-
Camera tracking, Placement of new 3D elements in the scene – Next steps?
Hi everyone,
I camera tracked my scene in Boujou and imported all the camera data and Cloudpoint into Blender. I then imported the 3D model I previously modeled and placed it in 3D space as close as possible the points so that the parallax would be accurate (The cloudpoint acts as a whole. I can select single points in Edit mode but I can’t make empties/nulls out of the selected points so I can’t really send the object to that position but I think that it sticks pretty well when I play back).
Now I have some questions as I have never composited 3D into live action footage:
1) How can I make the tracked tunnel entrance, look like it is sticking out of that far away from the cliff instead of sitting in front of it? Is it something I need to do in my 3D environment or is it something to be done in compositing?
1.5) I was thinking of duplicating the clean plate and masking foreground and background, make them 3D layers and my exported set extension layers in between to make it fit better
2) I don’t think that simply adding a shadow, matching noise and color in After Effects will help blend this particular shot. What else should I think about before exporting this scene from blender?
3) When I’m done creating my scene (I will probably add other elements such as bobbed wired fences etc..), would I need to export a separate EXR for every object that is in the scene since they use different materials? Also, EXR doesn’t seem so compatible with After Effects – which is the best export format for After Effects?
4) Is there anything else I need to keep in mind?
I tried looking for tutorials in which 3D elements become part of or modify a real action element (Like having the 3D tunnel stick out of the live action cliff) but I can’t find anything other than simply sticking stuff on flat surfaces or building faces.
screenshot notes: I have attached a screenshot of the rendered viewport, a clean plate, shadow catcher (which doesn’t follow the mountain very well, I know), and a side view (slightly moved otherwise you wouldn’t be able to see the shadow plane).
-
2 Replies
-
You’ve got your tunnel in 3D space in Blender,which you’ve rendered with a matchmove camera. So, now is time to stop thinking in 3D and think of the comp as containing two 2D videos, which you now have to blend together.
If this was a weather map, you’d do exactly as you just in 1.5). So just think of it that way.
In terms of help “sell” this effect, think about the contrast and the blurring on the object. Try to colour match the darkest parts of the tunnels to the darkest shadows you can see.
You might also want to replicate part of the cliff wall, so that the join with the tunnel is easier to hide. And think like a magician, by which I mean draw the eye away from any edges where your VFX meet reality.
Once you’ve got the tunnel blended into the shot, that’s when you start thinking in 3D again and add all the extra bits like barbed wire. Blender can export OBJ files, which you can import into either C4D Lite (or if you have the cash Element 3D from Video Copilot).
My last bit of advice, I have a 4-second rule for VFX. Anything after 4 seconds and your eyes start to pick up on the little, telltale details. I’m sure plenty of professsionals disagree with me on this, but it works for me.
-
Actually, that rule helps a lot in post.
Why linger if you don’t need to.
Sure, if you’re on a bigger production you can have that leeway but in most cases, it should be as little as possible to not give away the magic but to lower the cost of it 🙂
Reply to this Discussion! Login or Sign Up