The Theory of Uncanny Valley

Whilst becoming more acquainted with the world of 3D computer animation, I found it is also very important to be aware of its limitations and consequences. One of these limitations is the theory of 'Uncanny Valley'. Whilst I had not heard of this phenomenon before, I have experienced it myself and thus found it extremely fascinating, along with the studies and plausibilities made by a variety of academics that try and find its reasons for existence.

The theory was proposed by Dr. Masahiro Mori in 1970 who made a thought experiment that the more we made animations and robots closer and closer alike to the human form, would our affinity to these increase the more realistic they appeared or would there be a drop between this correlation of affinity and realism? Mori himself proposed that there would be a drop, yet there is not a conclusive explanation in the world that proves that the Uncanny Valley exists fully. 

In simpler terms, the 'drop' that Mori talks about is conclusive to the idea of an animation being either 'too real' or having something 'off' about it, so much so that it makes the viewer feel a sense of uneasiness or eeriness. The viewer is almost reminded of the boundary between humans and machines, and machines can only go so far until it becomes so obvious that the robot or animation is not real.

            

I have personally felt this experience with The Polar Express (2004). I remember when it first came out, everyone talked about how realistic it looked yet there was something a little eerie about it. I think this is because although the graphic quality is amazing, the characters themselves appear to be realistic, but they don't have the natural fluidity in their movements that make them exactly like a human being. They have this robotic feel about them, which subconsciously reminds you that they are computer generated. The characters also do not make the most believable facial expressions, as with most animations that over exaggerate, The Polar Express does not, and perhaps this is why they got the drop response in which Mori was predicting.

"Perceptional mismatch between appearance and motion."

Overall, I find that this theory is very interesting because all in all it is just a theory and there is nothing concrete about it. But because I have experienced it, I would like to think that it is real and this makes the critical level of animation even more intense. We are now living in a world that produces some extremely high quality animations and robots therefore the standard that audiences are expecting is even higher! I think to avoid the Uncanny Valley, characters who look like real human beings would have to also act like them. Have quick responses to things around them as opposed to looking like they are being puppeted, and also their facial expressions would have to be believable. Though body language is important, eye contact and facial expressions are also, and I think that this would deter audiences from feeling this odd dissonance reaction.