- Observations on film art - https://www.davidbordwell.net/blog -

Motion-capturing an Oscar

[1]

Kristin here:

Six years ago, when The Lord of the Rings: The Return of the King was nominated for eleven Oscars, there was considerable grumbling over the fact that Andy Serkis was absent from the acting categories. Many argued that his pivotal role in creating Gollum, the first convincing human-like computer-generated character, should have qualified him for a nomination.

Now we’re seeing a similar debate over the lack of actor nominations for Avatar, with Zoe Saldana’s performance as Neytiri especially mentioned as unfairly overlooked. An intriguing article [2] on the subject appeared in the Los Angeles Times a few days ago. In it, James Cameron expresses annoyance with both the Screen Actors Guild and the Academy for the lack of nominations for his actors:

I’m not interested in being an animator. . . . That’s what Pixar does. What I do is talk to actors. ‘Here’s a scene. Let’s see what you can come up with,’ and when I walk away at the end of the day, it’s done in my mind. In the actor’s mind, it’s done. There may be a whole team of animators to make sure what we’ve done is preserved, but that’s their problem. Their job is to use the actor’s performance as an absolute template without variance for what comes out the other end.

Because of innovations in the motion-capture process, including a tiny camera hung in front of an actor’s face to capture its every nuance, Cameron insists on calling the new technology “performance capture.” In some sense it may be true that the performance is preserved, but once the film runs through the theater projector, can the audience really tell what that “template” was like? I think not, and that’s why there is a reluctance to nominate these actors.

Where is the elusive boundary?

[3]

Don’t get me wrong. I’m not saying that Zoe Saldana and Andy Serkis aren’t fine performers or that their acting did not contribute enormously to the characters they played. Indeed, Serkis’ mo-cap contribution to the creation of Gollum was originally intended to be far more limited than it turned out to be. His facial expressions and gestures were so useful to the special-effects people that he was involved for a much longer period, and techniques to allow him to perform onset with the other actors were developed.

But however fine the original acting and however great the aid it provides to the special-effects team are, the process doesn’t stop there.  To a notable degree other factors intervene between the actors’ original performances and the characters’ final appearance on the screen. Let’s do some comparisons, using the publicity images that the studios themselves considered good indications of how close the expressions of original actors were to those of their characters.

Take the widely circulated image of Saldana juxtaposed with Neytiri shown above. There are numerous differences. For a start, the filmmakers obviously needed to make the Na’vi look like an alien species. They didn’t just give them tails and make them blue and really tall. Human as the creatures seem in many ways, their faces have a subtle suggestion of large felines.

The effectiveness of Neytiri’s snarl has a lot to do with the fact that she has been given exaggeratedly long canine teeth. Moreover, given the changes in the shape of the face, the mouth is not as large proportionately to the entire head as Saldana’s is; the tongue is not nearly as prominent or noticeable. Both tongue and lips are blue as well. All of these features allow the teeth stand out more by contrast.

Na’vi ears are pointed, and some of the lobes are apparently pierced with a small dark disk in the hole. Saldana’s ears played no role in her performance, but the laid-back ears in the Neytiri image, mimicking those of an enraged animal, contribute considerably to the shot’s impact. I remember noticing them while watching the film.

The nose and the wrinkles on and above it have been considerably changed. Unlike human noses, those of the Na’vi are smaller at the bottom than at the top, somewhat resembling lions’ noses. The wrinkles seem to be derived from canine or feline faces as well, extending from the inner end of the eye and arcing down toward the tip of the nose. The human frown lines at the lower center of Saldana’s forehead are transformed into larger, longer, curved wrinkles at either side; these start between the eyebrows and move up and to the sides. There they get extended by the curved areas of darker blue that radiate across the upper forehead, so that the lines of anger seem to cover more of the face. I suspect that relatively little of what the actors did with their noses has survived the special-effects processing. (In the image below, even the shape of Saldana’s naso-labial folds has been slightly altered.)

The change in the eyes is particularly important. Saldana’s eyes have dark irises within which the pupil is barely, if at all, visible. Na’vi eyes are much larger, to begin with, and the irises are light in color, a sort of yellowish tan. The irises fill more of the visible part of the eye, so that the whites of Na’vi eyes are minimized. As a result, the black pupil stands out dramatically. In terms of color, the model seems to be cats’ eyes, though the pupils remain round rather than slits, to avoid making the Na’vi too alien looking. Since the nose has been widened at the top, the eyes are also further apart than on human faces. (The norm with humans is for the eyes to be separated by a distance roughly equal to the width of one eye.)

[4]

Even in a less dramatic scene, when Neytiri is relaxed and smiling, some of these differences remain. The pointed ear is not laid back, but it sticks out from the side of the head at an angle that draws the spectator’s attention and makes the human-shaped face seem exotic–especially given that the Na’vis’ ears are placed higher on the skull than human ears are: while the human ear canal is about even with the cheekbone, in the Na’vi it opens at mid-temple level. Although the points of the canine teeth are not visible in this image, the teeth remain prominent because of their bright whiteness against the blue skin. Though partially masked by the headgear, the vaguely feline nose still differs considerably from Saldana’s. In keeping with the extraordinary height of these beings, Saldana’s neck has also been lengthened.

Again, I’m not saying that Saldana and the other actors in Avatar did not contribute enormously to the believability of their characters or that they did not aid us to empathize with them. On the contrary, although the big blue creatures did seem very odd in the trailers and posters, I have to admit that they quickly came to seem like real characters. Their design’s balance of human and alien is remarkable. I did not continually think of them as walking combinations of numerous elaborate special effects. The new facial-capture system renders expressions very well, as the frame at the bottom shows.

What I’ve pointed out here with relation to Saldana’s contributions to the creation of Neytiri applies as well to Serkis’ earlier contributions to Gollum. In the comparison images below, similar changes were made.

[5]

While Serkis’ ears were covered, Gollum’s are pointed and prominent. Here, too, the eyes have been enlarged and made a light blue so that the pupils stand out. Where the actor’s teeth are straight and even, Gollum’s are pointed, crooked, and separated by gaps. The cheeks have been hollowed and the eyebrows arched nearly to a point near their outer ends. Crucially, the body has been made inhumanly skinny, with long bony legs and arms that are not apparent in this image. I suspect that some naive audience members believed that a real actor had played Gollum, but to most the scrawny figure was a guarantee that no human could have performed the role. (The very thin waists of the extraterrestrials in District 9 served as a similar guarantee that these were not just guys in monster suits à la Invaders from Mars.)

With all the kinds of changes that I’ve pointed out, how would Academy members be supposed to judge these performances were they to be nominated in the traditional acting categories? Where is the boundary between acting and special effects? Despite actors’ and directors’ claims to the contrary, the movements and expressions caught by performance capture are changed in many obvious and not so obvious ways. A close inspection of the comparison photos reveals the details of the transformation, but in watching the film, the viewer cannot necessarily gauge what sorts of changes were made. I can well imagine that actors like Meryl Streep or Jeff Bridges would be justified if they objected to competing in the same Oscar category as what are essentially hybrid performances seamlessly combining the original acting and the digital transformation.

Possible new categories

One way I can imagine actors competing for awards would be for the Academy to create a separate category for motion-capture performances. To judge such performances fairly, the members would have to see videos running the original performance side-by-side with the finished film. This method might allow them to make a reasonable assessment of what the actor truly contributed.

At this stage in the history of film technology, such a category seems unlikely. So far, not that many people have been spoken of as deserving an Oscar nomination for a mo-cap performance. Even Bill Nighy, who was widely praised for his turn as Davy Jones in the second Pirates of the Caribbean film, was not touted as a possible nominee—probably because his face was so thickly covered with tentacles and partly because comic fantasy films tend not to be Oscar bait. So far the argument has primarily been made for Serkis and Saldana. Plus if the Academy did take the approach of requiring the sort of comparison film I’ve suggested, it would be a difficult and expensive thing to produce. Who knows whether Academy voters would watch five such films?

Maybe, though, as performance capture becomes less expensive and more widely used, there will be enough actors to make up a separate category. We’ve seen the animated-feature category grow from three to five nominees this year, and the number of such films being made suggests that five will become the norm. Animated films and live-action ones heavily dependent on motion-capture are somewhat similar technically, so a new category makes some sense.

A simpler and more logical alternative might be to create a category specifically for vocal performances. As has been pointed out in relation to animated films, an actor who is heard but never seen onscreen could in principle be nominated. Such a thing has never happened, but it’s not against the rules. (It’s easy to imagine that it could have happened for a performance like Celeste Holm’s unseen narrating character in Letter to Three Wives.) But with animated-feature and motion-captured performances becoming more common, a best-vocals category seems to make sense. After all, Serkis and Saldana and others like them do speak their characters’ lines, and their voices are typically not altered or enhanced very much. The digital manipulation of sound still lags considerably behind that of images.

The idea is not exactly a new one. The Annies [6], the awards given out each year by the International Animated Film Society, has two “Vocal Acting” categories, one each for film and television.

Very similar motion-capture technology can be used to create films that most people would agree are animated (The Polar Express, A Christmas Carol) and others that embed animated characters in a live-action setting (The Lord of the Rings, Avatar). Creating an Oscar category for vocal acting in animation or motion-captured, effects-based performances would make sense.

If actors are not yet being recognized for motion-captured performances, the Academy has been quick to honor the top scientific and technological innovators of the area of motion capture. In 2004, when The Return of the King won its golden statuettes, the less celebrated Academy technical awards included one to the Weta Digital team for its new approach to the creation of Gollum. (The award was shared with ILM for its similar use of the technique in creating Jar Jar Binks.) This year, on February 20, the Academy honored [7] a team that included one Weta Digital member for “the design and engineering of the Light Stage capture devices and the image-based facial rendering system developed for character relighting in motion pictures” which was used on Avatar and The Curious Case of Benjamin Button.

Despite fears that motion-capture may someday make directly photographed actors obsolete, there seems little chance of that happening. These techniques are fantastically expensive and are likely to remain that way for some time to come. There seems little point to using such elaborate technology to make an actor look like a real person when traditional cameras can do it so much more easily, so motion-capture performances seem suitable primarily for fantasy beings who cannot be as believably created in any other way.


Thanks to Cathy Root for calling my attention to the Los Angeles Times article.

[8]

February 25: See also Mark Harris’ essay [9] in Entertainment Weekly. He argues that the ineffable qualities of an actor’s performance simply cannot be conveyed through motion-capture.

December 14, 2011: A member of the special-effects team at Weta Digital who worked on Avatar has sent me a comment on this entry: “Your Avatar article nails it–there were so many daily discussions about secondary ear and tail animation!”

May 18, 2014: Thanks to David Cairns [10]for alerting me to two recent items on Cartoon Brew: Amid Amidi’s “Andy Serkis Does Everything, Animators Do Nothing, Says Andy Serkis,” [11] which comments on an interview [12]with Serkis on io9, and Amidi’s follow-up interview [13]with Lord of the Rings animation supervisor Randall William Cook, which quite convincingly argues that the animators changed and extended Serkis’ performance.