(Written back in… oh… 2007? But cleaning out old posts after a blog migration. The “zooming through an object” effect still annoys me today…)
I’ve been thinking a lot about cutscenes recently, because I’ve seen a few particularly bad ones.
How do you know a bad cutscene? Because you’re painfully aware that you’re watching something akin to amateur Hi8 footage. Everything seems somehow wrong. In-engine cutscenes are particularly susceptible to breaking the viewer’s supension of disbelief (or shattering the ‘fourth wall’, I think I heard it called recently). Maybe the camera lens appears to be at waist height. Maybe the camera moves through solid objects. Maybe the movements of the camera are exacting and unyielding, or follow a perfect path, or perhaps the camera has perfect alignment with its subject at all times.
There’s something about a human-operated camera, from the visual language of TV and cinema, that’s almost reassuringly wonky. Not necessarily eye-tracking NYPD Blue wonky, just imprecise and inertial.
Say what you like about Hollywood (er, not in the comments, go elsewhere to do that), but they know how to build and shoot a scene, or an action sequence. It’s rare when watching a movie that you’ll be aware of the technical details that went into the production of a scene, especially the bread-and-butter dialogue sequences in most movies. I haven’t played the game The Movies, but I think I’m going to have to go get a copy after writing this.
With games, I think the fundamental problem is that it’s hard to convince an engine to behave like a movie. Then there’s the secondary problem, which is that if the average developer wanted to coax a game into producing movie-style (or even TV-style, let’s not be elitist here) dialogue or action sequences, they wouldn’t know how to start.
So, what do I think I can suggest that might help cutscene producers in general?
Treat the camera as an actor
I wonder if people in the movie or TV business even realize they’re doing it (they do it so well), but game developers in general need to get better about this:
In your average 3D world, the camera is just a set of parameters that define a view onto a scene. That’s fine – functionally, that’s all you need.
But consider this: in a movie the camera has mass, it obeys the laws of physics, and it arguably has a personality.
Some of the most effective visual effects I can remember in recent times have used the “camera as an actor” technique staggeringly well – Battlestar Galactica and Firefly leap to mind.
On the other hand, one shot in Talladega Nights – while way cool – instantly told me that I was watching an effects shot: the one where the camera seems to zoom down, through the windscreen of Ricky Bobby’s car moving at 200mph, then out through the back of the car. It’s a seamless blend of live action and CGI, and instantly I got to thinking about the shot itself; it ruptured the narrative (well, the racing).
When a camera has no presence, it’s not a camera, and it shatters the illusion.