-
Attaching text to a 3-D layer with AE cam truck w/ continuous rasterizing
Hi all,
I have a locked off HDV shot of a building. I want to “attach” some text to the building. The AE camera is trucking in on a slight arch. When the camera reaches it’s end point the text will fall off the building. The comp is SD 720 X 540 1.33 – this is allowing me to push into the HDV footage without killing the resolution of the footage and this comp will be my final output size.After having fun trying to get the text to stick to the building during the camera movement I wound up with a precomp of the building footage (3-d layer) and the text with it’s per-character animation all set up where I want it. I put that precomp into my final comp which has the AE camera and the animated truck..
I’ve got everything essentially working the way I want however when I enable the continuous rasterizing switch for the precomp with the building footage and the text “stuck” to it the text positioning as well as the footage positioning / scaling gets all messed up. If I don’t have the continuous rasterizing switch on then the text looks horrible. And, as I illuded to above, whatever I’ve tried in terms of “sticking” the text on the actual footage in the same comp as the camera movement, the text doesn’t really stick in the right place throughout the camera movement – though it’only really goes out of place a few times during the arch and each time it sort of creeps back into proper position.. I guess I could keyframe it but I have a feeling that I am doing something wrong and that there is a much easier way to achieve the vision.
Can anyone point me in the right direction – or at least let me know what to tell you about what I’ve done so you might be able to help? Thanks!
–KTFA
-
2 Replies
-
It sounds like your text is probably drifting when everything is in the same comp because you are doing 3d transformations on the text layer to match the building perspective which is actually not 3d, but a perspective image on a 3d plane. When you continuously rasterize a 3d precomp, it remembers the 3d data from the precomp and no longer treats the precomp as a plane. This can be handy, you can actually build a precomp of a cube and rotate it in 3d, but in your case its working against you. The easiest way to accomplish what you want is probably to keep everything in the same comp with your camera, parent your text layer to your footage layer, and make sure your text plane’s z position and rotation values are all set to 0 (the same plane as the parent footage layer). Then use a distortion effect like corner pin, or transform to match the perspective of the text to that of the building.
-
Thanks Matthew. Turns out that while I believed the footage to be locked off on a tripod, there were three short and very subtle places where the camera moved ever so slightly. I tried to track it but couldn’t get a clean track so I wound up using a combination of still frame and live action in different spots. It was tough to get blend the still in because there were a lot of leaves blowing around in the live action which I wanted to keep but it’s looking alright now.
As far as the text goes, I went a different route as I have to get an approval pretty soon. But I’ll go back and try your suggestion for the sake of learning. Thanks for your time in this – I really appreciate you and the rest of the COW that always take a lot of time to help others out..
–KTFA
Reply to this Discussion! Login or Sign Up