Hello ! I am a video game student @ILOI & I am very thankful, your speech is very motivating .
Except the dude clearly doesn't know much of anything about the 3D game pipeline. Yeah, if you're very skilled, a high poly sculpt could, certainly. But then there's retopology, UV mapping, texture baking, rigging, animating, other means of optimization once imported into the engine. Granted it wouldn't take anywhere near the production time of a AAA character (Which the High-poly sculpt took maybe 10-15 hours altogether, but the finished character took ~94 hours). And granted pokemon models aren't nearly as complex as that, but I think at least a 1-3 hours from start to finish to be a fair average expectancy of artists who know the work flow well enough. I just hate how people are so critical of artists when they clearly don't understand what goes into it.
GNOMON student and excellent 3d artist Christen Smith from Lightstorm Entertainment talked about the production of awesome 3d environments. Christen talked about asset production, material creation, lighting and his indie-project, created with the help of Unreal Engine 4.
Hey all, I’m Christen Smith. I’m an environment artist at Lightstorm Entertainment, and an indie game developer working on a slasher/horror IP in Unreal Engine 4. I’m originally from Dallas, Texas, and moved to Hollywood in 2012 to attend the 3 year program at Gnomon.
This is the living room in my game Limb Chopper, the game I mentioned working on earlier. It’s still very early on, but here you can kind of get an idea of my layout process when I’m starting to put together a level. Even in a game, I try to establish lighting and composition very early on (after the level is laid out and tested), and I can then build upon that and carry it over into the other levels. I love UE4 all around, it’s very artist friendly and the PBR workflow within it is phenomenal.
I quickly gravitated toward environment art. When it was time to pick a track, I went the Generalist route, loving the idea of being at least competent in multiple discliplines like dynamics, animation, etc. Being that the program has a huge modeling base attached to it, I kind of stuck with that as my primary skill. I think it’s mostly due to luck, but when I graduated, I landed a contract job for a couple of months at MPC LA, where I contributed some assets to The Martian VR Experience.
From there, it slowed down a bit, and I worked freelance for a couple of months before the HR from Naughty Dog contacted me about my portfolio, scheduling an interview with Teagan Morrison and a couple of others from the environment team. I was fortunate enough to help out during crunch time for Uncharted 4: A Thief’s End, just trying to do my best and leave a positive impression upon the company. My school Gnomon has a really cool event that it does during the breaks called Employer Preview day, to which I attended. A teacher (I can’t disclose the person’s name, by request, but that person is totally rad and I owe them a steak dinner) took a business card to bring to Lightstorm.
This is my process for dissecting a concept piece given to me and itemizing it. I break down everything I can to establish my plan of attack.
I think it’s all about telling an immersive story, whatever the context. If you’re making an environment for a game, you’re always answering “why is the player character here? What happened here? Who frequents this place? Who sat in that chair? What is the mood?” Things like that. From Naughty Dog, it kind of gets driven home at every Monday morning meeting “Is it awesome yet? No? Make it awesome.” We’re in the business of entertaining, at the end of the day, and I think that if you give people something commonplace or ordinary, your viewers will be disappointed. We have real life for that.
Story is king in any environment, so I’d place that at the very top of the list. I would say that nailing the composition and lighting early on is critical to the piece’s success. Thinking in terms of a cinematographer or director, you’re capturing the story single-handedly with these two factors, and all the subsequent elements layer on top of that.
Telling Stories in Virtual Spaces
First, an environment contains the remains or remnants of whatever or whomever occupied that spot before, be it dropped items on the floor, fond memories, a collapsed wall with the remains of someone’s home underneath, things like that. You can place hundreds of things into an environment to convey this or lead the audience down the rabbit hole of what you’re trying to convey. I feel like enviroments are the most all-encompassing and immersive, especially in games (although there are some awesome character and creatures out there).
This is what a typical reference collage of mine looks like before I start modeling for any given piece.
Playing an environment-rich game like Dragon Age: Inquisition kind of reinforces that idea. I remember when I got the game, kind of nerding out over how lush and impressive the Hinderlands were, or more recently in playing Dark Souls 3. The environments are all-encompassing and omnipresent, they surround the player with their vastness.
Complex Environment Creation
It really depends on the origin of the piece, whether it’s a concept given to me by the concept artist, or if it’s my own. Universally, the process always starts with me opening the concept in Photoshop and spending some time just staring at it, deriving from it what I feel the meaning of it is, or what the artist is trying to convey. Is he showing me the lonely emptiness of the quiet room? That sword sticking out of the tree stump, and the knight that left it there? Reference is also critical. A big saying at Gnomon is that “an artist is only as good as his reference.” Learn to gather reference, and make collages.
When I find a piece from a collage that I’m definitely going to use, I can instantly see how I would go about modeling it. Since the rest of the scene is so busy, unfortunately, I had to somewhat strip down some of the interesting harnesses and buckles that made this chair so cool, but in doing so, give the eye a resting place in a busy composition.
I think the big two things that you MUST always have in your scene working from the very beginning are composition and lighting, which ultimately sets the mood. Learn the golden mean, or the Rule of Thirds. The test to this is during the block in phase. If you can manage to make a frame full of polySpheres and polyCubes look immersive and fascinating with good composition and lighting, then you’ve already won. From there, it’s just a matter of replacing those block-in shapes with well modeled and textured assets.
The Deadspace Cockpit was actually made during my time in the demo reel portion of the 3 year program at Gnomon. This is a screen grab of Alex Alvarez‘s critique on the cockpit, still under construction, in which he gives the students individual notes on their work. It was derived that the piece had a few layout issues which needed adjusting, and I also noticed some inbalances and empty or awkward spaces.
In terms of hard surface assets, I rarely or never build within the master scene, that’s just where everything ends up for the final comp. I typically make collages of just parts of machines or anything that would be useful in informing me of what to make, and start building a library of parts, or a kit. I establish early exactly what objects are in my scene and how many of each, then go to work on modeling the assets. As you model more and more environments, save the assets that you think will be useful for a future piece and save it along with it’s textures and shaders, masks, etc.
This was the resulting model, shown in wireframe. At this point, I see a good balance of areas of detail and areas of rest, and everything is feeling much better. Once this composition blocking is complete, I immediately begin to light the scene, along with atmospherics, and start to set up my render layers: Diffuse, Specular, Reflective, Atmospheric, ZDepth, AO, and a couple of others. In Vray, I can immediately switch between the different maps within the frame buffer and monitor their quality.
As for any metallic materials that I create, I have to first give some props to my Hard Surface Texturing teacher from Gnomon, Eric Miller. In the class, we spent a lot of time learning the properties of many different types of metals, and replicate them in Maya. With metals, know fundamental things like the diffuse of most metals other than gold is black, and that what you’re seeing as far as “reflected color” is from the Specularity, or Roughness if referring to a PBR workflow. Metallic surfaces, by nature, won’t have quite as much bump information from a normal or bump map, given that it takes a lot more to alter the surface. You really want to keep bump information, as a result, to a minimum.
I use a PBR workflow for both Unreal Engine 4 and Vray in Maya. I’ve found that the Substance package is extremely powerful for seeing the effects of all your maps at once, lit in HDRI lighting.
Take some time just making a lighted sphere in Maya with a backdrop, throwing an uber shader on it, and trying to replicate different metals. Put some good time and practice into learning what every different property is doing. Focus mainly on things like diffuse, reflective glossiness, IOR, reflectivity, and specularity.
I build libraries, or kits, or hard surface parts, as shown in this image. I translate this directly across every environment I do. When it comes time to place assets into the scene, I like pulling from a pre-modeled collection, so my mind is strictly on scene layout at that point. This kit was mainly built ahead of time by myself for use on the Deadspace Cockpit, but is later stored in a library, since I found several reusable pieces.
As far as post processing stuff, I do tend to spend a bit of time in Nuke after getting all the render passes from Vray. I use Unreal Engine 4, so I very much switch to a PBR workflow if building a level for a game. For that, I throw in a Post Processing Volume and can achieve almost the same results as though I was compositing the level in Nuke.
Hmmm. I would say that a bad environment is simply an environment that doesn’t capture the imagination or tell a story effectively. Bring people back to that period of their lives where they’re a 12 year old kid again. If your environment is photorealistic, but doesn’t grab people or compel them to check it out for more than a cursory glance, then I feel that you haven’t done your job as an environment artist.
I spend a lot of time at the end in post, trying to capture fully the mood of the concept, and find opportunities to add a touch of my own aesthetic if permitted. In Nuke, I adjust the different maps derived from Maya.