How to track a simple wireframe box...its not working bec frame is empty.

[Stonehedge]

Hi Guys,,,,

Theres this video clip which is just black with a computer cabinet in the middle (its a simple box wireframe) like this...

But its rotating in the clip...slowly as in its a 3D rendering. I just want to put a text on the front of the box so that it sticks to the front of the box and moves aligned with it in 3d. It doesnt seem to be working. I put the video clip on a layer and a text on the upper layer. Should the text layer be set to 3D and the video be left as a normal 2d layer. Also under tracking should i do motion tracking or camera track. I was hoping i could mark the 4 corners and the text would stick to that face. It tracks but the text goes and sticks to it at a funny angle and tilted (rotation). Some help please ?

[Rick Gerard]

If you have access to the 3D file, you can do another render that includes the artwork you want to place on one of the faces.

If all you have is the rendered video, the only option I can see would be to use Mocha AE to track translations on all four corners, then take the four tracks to attach them to a Corner Pin effect applied to the text layer. No 3D data is going to be gathered from the wireframe. 

[Stonehedge]

Hi Rick,

Yes that would be the most easy solution. I can insert 3d text there....but then i need mutiple texts about the box to appear and whizz out with motion blur and certain properties and characteristics that are easy to do in Ae. Thats why i use Ae. Basic texts flippping and changing and call out titles etc are much easier in Ae. My ageing old eyes and brain cant do deep 3D work as its taxing on my head. :).

I did a 4 corner pin tracking in Ae and actually it worked. The track is precise as they lock onto the 4 corners of the wirebox. But im stuck with a simpler problem....that is the text doesnt stick flat to the corner pinned plane, it does at an odd angle and tilted and then moves with the wireframe box but albeit at a wrong angle....if i could correct that angle then it would stick on proper. Thats why i was wondering if my text layer should be set to 3d type ? And should i manually turn it and position it with correct orientation.

[Mylenium]

It would be super simple to just re-create the cube in AE inside a pre-comp and rotatet it in a parent comp.

Mylenium

[Stonehedge]

Hi,

I used the simple cube just to simplify what im tring to achieve for explanation on the forum, in real the box is not just a wireframe box....it has rotating wheels like transformer coils, glowing parts to high light each element etc...its a power conditioning equipment.

[Rick Gerard]

If you have a rendered movie of your Power Conditioning Equipement with nothing else in the shot you might be able to use Camera Tracker to create a virtual camera that you could use to add additional 3D layers to the shot that you could use as labels or pointers. 

If the 3D model of the Poser Conditioning Equipement has already been placed in a virtual scene with other objects, Camera Tracking is not going to work. You would need 3D object tracking and that is not possible with the standard set of tools available in After Effects.

Here are a few facts that may help you figure out a solution to your problem. First, perspective is controlled by camera position. Framing is controlled by focal length. If you have a 35 to 300 mm lens on a camera, and the camera is 10 feet from a box, the perspective you have will not change as you zoom in and out on the box. The vanishing point will not change, the parallas will not change, unless you move the camera.

Knowing that, if you have an object that you cannot track in a scene you could create a Camera and a simple 3D rectangle that has the same propotions as one of the sides or the bottom of your power conditioning device and then line up one edge of the rectangle sith the edge of the box and move the camera closer or farther away from the rectangle until the perspective matches. You will have to modify either the focal length of the AE Camera or the scale of the 3D rectangle layer to keep the size consistent and just fiddle with camera position until the perspective matches. I do this all the time using a simple expressions on layer scale. When you get the camera in the right position, you can then parent the 3D Redtangle to a 3D null at comp center and then set a few keyframes to reproduce the movement of the object in the scene or the camera. 

Without seeing the actual shot you are working with, or at least a similar shot, it's pretty hard to give you the most efficient workflow. I use a couple dozen completely different workflows to insert virtual 3D elemts, like the labels you describe into scenes. The workflow always depends on the shot. There is no one workflow that works for every shot.