We were introduced to tracking today in class and told about the five ways you could camera track a scene:
- 1-point, which uses an X and Y position in your scene
- 2-point, which focuses on the Z rotation and scale
- 4-point (there isn’t any 3-point), which uses a 4 corner-pinning (projected perspective) process to track a rectangle throughout the scene
- 3D camera tracking, which is where the camera tracks through the shot
- planar tracking, which takes the region of an image and follows it through
While there are many reasons to use any of the other methods, Peter chose to show us the 3D camera tracking in After Effects first – while in principle it’s complicated, the computer ends up doing a lot of the work for you. We would go over how to make 3D text and then add a 3D camera tracker to a pre-filmed scene. I didn’t really know what to expect when we were told about this exercise, so we just dove in head first.
We needed to set up our composition for 1080p, 25fps, and 8 seconds long. Finally, to finish the set-up, we had to add in our text and make it 3D.
To get the correct effect, we needed to change the composition settings so that the renderer was Ray-Traced 3D. While this is a simple part of the process, if you forget it then the scene won’t work correctly.
After the text was in and the option to make it 3D was selected (the cube in the layer options is on), we could go into the newly added Geometry Options and add some Extrusion Depth. This would give the previously 2D texture layer a 3D feel. I actually really like the effect.
Once we were satisfied with the basics (text, adding a light source, and changing the settings), we could then add the movie clip Peter provided into our composition. It was a simple clip, comprising of someone walking through a park; the important part was the ground, as that was going to be used in our 3D camera tracking momentarily. With the movie clip selected, we could go into Effects > Perspective > 3D Camera Tracker, which would added new options to the clip’s layer.
By changing the Detailed Analysis to On, the computer could then map out a dotted blueprint of the footage – thus, doing all the work for us. It did take quite a while to do and then I accidentally hit CTL-Z and undid the analysis, but I got there in the end. The result was a bit bizarre:
All the anchor points throughout the footage could be used for the camera tracking, so next we would need to select three ground points to create a target on the ground; you need only shift-click on the them as you’re selecting them:
The target could then be used to set the ground plane and origin. This is a reference plane used during the tracking – once you set the ground plane, it will be used for your objects in the scene, but if you tried to add/change it, the existing objects wouldn’t change (they’re not intrinsically tied together).
I think the end result of this looked FANTASTIC.
It really reminded me of intro credits to a movie, where it would have the producer’s, director’s, main actors’, etc. names moving through the scene as the camera dodged about a city location, or perhaps appear and disappear while flying through the sky/space. Overall I feel like the final look could really be developed on and, to be frank, the actual process of setting up the camera tracking wasn’t too difficult for this sort of end result. Obviously it could become more complicated, but to use it for credits or information on screen could really make a whole project look good.
I’m really looking forward to learning more about camera tracking, since we didn’t really have time to go into detail about shadow catchers for camera and lighting, which could add more depth to the scene and really solidify the existence of the letters sitting there.