Nebula, I'm confused. In your first sentence you say 'I am looking for a setup to capture 'accurate' 3d animations ', then in your last sentence you say the opposite: ' am not going to capture 3d animation'. Let's assume you do:
When you say '3d animations', are you talking about steroscopic animations (requiring special glasses to view) or are you talking about motion capture data obtained from live actors?
If the first one is what you mean, then, in theory, you only need two cameras (one for the left eye view and one for the right). But for best results the cameras would have to be perfectly snychronised, if you're capturing 'live' video...and that's not so easy to do. If the video is generated by a 3D computer program, then the stereoscopic images can be created automatically if the program supports it.
If, on the other hand, you're talking about getting data to create 3D models by using multiple cameras, then it gets MUCH more complicated, although there ARE programs out there that can create models from photographs. (Try ftp://ftp.nsk.su/.3/windows/graphics/pmlt31a.zip for a free version of PhotoModeller - normally $895!)
Motion capture is yet another complicated task, whereby 2 or more cameras are used to obtain 'skeleton\bone' movement from live subjects wearing ping-pong balls with reflective tape at key positions on their bodies. It IS possible to get very good motion capture using only one camera, but requires a lot of patience and a 3D program that supports constraints and Inverse\Forward Kinematics...
Good luck!
Russell
When you say '3d animations', are you talking about steroscopic animations (requiring special glasses to view) or are you talking about motion capture data obtained from live actors?
If the first one is what you mean, then, in theory, you only need two cameras (one for the left eye view and one for the right). But for best results the cameras would have to be perfectly snychronised, if you're capturing 'live' video...and that's not so easy to do. If the video is generated by a 3D computer program, then the stereoscopic images can be created automatically if the program supports it.
If, on the other hand, you're talking about getting data to create 3D models by using multiple cameras, then it gets MUCH more complicated, although there ARE programs out there that can create models from photographs. (Try ftp://ftp.nsk.su/.3/windows/graphics/pmlt31a.zip for a free version of PhotoModeller - normally $895!)
Motion capture is yet another complicated task, whereby 2 or more cameras are used to obtain 'skeleton\bone' movement from live subjects wearing ping-pong balls with reflective tape at key positions on their bodies. It IS possible to get very good motion capture using only one camera, but requires a lot of patience and a 3D program that supports constraints and Inverse\Forward Kinematics...
Good luck!
Russell