Back in 2011, I was just another VFX student. After being inspired by Pixar back in high school, I’d decided to begin studying computer animation and, like many others, completed a final graduation short together with a few friends in my last year. We called it ‘Mac ‘n’ Cheese’.
Ultimately, our hard work on the final short really paid off. To our complete surprise, ‘Mac ‘n’ Cheese’ went viral, and before long, we even had several potential clients asking whether we were a studio and offering us work. That’s when we decided to start our own creative production agency, Colorbleed, and we’ve been lucky enough to work on projects together as friends ever since.
In this guide, I’ll go through how our team completed the final look for one of our most exciting projects yet: our very first game trailer for post-apocalyptic single-player shooter, Hard Reset Redux. We only had six weeks to complete the entirety of our work on the action-packed trailer, which introduced the two main characters in the game through an exciting chase sequence and explosive battle, building up the tension before seamlessly transitioning to in-game footage. Using Fusion, we went from raw Maya character and environment renders to a look that included effects, such as depth of field, glow, and smoke, compositing them to create a final result that would impress our client in record time.
Step 01 – Render your character
The first step was to render several passes of our character to ensure optimal control in the final comp. Creating the character itself was made much easier thanks to the game development team, who provided us with their ZBrush character models, complete with textures to use as reference.
We would simply take the high-res model, reproject all the low-res textures on it, then repaint extra details to make the textures appear high resolution again. Here we used Fusion for combining all of our Displacement maps. If I needed to update the Displacement map on a character’s face due to different pore placement, for example, all I needed to do was update my Fusion graph. Ultimately, I was able to write out an 8K Displacement map for an entire character straight out of Fusion. These Displacement maps were plugged into various shaders, from SSS skin materials for flesh to metal, rubber and fabric for clothing and mechanical parts.
Step 02 – Merge with a background
The next step involved merging our character with the background. For the environments, we were provided with both maps and models that we could reuse by the game development team, which made the process of creating backgrounds much faster and easier. We would typically render a Color pass and a couple of shadow passes from Maya, then perfect the rest of the image directly in Fusion through some clever compositing, without any need to add extra complexity through specular highlights or SSS passes. For slow shots with no intense movements, we were sometimes even able to render the character and environment together, and use multiple Mattes to then separate the two from each other. We used simple RGB Multimatte passes and the separate channels as masks for Color Correction nodes, for instance. The main character would be in the R channel, the robot would be in the G channel and the environment would be in the B channel. Sometimes we would break it down even more and separate a character into different Mattes, too.
Step 03 – Add fog
We added fog next to give the scene a moodier overall feel. We did this by importing a fog render as a separate layer, which gave us maximum control in post. This was essentially a Maya fluid rendered with V-Ray’s Phoenix shader. In post we could then decide how much influence we wanted the fog to have. We also experimented by doing things such as bumping up the Contrast, or colour correcting the fog in Fusion before applying it to the final comp. Fog can also be created directly in Fusion by using the Fog 3D tool, Volume Fog tool, and Fog tool in Tools>Deep Pixel. This can be used in conjunction with SoftClip, which takes in the distance of a pixel from the viewpoint to affect opacity, allowing objects to gradually fade.
Step 04 – Play with motion
Once the overall feel of the environment was set, we then focused on effects, such as Motion Blur, to enhance the realism in the animation of the action sequence. First, we rendered a Velocity pass from V-Ray. Because V-Ray outputs for use in plugins such as ReelSmart as default, however, the middle value representing no motion blur is 0.5. This was an issue because Fusion’s Vector Motion Blur node will interpret 0 as no motion. To adjust for this, we subtracted -0.5 from our Velocity pass before plugging it into the Vectors input of the Vector Motion Blur node. I then used a Custom Tool, switched to the Channels tab and set the Red Expression to r1-0.5 and the Green Expression to g1-0.5. We then plugged the Custom Tool output into the Vector Motion Blur node and set the X channel to red and the Y to green. The last step is to play with scale to finalise the Motion Blur effect – I usually go for 0.25 for normal scenes down to 0.1 for heavy movements. Remember to always render your Velocity pass without filtering for proper results.
Step 05 – Depth of field effects
The depth of field illusion was added using Frischluft Depth of Field and out-of-focus effects with a specific bokeh to generate a more realistic final result through a very shallow depth of field. Keeping the focus close with a small falloff makes the shot more mysterious. Colours were lifted through added layers, based on Z-Depth channels to separate the character from the background even more. This created a nice silhouette effect.
Step 06 – Dust and sparks
Volumetric dust was then rendered separately, as well as sparks, to add some extra detail to the scene. For some shots, we even added 3D and 2D stock footage to enrich the scene. To blend these details into the rest of the comp, we used an Object Property set with a couple of overrides in V-Ray to tell all our geometry, as well as other elements including hair, particles and curves, to cut out the Alpha. This way we were able to layer elements like dust on top, for example, and they would blend perfectly.
Step 07 – Camera shake
Next, a camera shake effect was applied. I used Fusion’s Transform node for this. Ctrl/right-click the Centre option and choose Modify With>Shake. Here you can adjust the Smoothness – I decided to go for a low value close to 1 for fast shakes. You can see Min and Max value as the input value, not as an offset. If you want no shake at all then you need values of 0.5 and 0.5 to keep the frame centred. I like subtle shakes, so I went for Min: 0.49 and Max: 0.51. You can do the same for the Rotation value. Just click the Angle option and modify with a shake too. Keep this even more subtle at Min: -1 Max: 1. Back in the Tools tab, ensure you then select the Wrap or Duplicate option under Edges to avoid gaps on the edges of your comp. You can also scale the image up with the Size option. Lastly, I would recommend enabling Motion Blur, too, for added realism.
Step 08 – Add in glow
From the beginning, we were using Blade Runner as one of our main references and sources of inspiration for Hard Reset Redux, so we aimed to add plenty of vignettes, chromatic aberration and glow effects to the final shots. To do this, we used Fusion’s Soft Glow tool with the High threshold set to 0.9 and a small glow size of between 2 and 5 to add highlights, and a FuseGlow tool to blend the surrounding pixels together in a subtle way. This is basically a blur with a large value of about 20-50 with a low blend of about 0.05-0.2. This render element was used to exaggerate the neon lighting in the shot. Finally, a Highlight Filter node with Curve set at Standard, Length at 0.5 and Angle at 45 was used to boost the illumination.
Step 09 – Final colour correction
Our final step was to give our trailer a cinematic colour grade using a File LUT node, which allowed us to import a cinematic LUT, which you can download online. Some are free, others you need to pay for. We then added several Color Corrector nodes and a Color Curve node to ensure our final shots had a nice contrast while keeping blacks and whites intact.