Video Game Art – Realism – Climbing out of the Uncanny Valley

As the eighth generation of console gaming and an influx of new titles rapidly approaches, audiences are periodically being teased with glimpses of developing graphics engines which further close the gap between consoles and high-end gaming PCs. Admittedly, I am not a tech-expert (nor do I claim to be), so I do not feel comfortable attempting to discuss the technical details of the eighth-generation consoles. I do, however, want to discuss another rapidly shrinking gap: the gap between computer graphics and actual human likeness otherwise known as the “Uncanny Valley.”

You know it's serious business when charts are involved

You know it’s serious business when charts are involved

The theory of the “uncanny valley” does not pertain solely to computer graphics, but of all imitations of the “healthy” human form, particularly humanoid robots. The valley itself is described as a discomfort and lack of familiarity when faced with something that appears close to human, yet obviously not human. In other words, mimicry of the human form is acceptable until a certain point, at which point it becomes disturbing to behold. Examine the picture below of the male default model of Commander Shepard of Mass Effect fame. The viewer’s attention is immediately drawn to the eyes and the mouth; the eyes, in particular, lack life. Though this screenshot is an extremely exaggerated example of corpse-like animation, it is meant to highlight the failure of computer graphics to emulate human emotion externally.

The uncanny valley contains the stuff of nightmares

The uncanny valley contains the stuff of nightmares

In 2011, L.A. Noire undertook a unique and intensive project to recreate a large, section of the city of Los Angeles as well as detailed facial movements of the game’s voice actors using a newly developed technology called MotionScan. MotionScan relied on thirty-two different cameras to capture details of facial movement down to the smallest minutiae. Though it was still not entirely photo realistic, L.A. Noire solved one of the major graphical flaws that had held video games in the uncanny valley for so long; the facial dynamics literally represented that of each of the actors that voiced them with near-perfect precision. Yet, no matter how life-like L.A. Noire appears to be, it is only just that: life-like. Cole Phelps ultimately just looks like a plastic version of Aaron Station, something you might expect to see in Madame Tussauds.

The resemblance is uncanny

The resemblance is uncanny

The latest push out of the valley appears to come from a number of eighth-generation entries: Metal Gear Solid V, Watch_Dogs, and Cyberpunk 2077 have showcased impressive graphics and realism, the latter-most of which looks like it may address post-humanism and the troubles that raises with the uncanny valley (this is purely speculative, though, and an entirely different article). The FOX engine being built by Kojima Productions was announced back in 2011, but it was not until the recently that its graphical capabilities were given a proper showcase. Though the manner of Kojima’s, rather “Joakim Mogren’s” clever reveal of the Metal Gear Solid V trailer surprised exactly no one, the degree of photo-realism was shocking. Take a look at this still from the Ground Zeroes gameplay trailer. It is difficult to describe exactly how the viewer can tell this is a computer generated image making it similarly difficult to determine where in the uncanny valley it is.

Big Boss has come a long way since he first appeared in the first Metal Gear.

Big Boss has come a long way since he first appeared in the first Metal Gear.

Every few years, computer graphics manage to creep further out of the valley. It would appear that it is only a matter of time before computer graphics achieves a level of photo-realism indistinguishable from reality. We can’t help but ask ourselves, what happens then?

Advertisements