I wanted? I knew it was silly, but I kept going because at this point I couldn’t stop.
I hit engine limits. It got to the point where I couldn’t save the game anymore, because the entity count would cause it to crash, so I had to rethink my strategy. I realized the only way I was going to achieve what I wanted was to separate every layer into its own render, and then stitch it all together in Photoshop. I then realized that if I was already rendering each layer independently, I might as well give each layer fancier lighting. So that’s what I did.
I knew lighting was important, and I knew it needed to look great, but I had no idea how to actually achieve lighting that looked good. Before, lamps weren’t an option: Source had an eight lamp limit, and there was no way I could spread eight single lamps on an entire map. But now that every layer could be lit separately, using lamps sounded more feasible. I loaded up a relatively small ‘catwalk’ environment I built, and tried to toy around with lamps and lights. I tried different things.
No matter what I did, it didn’t look convincing. Something was absent. I couldn’t settle, because there was no way I was going to compromise on something so important as lighting after so much work. So I kept trying different things. I then asked myself what would happen if I simply rendered the entire scene in darkness with the exception of a single shadow casting lamp, each time from a different angle, so I tried it out:
At first I didn’t fully realize why it worked; what made the picture look so pleasing? But I knew I was definitely onto something, and immediately tried it out on the different builds I already had:
By this point I realized I had re-discovered the core of 3D light rendering techniques - the idea of stacking different light passes together to replicate how light behaves in the real world. For the first time I was getting well defined, and convincing shadows, and suddenly all the details I worked so hard to put into my scenes, popped out. The world I built realized, and my contraptions came to life. I was in love with what I had discovered.
Once you go down that path, you’re bound to start thinking of rendering scenes differently. I later realized I could combine the technique similarly not just with light information, but with depth information. By rendering a depth map, I could add fog and sky into my renders.
Or that I could use the power of Photoshop’s layer blending to add textures to a scene by rendering texture passes.
Or create isolation passes and use those to change the base colors of props to my liking at ease:
By this point I was completely distracted from Spawn Point Zero. I discovered something I thought was amazing, and wanted to practice it ASAP. Having a lot more work left to do on the project’s scene building aspect before I could get to rendering, I started working on more immediate projects without ever finishing it. I ended up improving in my overall scene building techniques and styles, enough so that Spawn Point Zero’s scene builds were no longer adequate, and at some undefined point I dropped the entire project.
I may have never finished it, but it rewarded me like no other I've made, and I felt it was important to bring up. Spawn Point Zero was what launched me forward in my rendering techniques (which I still try to perfect), and it was the ‘big bang’ of the Sandbox Universe I still focus on today. I’m still marking off ideas from that same to-do list.
This is where Spawn Point Zero ended up.
The most demanding picture I made was Path of Spawn.
It involved 329 effects, 679 props, and 116 individually posed ragdolls that were iterated upon several times over the course of seven months. The in-game scene building took me around five months, the editing took me around two more. I barely made any compromises, and the ones I did make were because of engine limits. I used every trick I knew back then to get the most out of the picture, and as a result the project was extremely demanding, technically speaking. Material blending, atmospheric passes, direct and ambient light stacking passes (holding light information equivalent to hundreds of thousands of lamps) all rendered and outputted to 25600x14400 raw captures which I later had to manually--and very slowly--process in Photoshop. I definitely bit more than I could chew with Path of Spawn, but I followed through with it till the end. And it was still mostly fun.
Yeah. My first poses were in existing maps, but soon enough they were of an insufficient quality and variety. Maps are optimized for gameplay purposes, so compromises are made when it comes to polygon count and textures. They were good to start with, but I had to abandon them if I ever wanted to step up my scene building game. Being dependent on maps means less control over your artwork. And control is important.
In my opinion, making good, convincing ground is an art of its own. It requires different skills than those of ragdoll posing or contraption building. It’s extremely demanding, technically, because it’s almost all in the geometry and polygon count. The believability of a picture depends on it.
For me, the ground, rocks, trees, and overall scenery is just as important, if not more so, than what would otherwise be the subject of the every picture. In my eyes, a well laid-out environment can portray more character and personality of a world than any other element could. It tells a story, and immerses you in it.
I like hiding a bunch of my V logos painted around in the scene, sometimes for watermarking purposes and sometimes just for fun. I also hid a drawing of a penis once in a picture that ended up in the old Garry’s Mod 13 backgrounds for everyone to find. That was humorous.
I’m not sure. I don’t think I have a single most important addon. Every addon is almost of equal importance, from model packs to tools. But I do know I can’t go without my posing related toolset. The first tool would probably be the advanced camera tool. It lets me position a static camera with all the parameters I need to define my frame. The advanced camera tool also lets me build my scene while looking through the final frame. It comes with other great features as well that hugely improve workflow. Then the more conventional tools come in, like prop/effect resizers, the stacker tool, advanced duplicator, ragdoll poser are all tools which let me build the scene. After that, probably the material creator too. It gives me complete control over any prop’s material. This one is important, because good material management is necessary if you want even more control and variety in your pictures. After that, soft lamps. The tool is still in perpetual beta, but it streamlines a good deal of the rendering process, and it’s good enough. Soft lamps essentially wrap the poster command and utilize it to output renders of soft light passes, by stacking a number of neighboring projected textures. A good picture has good lighting in it, no exceptions, and good lighting never comes easily. Soft lamps help with that.
I ask myself the same question a lot. I feel comfortable with Gmod. I know my way around its tools and menus. It’s still a game, and running around the scene in first person with WASD controls, physics, and guns makes me immersed in the scenes I create.
I think the most important advantage Gmod currently has over other art packages is the availability of Lua. Being able to code my own tools to create my own workflow is super important for me. If a tool I wanted didn’t exist, I made it, or had it made for me. Thanks to Lua, I can now render soft lights, light diffusion, and reflection passes, as well as control materials and textures like never before, and improve any scene building related workflow as I please.
I also enjoy the challenges Gmod imposes us with. The Source engine always fights back. I like stretching its limits. I like seeing how real I can get my pictures to look with the 2004 graphics.
When I finish a scenebuild, and I’ve decided on my frame, there is a fixed number of steps I take to render it: light, sky & atmosphere, textures, and final touches. I won’t go into depths (for that I wrote a lengthy guide), but the general idea is to take as much information as you can from the scene, and manipulate it in an image editing software like Photoshop. This ‘information’ is essentially different passes: light passes, texture passes, depth passes, reflection passes, etc. Where each pass is a separate render. The more you separate passes from each other, the more control you have over your work post-game, and the more you can fine tune it to your liking. Each pass has its own role, and eventually all these passes are merged into a final product. That means that if each pass is well rendered, so will the end result be. It’s more work, but with the right tools you can streamline the process, and in my opinion it’s definitely worth it.
Just 53 GB.
Thank you to Vioxtar for the illuminating responses. I thought your work spoke for itself, but hearing what you go through made me appreciate it in a completely different light. You're a real asset to the community.
I hope to do a few more of these. There are others out there working on scenes, add-ons, game modes, and more who have different stories to tell, both in game and in the real world. I have a list, and I hope to bring you more of those stories in the coming months. Thanks for reading.
If you want to follow this project you can sign up to the mailing list.
We'll only update you about this project, we won't spam you about other stuff or sell your email address.