This post continues the discussion from the previous article, relating to things to consider when adding realistic 3D people to illustrations and animations.
There is a broadcast commercial currently running on the major networks for an allergy medication that you may have seen. I’ve seen it many times and in the commercial, for a nasal spray to relieve allergy symptoms, a small group of uniformed digital people show up to kinda promote the product. The figures are all male, and are wearing white suites and a type of headgear, but at any rate they illustrate very well the challenges faced in adding 3D people to animations. Simply put, these figures look absolutely strange and rather creepy. Their skin tone is lifeless, mouths do not synch with the vocal track, and their facial expressions are, well, expressionless. Visually it is a total fail in my view. Yet one can only assume that the pharmaceutical company that put this together must have recruited the best talent they could find for this high-profile commercial and it looks rather terrible.
There is a simple reason why these people, and all digital people I’ve ever seen, have such a lifeless and weirdly creepy appearance. Simply put, we as humans know nothing better then the human face and figure. Think about it, from the moment come in to the world, we begin our study of the face of our mom and dad and anyone else who cares for us. Almost immediately we develop an expertise for interpreting complex human expressions. We very quickly learn, really, in only a few weeks time, how to interpret joy, smiles, sadness, anger, in our parents. This is an amazingly complex and finely honed instinctual skill that we are born with. . .and it is vital to our very survival. So, there is nothing that we, as humans, know better then the human face, and secondarily, the human figure.
And, human facial expressions are amazingly complex. I’m not sure exactly how many muscles control human expression but I believe it is around 45. There is no software ever developed for creating digital people that remotely approaches this complexity. And I’m sure someday software will exist that will allow us to create the human figure quite accurately and convincingly, but this goal seems to be far in the future at present.