2 Replies Latest reply on Sep 3, 2010 4:40 AM by Simon Bjork

    Orient layer to camera field of view? (not auto orient)

    Simon Bjork

      I have a comp with a 3D tracked camera from PF-Track. If I enable 3D to a layer, it moves off-screen. I guess becuse of the 0.0.0 points and so on. Here is the problem. When looking through the active camera (in regular 2D view), I move through the timeline, and at a point I decide that I wan't to enable 3D to a layer. When I do that, the layer moves off screen. How would I do to to move the layer back to the exact position? I can of course just use x,y,z position/rotation/scale to bring it back into position, but it's very hard to be exactly accurate. Basically I want to auto-orient the layer to the camera, but as it's auto-oriented, I want it to stay in that position and not orient towards the camera. Ideas?

        • 1. Re: Orient layer to camera field of view? (not auto orient)
          Mylenium Most Valuable Participant

          It's certainly doable, but involves major math. Aside from that, it's not quite clear to me what exactly you want to do. The problem may in fact be possible to solve quite easily by always using a 3D layer. Here's the deal: You would have to calibrate your track properly in PFTrack and then add a plceholder item that exports as a Null to which you can attach the 3D layer. I realyl think it#s more a workflow problem that you are trying to solve on teh wrong end.

           

          Mylenium

          • 2. Re: Orient layer to camera field of view? (not auto orient)
            Simon Bjork Level 1

            Thank you for your reply Mylenium. The reason for all of this is that I wan't to use the camera mapper plug-in from digi-effects. So as I scroll through the timeline and find a frame that I want to project, I just freeze frame the (projector)camera and the frame and apply the camera mapper effect. But the effect only works if the layer is in 3D. And if it does't fit the frame exactly like in 2D mode, the projection will be off-set. Makes sense? It seemed to work by copying the position and rotation of the camera and place the layer in the exact same place. Then offset it in z and manually scale the frame to fit the view. But it feels like there would be an easier/faster solution.