@hawk911
You are on the right track with some of the things you mentioned concerning combining realworld footage and max scenes but there are a few other things a s well. It is best to have the real world footage shot against a greenscreen/bluescreen for this purpose and you need to think in layers of what is called Z depth in compositing programs (i.e., distance from the virtual or real camera) if you are going to integrate the footage between cg background and foreground elements or for other effects (e.g., Depth of field, Fog, etc).
Do I use a real camera and set it the same as the virtual camera in Max... ie same aperture, focal range etc?
Yes, to match the greenscren/blue screen portion shot with the real camera to a Max camera you need to match the following camera parameters:
1.Camera height off the ground (use distance between the ground and the camera nodal point)
- Camera tilt angle (angle of camera vs absolute horizontal) I use a protractor for the realworld camera and a maxscript for the max camera.
- Camera FOV (In practice this can be a little tricky depending on your camera; I have a Panasonic AG-HPX170P which has a variable zoom lens so I had to calibrate various max FOV settings with what the camera calls “Z” settings.) This is a one time calibration process which will produce a chart that you can use for all your shots. Depending on your camera, you may need to do something similar to equate Max FOV with your camera FOV.
- Distance of camera to the subject (I use the distance between the actor’s nose and the camera nodal point in most cases, but you could substitute the vehicle door you have modeled for the actors nose in your case).
What are some tips to go about this?
Perhaps obvious, but include a ground plane also in your Max scene if you haven’t already done so for measurements mentioned above. Plan the lighting to carefully match both in terms of intensity, angle, shadow softness, etc. Your described scene sounds like it has outdoor elements and interior (car interior) elements. So the lighting here has to be planned for potentially 2 different shots ((1) The 3d vehicle pulling up (pure outdoor) and (2) camera zoom in to reveal person inside (indoor light so you can see character’s face against the background outdoor light)
A 3d vehicle slides up to a stop and the window rolls down to reveal a real person talking inside as the camera dollies in for a closing medium shot of the door and person inside.
A moving camera always introduces another layer of complexity. In this case you will need to learn about camera matchmoving which is a big subject by itself. Also, you may need to create a CG double of your actor for the first shot of the 3d vehicle pulling up, unless you plan to shoot an outdoor scene of an actor driving up in a car, in which case, I am not sure what part would be max modeled. So, you probably need to formulate/storyboard the specific elements of your shot in more detail using the info above and specifying exactly how you plan to transition from what sounds like a stationary shot of a vehicle pulling up to a camera dolly shot in on the actor within a (CG) vehicle. You currently have 3 layers implied: 1. Foreground: Car door and Window (if rolled up it will need to be semitransparent), 2. Midground: Actor , 3. Background: Car interior behind actor and external scene.
Generally, I storyboard the entire scene in Max with CG doubles of my actors. Then I shoot the actors against a green screen using the 4 Max scene camera values mentioned above. Last step is compositing in After effects. But the live footage composites perfectly if you match the above parameters. Hope this helps.
