Motion Capture: Recreating Johnny Silverhand in Cyberpunk 2077

This project showcases a recreated animation of Johnny Silverhand’s powerful confession scene from Cyberpunk 2077 using Unreal Engine, Metahuman, and other advanced technologies. Johnny Silverhand, modeled and performed by Keanu Reeves in the original game, is brought to life here with high fidelity and a personal creative twist.

Project Overview

This project began as a personal experiment to explore the capabilities of Unreal Engine and Metahuman technology. As a fan of Keanu Reeves, known for his roles in The Matrix, Constantine, John Wick, and Cyberpunk 2077, I wanted to recreate the essence of his iconic performance as Johnny Silverhand. The focus was to blend technology and creativity while replicating the emotional depth of the original scene.

Workflow

Creating the Metahuman

The project began with Unreal Engine’s Metahuman Creator, a tool designed for creating highly detailed digital humans. I focused on designing a character closely resembling Keanu Reeves, the actor who portrayed Johnny Silverhand in Cyberpunk 2077. Although the Metahuman Creator has some limitations in terms of customization, I spent time refining facial features, hair, and skin textures to achieve a likeness. The result captures Keanu’s essence, particularly in specific angles and lighting conditions.

Building the Scene

To create a fitting environment for the recreation, I utilized the Safe House asset from the Unreal Engine marketplace, Fab. This asset features a detailed, atmospheric interior with a cyberpunk aesthetic, perfectly aligned with the gritty, neon-lit world of Cyberpunk 2077. I customized the lighting and props in the scene to enhance its realism and make it a suitable backdrop for the animation.

Facial Capture Performance

LiveLink, a real-time motion capture solution by Epic Games, paired with the Live Link Face app on iOS. By performing Johnny Silverhand’s dialogue myself, I was able to capture detailed facial expressions, including subtle movements of the eyes, mouth, and brows. These animations were directly applied to the Metahuman character for a lifelike performance.

Body Mocap

With the assistance of Topher, I employed the motion capture kit from Noitom, a wearable system that tracks full-body movements. This allowed me to act out Johnny’s physical gestures. A key sequence involved recreating a scene inspired by John Wick, where Keanu Reeves’ character smashes the ground with a hammer to unearth a hidden chest. The captured data was processed and fine-tuned in Unreal Engine to ensure smooth transitions and realistic body mechanics.

Animation Refinement

After capturing both facial and body animations, I spent additional time refining the data in Unreal Engine. This included tweaking keyframes, smoothing transitions, and ensuring that the animations were synchronized with the dialogue. The refinements enhanced the believability and emotional impact of the performance.

Voice AI

Achieving Johnny Silverhand’s voice was crucial for authenticity. I used FakeYou, an AI-driven platform with pre-trained models of celebrity voices. By inputting my recorded performance, the platform transformed my voice into one closely resembling Keanu Reeves. This technology not only added realism to the project but also aligned the audio with the visual representation of the character.

Rendering and Editing

Using Unreal Engine’s Sequencer, I set up cinematic camera angles to capture the scene. This involved experimenting with focal lengths, depth of field, and lighting adjustments to create a visually compelling composition. The rendered video was then imported into Premiere Pro, where I synchronized the voice, added sound effects, and applied final color grading for a polished, cinematic look.

Credits

https://www.cyberpunk.net/us/en/
https://metahuman.unrealengine.com/
https://www.fab.com/listings/0c106ab4-e1cd-40c2-8731-b7940e52b4b6
https://neuronmocap.com/

Instructor

Topher Maraffi
cmaraff@ncsu.edu
https://tophermaraffi.com/

Leave a comment

Blog at WordPress.com.

Up ↑