Performance Motion Capture for 3D Animation

Project Abstract:

This research project centers on exploring issues surrounding the digitalization of real-world motion for use in 3D animation and digital visual effects. Specifically this project will concentrate on the digital capture of dramatic performance and its application to computer generated 3D characters. The project will utilize the motion capture facility in Rehabilitation and Occupation Studies of performance capture, and Digital Media in Communication Studies for the 3D animation.

Theoretical analysis of the fully computer-generated imagery (CGI) in moving image, including analysis of the ‘performance’ of CGI characters, and the examination of codes, conventions, and representational and aesthetic issues specific to CGI, is a relatively new addition to moving image scholarship. This project aims to engage with some emerging questions in this field using a practical motion-capture project as a starting point.

Both the process and outcomes of this translation/transference of performance from actor to digital character will be examined with regard to the following related conceptual/theoretical frameworks:

  1. The digital translation of actorly performance – analysis of what elements of ‘live’ performance capture translate when mocap data is applied to digital models.
  2. Issues of performance modification for actors working with digital characters.
  3. The problematics of integrating mocap and traditional animation technique.
  4. The post-photographic film and 3D character animation’s relationship to a range of new cinematic modalities.
  5. The perceptual ‘disconnect’ known as ‘The Uncanny Valley’, a term coined in 1978 by  Japanese roboticist Masahiro Mori, where the more human-like a robot (or CGI character) becomes in its appearance and motion, the more its non-human characteristics will stand out, creating an uncanny effect for the human viewer. Combining human-like photorealistic 3D models with motion captured movements can multiply this effect where the captured ‘real’ human movements as perceived by the viewer are not always fully integrated with the CGI form.