Gart230 Environment Blog
Week 1 Idea Planning & Research
We have been set a task to create an environment scene based around either a movie or series. This allows for total creative freedom and a chance to work to strengths, or to set some challenges! Through brainstorming cinematography this week I have streamlined down to a few thoughts which I will get to in a second. I started with these movies and series as candidates to create:
- Samurai Jack (Feudal Japan – Heavily Stylized)
Pros: simple texturing, strong read of characters, clever colours portray personalities, unique style and scalable.
Cons: 2D art to 3D render, plain environment makes for boring read and unambitious creation, would have to hand texture.
- The Expanse (Spaceship – Futuristic Realism)
Pros: unanimous colour scheme, chance to show off texturing, hard surfaces, very tillable.
Cons: creation of a whole ship too ambitious for time frame, making 30 second show-real difficult under limited modelling conditions/created environment.
- Pan’s Labyrinth (Dark Garden – Mythical Surrealism)
Pros: unique piece for portfolio, lighting and art direction are very good to work from, easily adaptable for style of work.
Cons: character creation doesn’t work to my strengths, hard to find references for environment backdrop as limited to heavy fantasy surrealism.
- Into the Spider-verse (Cityscape – Conceptualized Comic)
Pros: art style plays to my strengths in texturing and would be ace with chromatic aberration, tilable and roads/pavements/buildings and reusable block assets.
Cons: over-commitment with a whole city block within time frame, would require many texture sheets and passes, scene might look empty.
- The Book of Eli (Wasteland – Post Apocalyptic Realism)
Pros: easy simplified background creation, high to low poly and texture resolution assets based on distance from scene implementable, sepia black and white simplifies texturing and highlights contrast of important assets to non-important.
Cons: barren with little props to demonstrate environment art abilities, complete realism hard to master, difficult to add world borders to refrain player as so flat, little bit uninspired.
- Love, Death, Robots (Stylized Animation)
Pros: easily adaptable to any art style or creative approach due to animation shorts, lots of visual prowess to work from, pop colours allow exciting texturing.
Cons: common theme among peers therefore difficult to stand out, hard to approach as many plots and styles of work within itself, lots of variant assets to create under limited time.
- The Dark Tower (Dystopian Fantasy)
Pros: would be confined to a room making viable for project, opportunities for illustration design for signs and magical assets, mute dark tones add atmosphere.
Cons: need character(s) to add interest to the room, architecture of realistic non stylized building would be difficult, little room for playful particle effects or physics from the game engine.
- Avatar (Atmospheric Otherworldly Surrealism)
Pros: allows for playfulness with game engine experimental features like particle physics, really vivid colour scheme, scalable if within cave etc, minimal plants replaceable, plays strength to my art style, feel motivated as looks cool, scalable.
Cons: organic environment challenging to create in keeping with a natural feel, randomness allows for mistakes, 3D sculpt of plants and trees hard to unwrap good uv’s and topologise, opacity leaves potentially cause render issues.
‘Avatar’ and ‘Love, Death, Robots’ ideas are coming off as the strongest contenders for me simply because of the beauty and diversity of conceptualized assets I could create, and how I could playfully experiment with in engine effects. They are the strongest likeness to my art style whilst keeping a healthy, yet not overly ambitious challenge for my 3D environment project. The biggest downside to Love, Death, Robots is deciding upon one of the several completely different, and I mean in every aspect, animations from the series, they’re all so cool yet are so short portraying little reference. The biggest downside to an Avatar world is how I’d overcome the many natural complications that come with the world, yet it’s strengths far out way the cons, for example due to the cult popularity to the movie and award winning cinematography there are dozens of references and articles upon the art, even a book ‘The Art of Avatar.’ Hence my reasoning to peruse this world as my environment project. The actual movie which I am going to watch and refer timestamps has little small scenes, giving me the idea to create a cave system, allowing scalability to this time sensitive project, but I will explore concepts later. To start with I have created a moodboard exploring colours, real life sources like jellyfish, aloe vera plants and sakura trees to get a sense of what everything should look like and to adapt that to surrealism, along with the ‘Zhangjiajie’ mountains of China that inspired the floating islands in Avatar, to the glowing plankton of Cambodia island ‘Koh-Rong.’ Furthermore I’ve included some art station concept artists that are inspiring for this project, namely DongbiaoLu and Nicholas Theophilus.
Week 2&3 Concepting and Further Research
In order to establish an overall artistic design that works well as a single environment I have decided to develop a stylistic art approach as a realism method will take too long for the time constraints of the project. That being said it is important to use real life influences for my assets and to get a feel for how to adapt them into my surreal stylistic approach. Here is a contact sheet of some pictures I have taken in and around gardens of Cornwall for this project, specifically looking at natural plants and trees. I have also created light paintings as some specific research on emmisive lighting using long shutter speeds in the dark with some lights.
In addition I have also collated some specific research detailing the shapes and curves of trees and up close photo analysis of the textures and elements of trees. Moreover some photos detailing water flow and physics and a jellyfish: the most similar real creature (Scyphozoa) to that of the flying creatures of the avatar movie. This is all to help with my asset sculpting, texturing and particle based water features.
The next important step is to develop a layout and map of how my avatar world will look and feel and play out if it were to be implemented as a game level. I’ve started to blockout this in UE4 using BSP’s. As it has been drilled into my head last year I’ve been sure to turn on grid snapping to 100 as ‘grid is king’ and have worked in BSP multiple sizes of 200.
In this photo I have play-tested this element and have realized that a player would be able to fall out of the world through here, however I will demonstrate later in a video how I have fixed that by simply lowering a fallen tree block.
In this early iteration of the blockout I have asked for external feedback in which I have been told that the area seems too circular and unnatural. In addition to this the curving ramp up seems to lead to nowhere, and the tree seems too big and crams the room. I have considered these points and want to expand the cave out into another section to the left which will also create a tunnel from there to the top of the ramp, and will decrease the size of the tree.
As I previously stated I wanted to experiment with in-engine tools and thought the best way to do this would be to attempt a particle physics based waterfall. I have considered where in the design this would work and have created the blockout around that, I figured it would be really cool to have a half blocked cave entrance with this, which would reveal glowing crystals behind it and would make a great camera shot, whilst allowing the player to walk into the cave next to it. The water would logically flow out where I’ve put the fallen tree block to stop players leaving the space.
Here is a mostly completed block out allowing visualization of the environment and to plan out what props and how many things I need to fit the space. Due to this being a cave system it allows for scalability within the design.
Week 4&5 Asset Making and Art Style
From analyzing the blockout I’ve put together a list of some assets that would be customizable within engine via size and hue shifting that can be replicated throughout the environment. I aim to work on prototypes in order;
- Mud hut
- Crystal clump
- Light pole
- Various rocks
- Various big plants
- Jellyfish-like creature
Being a bit rusty with the software Maya my first hut prototype will probably not be permanent as it has developed some issues, including a weird lighting bake error in Substance Painter and the unwrap is particularly unclean compared to what I would have liked it to be. This is something I will potentially come back to later on dependent on how the project develops, however I will not get stuck on it now. In terms of texturing I was quite happy with the simplicity of it, considering it is meant to reflect a shoddy mud hut.
Real crystals have sometimes been known to be a bit opaque in nature, this will be easier as a workflow to demonstrate later on in engine rather than through the texturing. However I do want to demonstrate a glowing emission which I attempted to add in Substance as shown. The sculpt was super easy and therefore low poly, I suspect that might be different when I create the clump as it will have a lot more points and be a bit messier which I will have to take into consideration. This asset is the perfect example how I can replicate one model and one single set of textures throughout the cave system in different colours and sizes that’ll create a unique feel for each individual crystal and tie the cave system together.
A NEW WORKFLOW – Mushroom
This asset involved learning a whole new fantastic workflow for me, I wondered how to create a low poly but natural looking mushroom? I have attached a workflow sheet below of how I achieved that. Essentially this workflow involved quickly making a very high poly and completely un-topologised mushroom in ZBrush, generating the form and details I wanted to achieve, this took 5 minutes. Then porting it over to Maya as a cached file. Then making the object live I used a tool called ‘quad draw’ which I’ve not used before to manually reconstruct the mushroom as a low poly neatly topologised model, which then get ported back into zbrush over the top of the original. I then subdivide the low poly once and project the bad high poly onto it, and repeat monitoring that the topology is still good until I’m happy with the smoothness, form and poly count of the new high poly model. Delete the original mushroom, export the new back into maya. Then It’s back to the basics I learnt in first year; unwrap, delete history, pivot points for placement in engine, freeze, and exporting both versions of the model for baking and texturing! Overall I am very happy about learning this workflow as it will be infinitely easier to create natural models in the future of the project.
The final baked and textured mushroom is 800 verts and 1.5k faces, however it needs to be a decent size as the plan is to have a varying scale from small to very big in the scene.
I’m working with a coursemate Zion Love to create this environment who has sculpted a feature tree, in combination with my mushroom and hut the crystal has become to look out of place. Upon feedback from my game team we have come to the conclusion that the texture needs to be reworked to become simpler and to take the emmision out as it’d be better off left for the shader in UE4. I’m happy with the new look of the crystal I’ve made as it looks like it belongs with the other assets now.
The crystal clump turned out really well I think as the texture closely mimics the original crystal and the random variants and sharp angles look really crisp. The modular bottom makes it easily blendable into the ground or walls it’ll be placed.
Unfortunately time has gotten the better of me over these two weeks with working on the other modules and have not been able to complete any more assets however currently I have plenty to experiment with and move about the environment.
Week 6 Walls and Rocks
A big part of my scene is simply the walls and rocks of the cave, therefore it is important to find a quick way of putting all this together. The basic idea was to create a few shapes and sizes of rocks to stretch about and duplicate around, and to create design consistency like trim, and to create bits of walls that are low poly and modular in construction to attach to each other. The first step was to create a smart material to drag and drop onto all rocks that looks really good no matter the shape etc.
This week is pretty boring but the aim is to create loads of rocks and walls to completely cover the blockout inside the cave at it’s basic form to start week 7 adding details and making it flow. I’ll just include a few screenshots of progression for this week.
As much as possible I’ve implemented the assets with a clear tiling pattern to minimize draw call use and optimize the ability to edit the scene effectively. I’ve playtested to delete any unnecessary elements and make sure all viewpoints have clean and polished results.
I decided upon creating a character for this scene to better relate it to the Avatar world. The Na’Vi is a species within the universe which I want to really demonstrate in this scene. Henceforth I will be making some environmental storytelling props like cave paintings and the like.
Week 7 Particles and Mists
I’ve decided to delve into the UE4 particle system, I would like to create firefly and a waterfall/mistfall effects. I used this tutorial for the fireflies to get a basic outline of how the UE4 system works for this stuff. Thereafter I experimented on my own to create the mist. I found that the more I used the tool the more I got used to what means what and how so many effects can be implemented. I found the refraction tool which I plan to create a water material with. Here is the firefly outcome. With some nitpicking I can see this being a really powerful part of the environment, adding movement and life.
The mist was much harder to work out as when it came to the particle emitter it didn’t seem to want to work with normal gravity, randomly trailing off to a random direction, however after an hour trying to fix this I admit to a shortcut which was simply to rotate the emitter in scene and lower the lifetime, which worked just fine!
Week 8&9 Landscaping and Materials
In order to create the terrain I’ve used the in-engine landscape mode which uses a heightmap and some sculpt tools to manipulate it. I am very pleased with the results and a starter material for moss works really well with some saturation and brightness editing. I tried to make the terrain nice and smooth where the player would walk and bump up at the edges to blend in better with the rocks and other hard surfaces. Furthermore I had to mess around a lot with the collision and make sure there were ramps in areas so players would not get stuck anywhere. During this time I also began to move around the crystals throughout the cave.
The water became increasingly complicated and the refraction could definitely use some work! However the overall results are great. I used a moving river tutorial as a reference and optimized and changed elements to fit my scene, such as making it more see through for our very shallow pools of water emanated from the mist.
Week 10 Assets, cave-paintings and post process
Firstly, I have bashed together a cubemap for a spacey sky sphere. I hand painted assets and used a software called ‘Spacescape‘ to compile them into a seamless sky.
In addition to this I have created various 2D environment art as cavepaintings on the walls to add an element of storytelling and add more compelling colours and pop to the scene. Some of these are black and white as when it came to making the decal actors I needed to create 3 versions for opacity, emmisive and base colour.
New assets! The scene seemed a bit bare so I created a couple floating seeds as seen in my moodboard research and added some low poly background trees. For design consistency I made another smart material to add for the trim tree I designed at the front of the tunnel. In Zbrush I’ve learnt many tools to optimize my topology such as projection and how to correctly use the ZRemesh tool.
I’ve also polished some pathway tiles I quickly made to further direct the scene and guide the player around a specific path, I found that you couldn’t see them amidst the grass so I manually made an emmisive opacity map in Photoshop, I also did this for the crystals giving them the extra glow. Furthermore, I’ve designed what blocked the player jumping out of the map at the water to put a fallen rock which I believe works really nicely within the scene and storytelling.
The final polish was to add post process effects and colour grading, I decided to manually colour grade eyeballing everything with a couple people around for that external viewpoint, although it is important to note that it is possible to do this within Photoshop also. I’ve added some bloom to really make colours pop but not too much to overwhelm the other features and put some glare so fire and assets can really stand out in the cinematic. Speaking of which I filmed using the in-engine cinematic mode and camera’s, and edited this in Premiere, which I will include at the end. It is satisfying to compare this video to the blockout and how closely I’ve followed the original design. In conclusion the project has been successful and the workload has been split fairly and equally. There are optimizations to make and little improvements to make such as the decals actors sometimes projecting on props that it is not meant to. In addition to lowering object draw calls and more LOD’s.
Sincerely thank you for reading the blog and watching the project unfold and develop and I hope the final result is to your liking, if you have any queries or questions don’t hesitate to get in to contact with me.
-Cameron Jago Lis