Anyway, I'm hoping to test and expand my horizons with this, bringing CG models and effects into a real life environment, hopefully seamlessly with the aid of Syntheyes, which will be further aided by the steady cam and 720p resolution.
Well, I've run some live action tests in the past, bringing footage into Max, and doing all my CG work against that footage. However, this becomes most difficult when CG subjects are required to pass behind an actor or complex object in the scene.
My best solution has been to apply a shadow/matte material to a many-segmented plane and then animate and reshape the plane to go over the character/object in order to allow CG elements to pass behind them with a makeshift garbage matte. This however is a tedious and extremely inaccurate process. Depth also becomes an issue.
Now, certainly I could take my green screen footage, render certain 3D layers separately with an alpha, and do things that way, but I can't have actors being composited into a scene they could more easily actually be PRESENT in, every time a CG effect is required.
I mean, look at most films. You can tell the actors and the backgrounds were in the same place at the same time. And yet CG effects are still matted in behind them and in front of them. I have to figure there is a more efficient way of achieving this. Not automatic perhaps, but something that is easier to guide than 300 vertices all individually placed frame to frame.
Anyone out there have some knowledge I might find useful?