Any help, people?

[KristianM]

Eh… no, my markers didn’t quite match the plane… and I’ve tried the rotation… my camera ends up upside down when I rotate to the point the markers touch the base in a line!

Any ideas on how to fix this? Or will I just have to deal with the odd perspective?

[FedPete]

Funny I had exactly the same problem. I discovered it at the end of this (difficult) project. I didn’t solved it. I got the impression it has to do with tweaking and deleting adding markers. Because in the beginning of my project these were some what aligned. Then I started tweaking to lower the solve error. And after that, this problem occurred in my project.
It looks like if the camera is looking under the plane. Never solved

[Hiju]

Hi,

I just managed to reach this section, it may be a little late to reply to your post and I am no expert but here is what I think is happening:

The camera solver doesn’t seem to be able to tell which markers are further than others, it may be caused by several factors:

1. The keyframes A & B the solver uses, if the movement of trackers between those two keyframes isn’t ample enough, the solver may have difficulties calculating effects due to perspective.
2. The camera’s settings in the motion tracking tab isn’t good. I think the solver uses this data to compensate for distortions and apparent distances between points, remember when Michael was tweaking the lens’ focal length to have a better solver. But if you followed along with him, you should have the right setup.
3. Some bad luck might be the case too

I hope this will help

Cheers !

[bjhackne11]

Just finished up this section, and had the same problem. I realize its a year after you posted your issue, so I am unsure if you still are following this, or if you solved it, however I think I know what happened.

I had this same issue, the below attached image is my tracker setup after setting the tracker scene up and switching to the default view.

After adjusting it to match the x-y plane through rotation and adjusting Z values, I got the same result others complained of where the camera perspective is just odd and difficult to use. In my case the camera ends up upside-down. See below image:

After some googling, it seems the issue comes from using footage that does not have a lot of forward/backward movement, and is mostly panning/left to right movement. The solver needs to use the forward movement to determine depth of the trackers. So in order to fix this, I recut my video to include a bit at the start where the forward movement happens, and retracked the camera. after setting up the tracking scene I got the below result instead of something similar to what you and I got on the first go around.

After rotating and changing z coordinate it looks like Mikeys example (except that one random tracker around 1BU too high):

Hopefully this helps you or anyone else that runs into this issue. The unfortunate part is that it seems like in order to fix the issue, I think that you would need to recut the video, retrack, and resolve your camera.