Hi! My name is Piet Braun, I’m a Lighting Artist at Gearbox Software. We recently shipped Borderlands 3, which I had the incredible opportunity to work on. Before Gearbox, I interned at Steelcase creating ArchViz virtual reality tours in UE4 while studying Game Design and Development at Michigan State University. I’m really excited to get to talk about my last project with you here at 80.lv! Feels like not too long ago I was just starting out, admiring articles, wondering if I would ever work in the industry. A lot of what I know came from reading articles here, so it’s awesome to be able to contribute!
Fallen Trees: Goal and Inspirations
I loved the rich colors and vintage film look Bleda has in some of her night photography. The layering of the soft fogged out tree silhouettes is so great. You can see her work below:
Roman Gaydakov’s Midnight ride #1 really inspired me in terms of composition and lighting. I loved the way the headlights are scattered in the fog, and the brake lights are reflected on the wet ground.
These two pieces sold me on the ominous foggy forest setting. I really wanted to capture a dramatic cinematic look. I have been teaching myself a little bit of color grading in Davinci Resolve, so I applied what I’ve learned to achieve a similar look.
I really don’t like to spend a lot of time modeling and texturing and often use asset packs so that I can focus on lighting. I primarily used the Environment Set for the setting, the car from Desert Gas Station, and Soldier Gas as the guy running down the hill. I scouted around the Environment Set map until I found an angle I liked. Once I found a good location, I placed a camera to lock down my working angle. I think this is an important step for keeping your adjustments consistent for the rest of the project. From there I knew I wanted to get a composition similar to Gaydakov’s work: the car stopped, headlights on the fallen trees. I placed other cameras around the forest for fun and to see what the environment looked like while I made lighting adjustments throughout the project.
For the natural lighting, I’m only using the Directional Light and Exponential Height Fog. Though the scene is set at night, the directional light ended up being surprisingly bright to express the moonlight. Since I wanted this to be a dark foggy forest, I allowed myself to use the Volumetric Fog’s Emissive as a way to brighten my shadows instead of using a Skylight to bring up the ambient light.
For the car lights, I used Spotlights for the headlights with the volumetric intensity turned up, and Point lights for the brake lights with a stretched out light source length to get the nice reflections on the wet road.
I faked the lighting in a couple of spots: 1) a Spotlight for the cool moonlight rimming the edge of the car to bring the shape out and contrast the red brake light color 2) and a large Point light to illuminate the trees being lit by the headlights, but more importantly to reflect off the side of the car in order to highlight the body shape.
Camera and Exposure
Once I had my basic lights in, I adjusted the exposure settings in a Post-Process Volume to darken the scene trying not to lose detail in the shadows. I ended up with a Min of 1.2 and a Max of 2.0.
While I’m zeroing in on the exposure, I’ll brighten the base color texture of assets that are too dark. For instance, the tires on the car were very dark, so I brightened them.
I am a big fan of wide aspect ratios, like, a BIG FAN. I really dig the look of Ultra Panavision, so I like to set up my CineCamera’s “Sensor Aspect Ratio” to 2.35 or above. I like how large it makes the environment feel.
One of my guilty pleasures for composing shots is using the Fibonacci Sequence as an overlay in the camera’s dirt mask. It’s all bogus, but I like it.
I try to angle elements towards the focal point (in this case, the hazmat guy). I angled the car and the road slightly to give the scene some movement.
Now that my composition and exposure are getting close to the final result, I take a screenshot and bring it into Davinci Resolve.
Here, I wanted to share some things I learned about the different adjustment tools.
It’s a little weird figuring out how to get your image into Resolve. Load it into the media pool, right-click it and hit "Create new timeline". All the fun is under the color tab at the very bottom, hit that.
One of the reasons I like using Davinci Resolve so much is that it has different scopes. I like to use the parade scope. The parade scope visualizes the light information range of your image, the bottom being the darkest, and the top - the lightest. You can check if your image is losing any light information, either by clipping off the top or crushing off the bottom. It’s also a great way to see the relationship between your R, G, and B channels.
This is the parade scope of my screenshot. You can see that there are some red highlights clipping off the top from the bright brake lights. Ideally, you want to keep all your light information within the scope, this gives you more flexibility when you’re grading.
You can see how the color channels are not all similar here. Blue and Green are kinda close, but Red is stretched out. I want to get these balanced out before I do anything else, and I do this by White Balancing. I use the Color Wheel’s black point and white point picker tool to select the darkest and brightest points in my image. This will make adjustments on your Color Wheels and get all 3 color channels balanced.
The Color Wheels are powerful adjustment tools that control different parts of your image.
Lift = Shadows
Gamma = Midtones
Gain = Highlights
Offset = Moves all 3 up or down in value
By pulling the center point closer to a color, you can start coloring the specific parts (shadows, mid-tones, highlights) of your image. Keep in mind that adjusting the intensity of one color affects the intensities of other colors.
I was trying to match Bleda’s grade. Her shadows have a warm quality to them, so I adjusted the Lift of my image towards red slightly. This raised the red channel up off the bottom, making the shadows brighter with a warm tint. I think this is the key to the look: brightening and warming the shadows, and thereby flattening the shadow’s depth. This makes it feel like it’s foggier than it really is. The highlights in her image are warm as well and not very intense. So I inched the Gain toward red to tint my highlights slightly. I also turned down the overall Gain intensity to reduce the power of the highlights in general.
Once I like the grade, I right-click the clip, hit “Generate 3D LUT” and save it. In order to bring this into Unreal, you have to download their Neutral Color LUT.
Bring this color strip image into Photoshop. Add an adjustment layer, select “Color Lookup…” and load the 3D LUT just created in Resolve. Save the color strip with the 3D LUT applied and import it into Unreal. In the texture file, be sure to set Mip Gen Settings to NoMipMaps and Texture Group to Color Lookup Table. In the Post-Process Volume, apply the new LUT to your scene.
If you need a more thorough explanation of this process, Unreal made a handy guide for bringing LUTs into the engine. Check it out here.
I’d like to add a disclaimer about LUTs in Unreal. At the moment, LUTs are not consistent across different monitors and will not appear the same in HDR. I use them because I found it to be a fast and simple way to bring in my color grades. A good scalable solution would be to copy your grade values into Unreal’s Post-Process Color Grading settings manually. Brian Leleux has a great article where he covers that!
Here is the final result!