The Present of Space Vamps

is a unique journey, transporting the practical effects of early silent-era film into a 3-D animation context.

The majority of sets and props were developed in Blender. Blender’s excellent material node shader system allowed us to create a densely textured world, decaying and falling apart.

The Evee render engine allowed us to get instant feedback on camera placement and choreography, and to sample texture designs in realtime.

One of the main drawbacks of using Evee was that the camera depth-of-field was not accurate, which was particularly noticeable when foreground and background pixels were adjacent to each other. This was corrected through compositing in After Effects: a blur effect was added through an exr depth map, or keyframed by hand depending on the camera movement of each shot. This allowed render times to be kept down in relation to a ray-tracing renderer, which maintaining quality of image.

All footage of hands in the film was captured using an iPhone, and either rotoscoped by hand or keyed out using After Effects Keylight. While it did serve to solve the problem of complex hand animations, the primary goal was to create the aesthetic feeling of “presence” to the characters. The black-and-white look of the film allowed this iPhone footage to be “relit” with simple 2-D masking techniques.

After Effects was also instrumental in achieving a cohesive look to the film. We had two major constraints to account for on this project: the technological limitations of a 2013 iMac, and the craft limitations of a first-time 3-D animator. Knowing sheer animation talent wouldn’t get us far, we focused on the physical artifacts of cameras of the silent era. This had an added and unexpected benefit: it allowed us to soften the video-game graphics that came out of the Evee render engine, it allowed hastily modeled detailed to remain suggested in darkness, and it and obscured artifacts from rotoscoped iPhone footage in suboptimal lighting conditions.

Marvelous Designer was responsible for all clothing designs and cloth simulations on animated characters.

Characters and animations were sourced from Adobe CC Fuse and Mixamo. Blender was used to clean up the keyframe animations, add a T-pose, and send to Marvelous Designer for fitting and cloth simulation. The cloth simulations were re-exported to Blender as alembic files, and joined to the original character animations. These were then imported as .fbx files into the Blender scene.

Realistic face meshes were generated from still photographs using KeenTools FaceBuilder.

The Future of Space Vamps

These technological developments allowed us to get over the hurdle of a proof-of-concept, but our dreams for the feature-length version of Space Vamps are so much grander.

Here’s where we see the technology going next:

With Epic’s Unreal Engine, we will be able to push the integration of live-action and animated footage further, creating hybrid live-action/digital performances with real actors.

We are upgrading our hardware to Puget Systems RTX Edition desktop to model, previsualize and render more sophisticated scenes.

Using the Rokoko Motion Capture suit, we will leave behind Mixamo’s pre-fab animations to incorporate more nuance into physical performances.

Virtual production tools, such as the HTC Vive Tracker, will give us the ability to record camera choreography in realtime on a previously animated scene. It will also be instrumental in hybrid virtual production, capturing hand performances in-camera while placing the footage into a virtual scene.

Unreal Engine’s advanced post-processing tools will also remove an entire workflow from the pipeline. With color correction and lens effects, we will be able to create the look of Space Vamps as a virtual LUT instead of adding this post-render with After Effects, in effect reducing the amount of renders for each from from 2 to 1.

Sets will continued to be modeled primarily in Blender, but we will expand our texturing capabilities with Quixel Mixer, infusing each object in the world of Space Vamps with lived-in wear-and-tear, and creating more sophisticated texture maps.

Quixel Bridge will allow us to more rapidly prototype scenes as we develop the script, and in later stages will add a level of detail that would otherwise require a much larger animation team.