Skip to main content
KBS_Icon_questionmark link-ico
;

Face to interface

How could the latest lip-syncing software transform animated film? Dr Christopher Holliday discusses manipulative technologies and novel machine learning processes, focusing on the TrueSync software – an AI engine that creates lip-synced visual translations and presenting this at Animafest Scanner.

Subtitled the “World Festival of Animated Film” and first established in 1972, the Animafest Zagreb film festival comprises what the organisers enthusiastically describe as a dizzying array of “competitions, retrospectives, presentations, exhibitions, pitching forums, lectures, awards, panoramas, [and] workshops” collectively aimed at the exploration and, crucially, the expansion of animation as an industrial and artisanal medium of creativity and craft. Yet amid the busy calendar of activities that has helped make Animafest the second oldest film festival in the world completely dedicated to animation,an emergent element of its growing animated agenda has been the two-day academic conference ‘Animafest Scanner’, which this year for its 10th anniversary took as its theme four interconnected topics that were designed to plot the scope of animation past, present, and future (“what does ‘animation’ mean now?”; “sci-fi in animation”; “animation assisting other art forms”; “digital technologies and originality”). The structuring aim of these key questions was to develop animation as an interdisciplinary “superpower” of representation whose everyday – yet increasingly pervasive – existence across moving image culture requires an ever-closer look at this most universal of communication media. As I say to the King’s students enrolled on my From Innovation to Illusion: Topics in Animation module, the animated medium is everywhere, from feature films and advertising to music videos and medical imaging. It exists between us and the world, an intermediary that holds the ability to comment on and literalise that which is material, to visualise certain psychological and physical interiors, and to act as a device of catharsis that creatively captures subjective experiences, complex identities, and multiple states of being.

Chris Holliday Animafest

After watching Animafest Zagreb from afar for several years (and having written a review of the Global Animation Theory: International Perspectives at Animafest Zagreb collection for the Slavic Review back in 2019), I was naturally very keen to experience first-hand the kinds of intellectual synergy between theory and practice among filmmakers and scholars for which the conference has become best known. Too often placed as an adjunct or afterthought to film and media enquiry, animation was here positioned front and centre in ways that interrogate animation practice-as-theory and theory-as-practice. In doing so, the Animafest Scanner effortlessly sidestepped the well-rehearsed statement of intent that animation is nothing more than a medium ‘for children’, an assumption that is both wide of the mark to anyone familiar with the game of professional justification and increasingly overemployed within educational circles as a default claim for animation’s relevance as an object of study. Thankfully, the interdisciplinary flavour of the Animafest avoided these ‘same old’ starting points from the off, and instead refreshingly ‘took for granted’ animation’s vibrant creative affordances to paradoxically confirm its status as a pluralistic and heterogenous medium.

Nowhere was this expansive and provocative approach to animation more evident than in the Animafest Scanner’s particular focus on the stakes of originality and authenticity in relation to digital technology, which of course remains underpinned by the multiplicity of cautionary tales of the kind that I have found in my own ongoing research into digital de-aging and Hollywood’s virtual recreation of youth and the online video production of Deepfakes as the culmination of the computer’s growing powers of distortion. Thanks to funding from King’s Digital Futures Institute as part of its Microgrant scheme, I was able to present my own paper at Animafest Scanner on manipulative technologies and novel machine learning processes, looking specifically at the TrueSync software – an artificial intelligence engine that creates sophisticated lip-synced visual translations that so far has been used primarily when adapting films for foreign release (whereby the performers’ original mouth movements are digitally retouched to match new language dubs) and also to remove verbal expletives from footage in post-production. As a computer programme that relies on the transformation of facial features for the global transfer of popular film, TrueSync certainly swerves the clumsy synchrony of prior dubbing practices across cinema history, and in doing so provides flexible opportunities for multilingual film production. Yet its possible set of applications also raises timely questions around where the real-time manipulation of ‘live’ photographic footage (as offered by SyncWords and EzDubs) can go, if not the collapse (or perhaps sharpening) of cultural specificity and national identity if images now have the ability to speak in tongues.

From image-generating interfaces (DALL-E, Stable Diffusion, Midjourney) to lip-sync visualization software and synthetic speech platforms (Resemble AI), deep learning algorithms and artificial intelligence like the Neural Network Lab Flawless’ TrueSync undoubtedly now form an emergent part of the contemporary media and entertainment industries. Furthermore, the collision between technology and the face (as other authors on this blog have noted) has meant digitized physiognomies have regularly been the bearers of how and where AI technologies are being applied within the context of new media production, and as such are central to how we might better understand the new encounters that the human body is having with the digital due to its ongoing technological mediation. The specific implications of the TrueSync software are yet to be fully felt across the film industry, yet recent news in Variety and elsewhere the Hollywood trade press that Scarlett Johansson is to take legal action against an Artificial Intelligence App that utilises her facial likeness in the creation of a digital doppelganger, suggests that it may not be immune to the kinds of crisis narratives that continue to define technological manipulations of the body. TrueSync’s story as a tool of digitally-assisted augmentation and its techniques of video dubbing (or “vubbing”) is thus likely to be told through the familiar narratives of loss and gain – what affordances does digital processing offer, and where might its limitations and untruths lie.

With speakers, filmmakers, and creative practitioners from across the globe calling for fresh ways of understanding agency, performance, and presence in the digital era, the conclusions drawn around the acceleration of synthespians, proxies, and avatars surfaces a set of anxieties around the very disposability of body that should be of interest to those working in a number of fields. Including, of course, at our very own Centre for Technology and the Body that – like the Animafest Scanner – asks us to consider both how we might live with technology, but also the potentially worrying terms under which technology might be living with us, face to interface.

Animafest Scanner

Animafest Scanner X – Symposium for Contemporary Animation Studies takes place annually at the University of Zagreb Student Centre in Croatia’s northwest capital.

In this story

Christopher Holliday

Christopher Holliday

Lecturer in Liberal Arts and Visual Cultures Education

Net Gains? Living Well With Technology

Our experts discuss some of the challenges of living well with technology: showing what we can learn from a long history of technological innovation as well as addressing the challenges of…

Latest news