With the improvements in processing power of modern graphics cards the characters in video games are appearing increasingly lifelike but have now possibly hit a wall. Human beings have over the millions of years of evolution developed an amazing ability to recognize other human beings and are able to discern the difference between real people and fakes. No matter how good the waxworks are at Madame Tussauds they could never be mistaken for a real person because they lack the ‘spark of life’ that humans are somehow able to recognize. The same is true of video game characters and computer generated animations no matter how close to realistic they are they will probably always lack that je ne se quais. The characters appear like dead people that are moving, puppets made of human flesh.
This phenomenon has been termed the ‘Uncanny Valley’ by Japanese roboticist Masahiro Mori and is written about in the article The Undead Zone: Why realistic graphics make humans look creepy by Clive Thompson. Mori used the term in relation to robotics but it is equally applicable to any situation where a representation of life is produced.
The comic book writer and artist Scott McCloud tackles this same problem in his book Understanding Comics. People prefer art in comics where the characters are detailed enough that you can distinguish between the various people but also lacking sufficient detail so that the reader can fill in the blanks and imbue it with life in their minds. But when something is too realistic the reader has less work to do to add life and so focuses on the minor details that detract from the realism.