I have a video where I walk through a hallway into another room, and I would like to track the whole video.
I do this, but going through the video and add tracking points whenever new areas/rooms appear.
But now I have to set the keyframes, and they need to have common markers. This means they can only be within a small part of the video (not to far away). I also notice that markers, that are not visible in the keyframes do not appear as empties when in the 3d scene with the tracked camera.
All of this makes me assume, that I do not have the Ideal workflow yet. So what is the proper workflow for this?
Have you thought of adding a path for the camera in the form of a Bezier curve, then adding a curve constraint or follow path constraint for the camera, then key framing it’s location and rotations along that path? This may be a more precise and easier managed solution. If not and you want to go this way, I can help here if you need it.
thanks for the reply. I want to track the camera because i want to add 3d objects to the Video Material. For this i need a very exact track of the camera. That is difficult if not impossible with a manual path, or am i wrong?
A bezier curve - bear in mind that a “curve” can be a series of straight line segments - would be the most accurate way I can think of to position the camera. You would know for instance that at any point in time it is, for example, exactly 0.56734 of the length of the curve from the start point or that it had moved, again for example, exactly 104.5436 units down the curve depending on how you set up the Curve or Follow Path modifier/constraint. Would it be helpful if i made a sample blend file so you can see how it works? The other good thing about using a “track” for the camera is that the movements of the camera are endlessly and exactly repeatable within not only your current project, but also any project you append the track and camera to. This is the Blender equivalent to a “Motion Tracked” or “Rail Mounted” cameras in real life, where the exact track of the camera can be reproduced to enable shots to be overlaid with precision fit.
You can of course always make edits to the track to refine the camera position, these are stored and will be reproduced next time you use the track, and you can have multiple tracks for one camera, then you can keyframe the influence of each constraint to pass the camera from one track to the next.
Doing it manually, particularly trying to reproduce the track in separate blend files would be close to impossible IMHO.