Versatile Actors & Negotiating The Uncanny Valley

ABOVE: 156 LED lights and Emily O’Brien, an actress undergoing high-res face scanning via Image Metric‘s Light Stage 5 device. The process is explained in this brief video.

The advent of digital cinema has brought more troubling (yet arguably unwarranted) fears to the fore. One of the worries spurred on by the rise of digital media is based around the idea that human actors will one day be displaced by their digital ‘superiors’. I think a lot of us tend to treat this particular anxiety with a certain degree of levity given that digital media’s now long-existent (and ever increasing) ubiquity has not yet dramatically compromised the necessity of actors. They’re still pretty popular and more or less essential. However, as the image above hints, this is not to say that absolutely nothing has changed…

In his article, “After Arnold: Narratives of the Posthuman Cinema”, Roger Beebe discusses the gradual death of the action hero at the hands of his tenacious nemesis, the digital morph (AKA the T-1000 from Terminator 2, who Beebe describes as the film’s “effective and affective center of interest and pleasure” [170]). T-1000 and his various successors and predecessors unsettled the once stable notion that wholly real human beings were the primary draw for moviegoing audiences. As Beebe reaffirms, “It is not the humanly embodied Patrick but rather the T-1000 (i.e., the morph and morphing technology) that is the center of interest in T2” (169). T2 can thereby be understood as just one out of many seminal moments in cinematic history that recur every now and then, if only to momentarily rekindle our latent anxieties about obsolete actors.

Looking toward another one of these seminal moments in Jurassic Park, Beebe notes the following:

“Jurassic Park (or, more properly, nearly a billion dollars in international box office receipts) shocked Hollywood into the realization that a blockbuster need not be either a star vehicle or even a narrative centered on the plights and adventures of a single human subject [...] when we return to Jurassic Park, we simply want to see more of the computergraphic dinosaurs” (171-172).

However, similar to the case of Final Fantasy: The Spirits Within, Jurassic Park may have been revolutionary but it didn’t turn out to be a devastating blow to actors everywhere. It is certainly true that both films showcase the allure of digital cinema, wherein actors are not always the primary attraction (rather, we have: prehistoric beasts, spectral aliens, and Uncanny humans being brought to ‘life’). However, that doesn’t mean that either film totally rejects the importance of an actual, human element. Thus, while Sam Neil may not be responsible for Jurassic Park‘s record-breaking revenue, he and a variety of other actors were nonetheless integral to the film. That is, I think it is fair to say that a significant part of the film’s attraction was based on the collision between dinosaurs and man. As fantastic as the former may have looked, the film wouldn’t have been quite as exciting without its motley crew of disposable humans. This is a trend that is still continuing today, ensuring that human actors remain consistently relevant and alluring… even if they happen to be located within primarily digital expanses every once and a while (the latest example of this being the forthcoming Avatar). And when it comes to wholly synthetic ‘realities’ like that of Final Fantasy, wherein live-action actors are pretty much nonexistent (visually), the importance of high-profile voice actors ensures that a link to actual people is nonetheless maintained.

As for the practical benefits associated with replacing a costly star-system with a digital ersatz, they seem dubious at best. Not only is it highly doubtful that the public would want to wholly relinquish their fascination with a good, ‘old-fashioned’ performance, the cost-effectiveness of such an actor-less scenario is fraught with flaws. As Bebee notes:

“One of the more obvious of these dystopian aspects of the posthuman is that the shift away from the 1980s star system may simply represent Hollywood’s attempt to overcome its reliance on the troublesome human labor of stars. (Indeed, much of the panic over the use of computergraphic technologies seems to center on this possibility of the succession of actors and acting). Of course, this strategy has already begun to fail as, ironically, a number of ‘star animators’ have emerged who require the same costly investment in human labor as the action heroes that they helped to displace” (173).

Having considered the resilience of the human actor, I’m thinking that digital cinema is best understood as a partnership between actors and synthetic creations (rather than a straight-up, or inevitable, usurpation of the former by the latter). Even when it comes to those films dealing with exclusively synthetic ‘realities’, actors aren’t always demoted to voice-acting. As an example, consider the possibilities of Contour, a semi-recent step forward in motion capture technology developed by former-Apple engineer Steve Perlman (its possibilities and advantages are discussed in this article). The process involves applying a generous amount of phosphorescent powder onto actors’ faces, and then putting them before a couple of cameras that attend to the various details and nuances of their appearance. Then, as the article explains, “the captured images are transmitted to an array of computers that reassemble the three-dimensional shapes of the glowing areas”. This is the kind of technology that allowed Angelina Jolie to not only voice a digital character in Beowulf (Grendel’s mother) but also end up informing her look (however, a marker-based method of motion capture was used for that particular film). In other words, even if there were to be a sudden dearth of films oriented around live-action footage, actors seem to have already resiliently ingratiated themselves with the possibilities of digital cinema.

Thus, perhaps we can rest easy when it comes to our concerns regarding actors and labour vis-à-vis digital media (for now). However, that’s not to say that there aren’t a variety of other interesting issues related to digital cinema that are still worth diving into. This brings me to the second concern of my post, the Uncanny Valley, and its relation to the actress pictured above, Emily O’Brien. On one hand, she represents the adaptability that I have been talking about, wherein actors take digital cinema for the accommodating business opportunity that it is rather than some immovable obstacle. On the other hand, she also relates to a lot of interesting questions regarding the seeming insurmountability of the Uncanny Valley. Allow the following video to introduce her (or, moreso, her digital doppleganger), Image Metrics, and how the folks behind the company were able to capture her digital likeness without resorting to the old-school methods of powder or face markers:

Note the transition at the 1:30 mark, wherein the real Emily O’Brien is briefly revealed. When this video first hit the net a while back, I remember a lot of unchecked hype surrounding it. There was a healthy dose of doubt and cynicism, of course, but this video was treated by many as a significant blow against that seemingly indomitable force, the Uncanny Valley. For instance, check out this article and the exuberant reactions therein: “I’m one of the toughest critics of face capture, and even I have to admit, these guys have nailed it. This is the first virtual human animated sequence that completely bypasses all my subconscious warnings. I get the feeling of Emily as a person. All the subtlety is there. This is no hype job, it’s the real thing.” Really? It’s undeniably impressive and all but it’s certainly not perfect. Admittedly, the video I linked to in the caption underneath the image up-top really showcases the amount of detail Image Metrics was able to create. It was evidently a pain-staking process, performance analysis and all (i.e., analyzing Emily’s face to see how it moves and changes). It is all pretty astounding and their techniques are commendably innovative. But I’m still left feeling that something, unfortunately, falls short. A testament to the tenacity of the Uncanny Valley obstacle, perhaps.

I find that the video of Emily recalls the lesson learned from Michael Najjar‘s series of photographs (titled “Nexus Project Part 1″) that Joe introduced to us last class. Having edited their eyes, Najjar’s photographed subjects straddle that fine line between the real and the synthetic (wherein the eyes suddenly become a gateway to keenly felt eeriness, rather than the soul). While Digi-Emily’s eyes aren’t quite as vacuous as Najjar’s disconcerting creations, something about them seems nonetheless artificial and grafted on. A lot of Emily’s facial expressions were awkwardly rendered at times, as well. It’s pretty good, and I know that Image Metric was able to capture places (lips, tongue, teeth, etc.) with more finesse than prior modes of motion-capture, but the general awkwardness of the animation is still there.

Most of the problems I’m noticing with Emily’s video more or less cropped up in Final Fantasy, as well. While Final Fantasy looked arguably better than Emily’s digital clone at times, the film more often mirrored her off-putting, stilted quality. This NYTimes article notes a similar problem in regards to Beowulf, claiming that it is impossible to watch the film “without sensing that the ‘actors’ are being pushed around by invisible forces, not living and breathing on their own”. That would serve as an astute description of Final Fantasy as well, as I believe the film falls victim to the second Uncanny Valley pitfall mentioned during Tuesday’s class: mismatch of cue realism. This refers to an instance where digital characters fail to compel us because of the dissonance between the rich detail of their appearance (although there are problems there, too) and the awkwardness of their movements. Indeed, it often felt as if the attempts at natural movement and overall verisimilitude were more foregrounded than those intermittent instances where something actually seemed authentic. Dr. Sid ended up impressing me the most (although he wasn’t without his problems) — I attribute this partially to Donald Sutherland’s voice acting, and the fact that he, unlike Aki, wasn’t conspiculously clean and virtually free of any nuance or facial blemishes. In regards to this last point, I think Roger Ebert put it best in his review:

“Not for an instant do we believe that Dr. Aki Ross, the heroine, is a real human. But we concede she is lifelike, which is the whole point. She has an eerie presence that is at once subtly unreal and yet convincing. [...] The first closeup of her face and eyes is startling because the filmmakers are not afraid to give us a good, long look–they dare us not to admire their craft. If Aki is not as real as a human actress, she is about as real as a Playmate who has been retouched to a glossy perfection.“

Sounds about right to me. This is, of course, a problem shared by entirely synthetic, ‘original’ characters like Aki and those, like Emily, who are definitely informed by specific people (then again, Aki’s look may also be based on a specific person or several, I’m not sure). It seems that, while actors are not on their way to becoming obsolete, the technology that is bringing them into the digital realm still leaves something to be desired. Yet, perhaps the incessant comparisons to the real world only serve to obscure the possibilities already provided by our digital image-making technologies. As Ebert notes in his review: “Maybe someday I’ll actually be fooled by a computer-generated actor (but I doubt it). The point anyway is not to replace actors and the real world, but to transcend them–to penetrate into a new creative space based primarily on images and ideas.” The implication here is that we can move beyond exhausting, perpetual comparisons between the world we know and its synthetic counterparts. Such comparisons may be entirely futile exercises that keep us from enjoying the fascination, visual richness and creativity already awaiting us in the Uncanny Valley. Then again, it’s typical of human nature to tenaciously strive toward such ideals. Even if our continued attempts happen to be repeatedly beset by the dissonance between that which is real and that which can’t help but subtly betray its own artifice.

Works Cited:

Beebe, Roger. “After Arnold: Narratives of the Posthuman Cinema”. Meta-Morphing: Visual Transformation and the Culture of Quick-Change. Ed. Sobchack, Vivian. Univeristy of Minnesota Press, 2000. 159-182.