Building Stylized Scenes With Lights & Colors in UE4

Have a look at a brilliant breakdown by Peter Tran that shows some of the ways you can use smart lighting scenarios and other tricks to set up stunning stylized scenes. 

Introduction

My name is Peter Tran, and I’m an aspiring lighting artist for games. I recently studied 3D animation and CGI at Collège de Bois-de-Boulogne in Montréal, Canada. Currently, whenever I have free time, I freelance at NOWWA as a lighting artist. We are working on a stylized multiplayer hero shooter game called BulletVille.

A little bit about my background, long story short I’ve had my sight dead set on becoming an environment concept artist. I was mostly self-taught with online contents such as Feng Zhu‘s youtube channel and many others. Soon after starting college, I found my true calling as a 3D artist and eventually I found out about the role of lighting artist which really piqued my curiosity. I’ve never stopped trying to learn more and master my craft as a lighting artist since then. I firmly believe that lighting is the most crucial part of the pipeline to showcase the maximum potential of teamwork. Even now, I still try to bring out as much as possible my love for concept art into my personal lighting projects.

Memory Lane project

Usually, I never put school assignments into my portfolio. This is a rule of thumb that I’ve learned throughout my few years at college and from speaking with artists from the industry. The reason is that if it’s a school assignment, chances are there will be someone else who from your school will put that same assignment in their portfolio. It would totally scream “Hey! Look at what we’ve done in school! This is an AK47 that we’ve all done, this is a brick shader that we’ve all done, and this is a sculpt of a head bust that we’ve all done!” To put it simply, school assignments aren’t original. It would be better if you dedicated your projects for as a unique portfolio piece. However, in this case, I had a lot of liberty and decision as for what I want to do; thus I did my absolute best to make it shine in my portfolio. For our final project of the major, we had to do a short cinematic, in my case we were in a team of three: me as the environment artist, Daniel Ferland as the character artist (and amazing rigger) and Brandon-Midge Innocent our Animator (and also a music producer!)

Pre-production

It’s most likely because I had such a significant interest at the beginning for concept art that I think the pre-production phase is so important. Well-Planned pre-production can go such a long way in the end. Be decisive and think of what you want to spend a good chunk of your energy and time on. Working on something that you’re not interested in is the perfect recipe to feel like you’re wasting your time and motivation. I also heavily recommend anyone to document your work in progress from the start to the end. I can’t stress it enough; it helps your morale a lot to see what you’ve achieved so far when you feel like nothing is going well.

First of all, we started with brainstorming what kind of cool environment we could use for our short. At this point we had already decided on our story but not so much for our mood and the backstory of our world. We went to Pinterest, Tumblr and google images to find anything cool. We stumbled on small, claustrophobic Tokyo market streets which had awesome lightings at nights and days. We were also playing Overwatch a lot at that time, and both the lighting of Overwatch and the market streets were very similar in certain areas. As a team, we decided to attempt a similar overall style like Overwatch.

References

After deciding which kind of environment we were going, I wanted to find an actual reference in real life to make our job easier and not improvise too much, especially at this early stage. I wandered around randomly on google map in Tokyo, and I found this place called Memory Lane / Piss Alley. I analyzed the environment and found a lot of things that I liked. There were a lot of repetitive props and buildings that I could separate into modular sections. However, I didn’t want to replicate this place exactly as is 1:1 but stylised it to make it a better fit for our project realistically to finish it (and with a touch of Overwatch game art). In real life, the whole street has about a dozen stores. In our 3D environment, it was about 3-4 stores.

At this point, we knew what kind of direction we were going to. This is also the perfect moment to analyze in depth what’s in our references. I probably took a few hours to list out every prop from the google map viewer and google each object to have better references. In the end, I chose which props were worth our time modeling, which props could be modular, which props would be actually seen in the environment and so on. This step shouldn’t be overlooked; I think it’s one of the most important steps. It’s better to do it now while we’re not in the production step because the moment we go into the production phase, I think we should focus on the thing we’re working on in front of us. We wouldn’t want to break the focus to find references. It could ruin the momentum, and it could distract us from being productive. It’s also worth mentioning that it shouldn’t be a hassle looking for references. You’re choosing which props inspire you to model, to texture and so on, you should be excited! In my case, whenever I jumped on Overwatch to play, I reminded myself to take screenshots and pay attention to references of interesting props and lighting scenarios.

Production

As for my modeling process, I kept it as simple as possible for me. I start with very simple primitives to get an idea of all the props. Having listed everything made it a very straightforward process; I would begin to modeling everything pretty quickly and then import them into ue4 to begin placing them to block out the environment. At that point, I was pleased with the density that I’ve achieved with only simple primitive geometry. As you can see in the image below, most props at this stage are cubes and cylinders. The most intricate props I made so far are probably the wires that are hung on the walls. In the end, I had around 8 types of modular electric wires that could connect to each other in every direction. I could flip it or rotate it, and any cable would still make sense together. I love cables because they really help with the details and the density of the scene.

This is also where I should spend a bit of my time on the lighting. Start big and broad and finish it later with precise details. Set up a mood first and leave it at that. Once you have a bit more to play with like textures and colors, you can jump back into the lighting. I’ll go into more details at the end of the article.

Once I’ve set stone the blocking it’s time to start modeling the game resolution or whatever you’re aiming for. In my case, I decided to go as a mix of film and game for the resolution. I had some traditional low res for a game with normal baked on and a few other straight up mid poly without much in the normal map. Sometimes it’s not worth dealing with a highly baked on a low poly when it’s only a simple rounded bevel for instance. Overwatch had a lot of nice rounded bevel on many corners which can be done by a simple soften the edges. If you plan to have a hero asset which is usually a lot denser in term of details and poly, then yes you should spend more time on it and actually make it the right way. As for my project I didn’t want a specific hero asset, I really wanted it to be good as a whole environment on its own without real attention to any props in particular. My real focus was aimed at the lighting from the start. I firmly believe that a scene with only very simple models should be plenty good with well-crafted lighting. It shouldn’t need a hero prop to capture the attention of its viewers.

There’s this place in the game Persona 5 (image below) where it’s a little corner of a school with vending machines. I thought that it was such a cool place to hang out for the characters in the game. This inspired me to build something very similar in my environment. In my opinion, little inspirations from your favourite games or movies help with creating a personality for your project.  

Here’s the little corner.

When I feel very comfortable with the current state of the project, I decide to iterate somewhat drastically. It’s pretty essential to iterate a lot along with your pipeline. Maybe you like it a lot as is, but there’s also a chance that you have a better idea or composition in a few iterations. You will never know unless you try different things, sometime super cool idea will come spontaneously. Through this iteration process, It’s also a crucial moment to realise what you don’t like either. There was a point where I wanted to add vegetation to the left wall.

I went all the way and did a post-apocalyptic scene like the game last of us. My idea at that time was the vegetation would be overgrown on the left side, and it would slowly creep up to the right side. I decided that it wasn’t a good fit for this project. I end up keeping the tree but turning it into a pink tree instead and replaced most of the trees with tall buildings and apartments. Although, now that I’m thinking of it, I’ll probably do a personal project based on this idea.

As for the textures and material, they’re rather simple and basic. I textured my props with Quixel, and I base stylized shader online as a starter and I kept it pretty simple. I knew that I would tweak a lot in UE4. I built a fundamental master material with a lot of simple parameters so I could change on the go the details of my texture maps with instances. This is what my master material looks like. I usually have a normal map (if the prop has one), a base color map and finally, an RMA map which is a roughness, a metallic and an ambient occlusion together on different channels.

If I didn’t like the color or roughness of my metal, then I could change it on the fly with my parameters.

I also built many decals to add a lot of details that were missing in my texturing. There was a lot of grunge and damaged surfaces decals. I also had much graffiti in my scene. I painted on Photoshop a bunch of graffiti in black and white on a 2k resolution map. So basically how I set it up was pretty simple actually. As long the graffiti is selectable in a rectangle without overlapping on the other graffiti, it would work wonderfully. I could change on the fly the colors and roughness of my decals with only a few texture maps.

I have parameters to move up and down, left and right the selection. I also had a few parameters to scale the size as well.

If I want to change the decal, I’d just have to move the selection and its size to fit the decal I want with the parameter values. Finally, I have to rescale the decal itself to fit on the wall I want.

This is how my decal material looks like.

Lighting

This is the fun part. Light plays an enormous role in the composition of your image. My philosophy about lighting is simple, guide your viewer and tell them a story through it. You’re the storyteller.  As a lighting artist for games, my priority is to guide the player somewhere. There’s a lot of ways to do so. For instance, shine the end of the hallway with a solid key light. If you want to make the player feel curious about the lighting, hide the source of the light. Play with different colors that would stand out. If you’re a knight wandering around an abandoned castle with tons of natural orange lights coming in, a bright blue light coming out of the throne room would be very unusual in a good way. Lighting the right objects can tell a story just by looking at it. A great example, lightings in Overwatch are amazing. They don’t actually have an in-game / campaign story; it’s only lore in actuality. However, by lighting certain dead machines, it can tell the mood of the map. That’s the beauty of lighting; we don’t need highly detailed models to light. That being said, if we have a cool environment all textured correctly, our job would be to bring the 100% quality to 120% with fantastic lighting.

First thing first I set a bunch of fixed cameras so I could take a screenshot whenever I change something. Doing this really help with creating a nice Gif at the end to see the progression. It’s also handy to compare the modifications and see which one is better. When I’m focusing on the lighting, I tend to stay on the view “detailed lighting only” so I won’t bother with the colors from the textures.

Dynamic lighting vs baked lighting

Next step is I have to decide whether I’m going to do a baked lighting or dynamic lighting. Personally, baked lighting always does a better job at the end while being super optimised. Dynamic lighting seems like a more comfortable way since there are many factors that you can ignore such as lightmaps and the daunting settings to bake. I can see the appeal to the newcomers to UE4 in those matters which tends to categorise dynamic lighting as for beginners only. That being said, I love lighting a scene with only dynamic lights. The reason is that it’s tough to get a result at the same level as baked lighting. If you managed to fool your viewers and made them think it’s all baked GI while it’s only dynamic, then you’ve done your job, GG. Don’t get the wrong idea though, in the context of a video game environment; you definitely have to take into consideration the optimisation. For my Memory lane project, it was all dynamic lighting. Sorry, not sorry. Although most of my other personal relighting projects are 100% baked lightings.

Composition

Part of it is level design, but lighting is a big part of the composition. A simple trick to see if your lighting is guiding your attention correctly is to blur your image.

You can see here that there are two points of attention from my lighting. The big open space in front blasted with my lighting and the bright wall in the background.

This way, I can guide the viewer from point A to B in a very straightforward way. The viewer walks in the street, stands in the light beam in front and sees another bright beam of light hitting the stairs further away. With this specific shot, I try to play with everything around it to point towards the stairs with the bright wall with bouncing lights. You can see on the right side; I intentionally let a crack open with the metal door to create an artificial light source pointing towards the stairs.

This is something I do very often in my projects, the two light sources from a single hallway. On this Victorian project, I also guided the attention from point A to B in a hallway type of environment.

Since I have this big blasted light in the middle ground, I balanced it with a darker foreground on the right.  

Shadows

When we talk about lighting, we also have to talk about shadows. Making the right place darker can also tell a story. Leaving an area dark can make the players assume what’s in there. Let them imagine what it could be there. It would create a very mysterious scenario. Lighting isn’t the only thing that can catch the attention of the viewer; shadows can do that job too. Here in my desert lighting project, the cat is looking at the darkness and also being in focus of the light. Not only it’s dark but we don’t see the end of it or what the cat sees. This can create an intriguing effect on the viewer. What is the cat looking at? What’s there? Is it a dark cavern? Only the cat knows.

It’s also worth mentioning that I indeed placed everything pointing toward the cat. The fact that the cat is looking at the dark, creepy shadowiness is the intriguing factor that tells the story. The old pillars are also bonuses to create a feeling that there was a civilisation powerful enough to place them there but something happened. 

As for post-processing, it should only be used as the extra 10% at the very end. If the mood can be done in the lighting itself, then you shouldn’t be doing it with the post-processing. I usually use post-processing to lift up the darkest black in the scene or to boost the brightness of the highlights. Play with LUT, if you don’t know how to create one yourself, there are plenty of free LUT online.

Tricks and advice other students

Here are some tricks that I can share:

• You’d be surprised by how much lighting in video game and movies don’t make sense at all. The reason is that there’s an artistic lighting pass to most environments. In short, we lighting artist very often cheat and add lights where it needs to. Sometimes some places need more light, or you want to add a rim light to an object even though it’s not logical. One way I create fake bounce light is with point light that doesn’t cast shadows, fog and without specular. Usually, you would want to avoid using point light that casts shadow anyway. Point lights are technically 6 spotlights in 6 directions like a cube. If you cast the shadow, that means you have an additional 6 shadow maps (if it’s baked).
• Use fog cards or god ray cards to fake the effect. It’s not the best for the game environment, but if it’s only for an image as an end goal, it’s totally fine. You don’t have to create it from scratch if you’re not comfortable with creating complex material, you can find it in the free “Blueprint” project on the Epic Games launcher.

• Relight an environment with a different lighting scenario from a photo or cool concept art. If you want to become better at lighting, you don’t have to spend an enormous time to build an environment just for lighting. There are plenty of free environments on the ue4 market and online to use it as a personal project. This one is the reflection environment.

• Better yet, use grey cubes only and try to get a good mood. As I said earlier in the article, you can get excellent lighting without any details or texture at all.

• If you have made some environment, use them and give back life to it with different lighting!

• Read about painting and photography compositions. Learn about color theory. It’s all going to be worth it in the end.
• Have a bunch of inspiration on hand like Pinterest, Instagram, Twitter, Tumblr or even better real art books from your favourite games and movie. This are my beautiful collections.

• Find a specialty that you’re proud of; environment, character, VFX, animation and so on. It’s like a game (Final Fantasy in my case). Don’t pick the mage class if you hate casting magic. Pick the warrior class if you like the sound of being a warrior. Eventually, you’ll have to pick a sub-class, dark knight or dragoon? You choose. Be proud and excited. Be open to change your sub-class mid-way if you find your true calling. I could’ve stayed being as an environment artist, but lighting piqued my curiosity so much that I just had to become a lighting artist. Did you know there was such thing as groom artist or even a rock artist?

• Don’t be afraid to ask for critiques or advice from the artists that inspire you. It may seem very farfetched to have a reply from a senior artist of your favourite game studio but I’ve done it many times and the worst case possible is that they ignore you. Be respectful and be professional about it. You’d be surprised by how willing they are to go an extra mile to help younger artists when they feel like it.
• Compare your portfolio pieces with the right people. You will be very discouraged if you only compare your work with the first page of Artstation. Don’t do that if you’re not in that specific skill level yet. Search for student works, search “art test” or “aspiring artist” to find them. Search artists from your local area, and not only your classmates. Look for senior artists from your areas that are willing to help you.
• Be active on art platforms online to build a connection with professional artists. Artstation is an absolute must in my regard. Polycount is a fantastic forum with an insane amount of information. Facebook groups like Level up and Ten thousand hours are very useful to seek critiques on your works and to gauge your work with other people.
• Finally, read a ton of 80.lv articles. Like all of them.

Conclusion

It was a long and hard path that I ultimately have to admit that I wouldn’t be able to be here without taking it. I was lucky enough to have met super talented guys and gals to keep pushing my limit when it was the toughest time. I also wouldn’t make it without learning to sometime enjoy well deserved long coffee breaks to joke around. So yeah, school wasn’t half bad.

I am actively looking for an opportunity to work at the moment in Montréal or anywhere else. I am open to relocation. You can contact me through my LinkedIn, Artstation or through email at peter.bv.tran@outlook.com

Many thanks to Kirill for this fantastic opportunity to write an interview and thanks to you all to read my article! Cheers!

Peter Binh-Vinh Tran, Lighting Artist

Interview conducted by Kirill Tokarev