Initial Concept Art for Sinners Mire by SIMON KOPP is dope. It was more like a mood board than something to precisely duplicate. We knew for instance that, among other things, blockers would have another look and the playfield (the tile arrangement) was not the definitive one. Player 2’s view, among other things, was also something to experiment with.
I remember first starting to play with the look of the fractured tiles and to experiment with muddy waters. Slowly more and more of the concept art’s key elements were built until we had a first rough impression of what the map would look like both from player 1 and 2 cameras. At this point, concept artists stepped in and drew paintovers to bring me a more precise direction (and new cool stuff were brought in like the Temple)
Why I love Unreal Engine… The tech stuff is just too good with UE. Vines were added using splines and I used material’s world position offset to procedurally hang the moss in a random fashion. Water droplets and ripples were also added.
We spent a bit of time experimenting with the circuity looking structures we call “Chromite”. Main challenge was to make it readable and to look like there’s still active glowing chromite emitting light from inside the structures
Thanks to Art Director Josh Nadelberg for the great support and help through this process and to Senior Tech Artist Ira Goeddel for dealing with my poor UE knowledge at the time and overall lack of optimization. He drastically improved the lighting since then as well as many other things, so download Shardbound and check all that in-game
This is “Embassy Gardens”, the tutorial map, the second map I did for Shardbound. We went through quite a lot of iterations with that one. The initial concept art, although giving very cool ideas and strong overall guidelines, wasn’t as precise as for some other maps. Just like with Sinners Mire, once we had an overall idea of the map in 3d, concept artists stepped in to do paint-overs and give Josh and I more ideas to toy with.
Readability is a constant challenge and our main focus when we work on these maps because, in the end, all the player’s focus should and will be on the tiles and on his units. The surrounding environment is here to tell a story and beautify the picture, but shouldn’t never affect his ability to read the tiles. In the other hand, both the tiles and environment should look like they belong together so it’s a constant battle between making the tiles blend with and stand out from the environment. Trial & Error was the key to success.
We did not have to bother with player 2’s view since the map is for now only used for solo tutorials. That was good news hehe. Just like Sinners Mire, for the sake of getting the map in-game asap, background was left temporary empty until a matte-painting or 3d background is added later on.
The Ivy and foliage were quite interesting to do. Since most of the scene was quite basic regarding tricount and material complexity, I took the liberty to spend quite a bit of geometry on the foliage. I love how much that green pops!
It was the first time I toyed with Epic’s Pivots Painter nodes/script but the life it brings to the scene is definitely worth the trouble looking into it. Also added some falling leaves with a simple particle system.
Thanks L04D3D! Regarding being in the right section, hum my work on those is done and they all were released to the public. So I guess Released Projects makes sense? Anyway here’s the last map I did for Shardbound. HighLands
This is “Lost HighLands”, the third map I created for Shardbound, one I am incredibly proud of. When I first started working on it, I felt more confident already with Unreal Engine but I didn’t know yet that, to this day, this would become the hardest thing I had to work on.
The initial concept art for this environment, painted by Simon Kopp, is amazing. Although Art Director Josh Nadelberg and I first set goal to closely replicate it, we found through many many iterations and experimentations the need to tweaks things a bit to better emphasize the playfield. There’s also player2’s view to figure out which is always very worrying at first and, sooner or later, very troublesome. The composition has to work well for both players, looking from both sides. We allow ourselves a degree of freedom to cheat and move things around for a specific player to ease some of the pain through this tedious process… (for instance, we swapped the rings to balance the composition, which was the most drastic cheat we did so far).
The amount of details and vegetation this map brought was new to Shardbound and very very tough to translate and balance in 3D. We had to constantly work on keeping the playfield readability to a maximum and stay with-in that stylized look already established for the game. The grassy tiles were especially hard to figure out, needing a lot of details but remaining very subtle to keep the focus on units.
HighLands was challenging in ways that Sinners Mire and Embassy Gardens weren’t. The dark-moody lighting on Sinners Mire permitted us to not be perfect and hide mistakes, whereas Embassy Gardens, being in bright daylight, was so architectural that it wasn’t necessarily too hard to get a sharper look. HighLands needed to be very bright but look natural, sharp and clean. It also led us to be smart and find ways to bring shadows to the scene and separate foreground from background.
Through the making of HighLands and the previous two maps, I learned so much about Unreal Engine that I started to focus way more on profiling, performance and ways to create a project as clean as possible to send to my team-mate and senior tech artist Ira Goeddel. I also thought a lot about ways to bring life to the scene. Grass, plants and trees are all animated, structures are glowing from inside, generating particles and lighting. Flicker of lights are visible here and there, birds are flying in the distance and so on.
The main technical constraint we have for the tiles is that we must stick to only one material index per tile. So achieving the material for the grassy tiles, with grass blades, grass and dirt on them, was very interesting to figure out.
I used World Machine & Geoglyph to quickly generated both the top and bottom surface of the distant floating shards. Both top and bottom heightmaps were assembled into a mesh and cut into individual Shards, then their edges were further refined in Zbrush.
I’m using a lot of XY world projected textures so grass blades can share the exact same treatment as the ground they sit on, and blend very softly. Grassy tiles master material is blending between a dirt and grass material using a world projected mask. Grass blades are procedurally pushed in the ground to hide them where there is dirt. That way I could get away with using a single mesh for all grassy tiles. Even the deadzone (the four muddier tiles in the middle) are using the same mesh and same master material, there’s just a switch in the material affecting how the dirt material is processed.
Interactive events weren’t part of the original plan, but I always love to get technical. I love to code (although I’m bad at it) and to get to know how a program works. It was only a matter of time until I started fooling around with blueprints in Unreal Engine. It’s so fricking powerful. I added interactive events to the map in my freetime which the dev team loved and implemented. There’s a bird that you can call and send away, the ring can be clicked to show a pulse of energy and so on.
Just like previous maps, matte-paintings were supposed to be added later on to fill empty backgrounds. I gave it a rough go on my freetime and we agreed that the idea should be pushed further in 3D so we ended up making those floating shards. Everything rendered here is all using unlit materials for max performance. Clouds were quickly generated using volumetric in Blender in a mask texture, one channel storing the main light’s direction, another storing lighting coming from underneath to get that blue glow on clouds floating in space. Vertex painting is added on top of it to add a layer of control. Volumetric shadows are billboards matching each shard silhouette and using a cheap multiply material. Cloud’s shadows are also baked into the shards’ albedos. All of it is super cheap to render but the downside is that it must be fixed and it’s bound to our lighting. If we change it, we’re screwed.
Out of three maps I worked on, HighLands was by very far the hardest one but I loved every challenge it introduced and in the end I’m super proud of the result. I look back at it and already feels a bit of nostalgia. Such a good work experience. Again and again, thanks to Art Director Josh Nadelberg for the tremendous help!