-
Shifting Nested Camera Issue
Have a look at this file. Its an AE 2018 15.01.
I have two cameras as children of a null. They are both identical, both in where they visually look like they are pointing and in their values. (one is a duplicate of the other)
If you go to the front view so you can see their wire cages, select one of the cameras and then unparent it from the null. The one you unparent will rotate ever so slightly along its axis.
It doesn’t matter which one you choose, either will do this. There is no animation on these elements.This came about because I was getting errors in camera position and orientation when exporting for C4D. If you bring just a camera in, all is well. But bring a camera that is being dragged about by a null, and C4d creates slight differences, especially if the AE camera has been rotated inside the null.
I think this has to do with the POI fighting with the rotation values in AE. I’m not sure what the advantage is with a POI that you can rotate away from – so the rotation values aren’t relative to the cameras world space but to the POI. And the POI values become local values (as one would expect) if the camera becomes a child of another object or null. But it seems the rotation values do not. Notice when you unparent the camera how the rotation values don’t change even though the null’s own rotation should add to the child’s rotation numerical values once it is unparented. I think this is because those camera rotation values are in camera space relative to the POI.
Because C4D doesn’t work this way (you can have a target OR you can rotate the camera but you can’t do both) I thought I could tweak the values in C4D to make the cameras match. I tried using a target in C4D to simulate the POI and the rotation offsets from the POI, but the position values of the Target were out by exactly a factor of 10 in the Y.
Even accepting that oddity and just going with it, the rotation values seem to be magnified a little as well in C4d.I was using the duplicate second camera in AE and uparenting it to get world values that I could transfer manually to C4D. And that’s when I noticed the shift.
You can also see from the Jpgs that, even by eye, the cameras have a different angle between aE and c4d even though the values say they are the same.
12429_nestedcameraissue.aep.zip

