3D tracking uses the 3 axes i.e X, Y and Z. X denotes the width, Y denotes the height and Z represents the depth, this helps us to establish a 3D point in the video.
But first, to tack, the video needs to be distorted. To do this, we added the Lens Distortion node and the distorting can then be done in two ways: the automatic way is to go under analysis>grid detect>current frame>click detect> click solve under editing and drawing, the other way to do it is the manual method in the Lens Distortion node, under analysis>editing and drawing>click add lines> draw lines in width and height on a building or the horizon which will help the software to understand the distort in the video and fix it> click on solve it.
To track, we need to add a camera tracker node. The node will analyse the video and then place points, it is also better to mask out any reflective surfaces so the tacking will be perfect. Add a roto if there are any reflective surfaces> added to the mask input on the node. In the Camera Tracker, there are a lot of options. In Source there are 2 options one stills, and the other, sequins. Still is one frame and sequins are the whole video. In Mask, there are various options but if anything has been masked out the setting should be changed to Mask Alpha. The other option is for the camera and its settings.
In the camera tracker setting, under features, the Number of Features is the number of tracking points and the preview features must be switched on so that the points can be viewed. Go back to Camera Tracker, click on Track then click Solve, the Error should not go over 1.0, then go to the Auto track and click on Delete unsolved and Delete Rejected.
After solving, the green lines will be visible, select any line and right-click it and there will be a toolbox opened. Select the ground plane>set origin, this will be the centre of the 3D world, now select a few points which are on the ground again go to ground plane>click on set to selected, this way the software will recognise the ground.
In the Camera Tracker, go to export in that Scene+ make sure you have switched on Link output.
In the Camera Tracker, select a point or more and right-click on create. There will be various options. If you want to add a 2d object in the face click on the card and place the card where you want the poster to be then add the image or video to the img input of the card. By pressing Tab on the viewer you can access the 3D view and move objects in 3D space.
For our homework, we had to play around with the camera tracker and this is what I created.
My biggest take away from here, apart from 3D tracking is, to change the anchor point in the manual in Corner points. So Roos and I are working on a group project, and I wanted help with it aligning the video in the right perspective, so I asked Gonzalo for help, and he showed me this easy way.
In Corner points setting, unselect all the enable and move then according to your requirements, you can then go to From click on Copy ‘to’ go back then click on Copy ‘from’ then select all the enable. This will allow us to edit to the right alignment.