-
Best workflow for CGIs and compositing into video
Ok… so here’s my test I shot in our front street:
https://homepage.mac.com/captmench/BikeCompositeOut.mov
I learned a lot in this little clip.
Here’s what I did — My question really relates to if this is best or not.
1) Shot the footage (duh)
2) Brought the footage in for camera tracking in an old program called Icarus. Gave me camera data
3) Got my model (thanks, TheForce.net)
4) Brought it into C4D and then loaded the Tracking Data to give me a new cameraTHAT step brought up a WHOLE bunch of issues that I’ll need to get worked out if my workflow is ok.
5) Rendered out the movie with alpha channel – thanks to the post below for saving me there
6) Composited in Shake
– The shadow was added in Shake and highlighted a problem from step 4 and 5 — I rendered out at my given frame size so I had NOTHING outside the frame to use… hence the ‘ripping’ shadow at the end of the clip.7) Then brought it into FCP for processing and editing.
—
Does this seem as the basis for doing this? I’d like to do more with building virtual sets and combining CG with Video more and more.
Again though — the camera tracking was a pain… but I’ll get to that later.
Thanks,
CaptM
-
1 Reply
-
You’d be better off generating the shadow in C4D using a shadow-catcher setup and some proxy geometry—in this case a simple ground plane. You’ll have trouble if you try to expand the rendered image to be able to generate the shadow this way in Shake—you’ll have to recalculate a wider field of view very carefully as well. I could generate the equation if I had to, but it would be a pain and ultimately this isn’t a robust technique for generating true shadows.
Paying the rent: you might like to check out our SynthEyes package (www.ssontech.com) for a state-of-the-art yet affordable tracking solution.
Reply to this Discussion! Login or Sign Up