Capturing emotion in virtual characters is a challenge. It limits expressional creativity in virtual productions–games, 3D animation movies, and the lot. This post is the second of the three part series that introduces some of the recent evolution in motion capture technology, or more accurately performance capture technology.
Last time we looked at what Quantic Dream was able to use motion capturing to bring an actress’s performance into the virtual world. In this post, we will look at what technology Rockstar Games used to bring the performance capturing to a new level. Following is a promotional video Rockstar created to explain their technology used in the game L.A.Noire to capture the details of facial expression to add a completely new dimension to detective games.
The technology is called MotionScan. It uses numerous cameras from many different angle capture the performer’s face to generate the mesh AND texture to most accurately capture every little detail of the performance. Because the 3D models look just like the actors, it really is the actors acting in the game. The detail of the facial expression is so great that you, as the player, are expected to read the suspect’s facial expression to tell if they are lying or telling the truth. This method is a step up from motion capture in a sense that instead of using a set number of reference points, it uses the entire face to capture the performance. It is much more accurate. However, it is not without trade-off. MotionScan accurately “scans” the actor/actresses faces–as is. The technology cannot be used to “map” an actor’s performance to a different looking virtual character (like in Avatar where the actor’s performance was puppeteering the virtual character’s expression.).
Nevertheless, the detail shown from this technology is proof enough that Video Games can be as compelling and convincing of a story telling medium as movies can be.