0

I'm currently trying to figure out the workflow for using the motion capture features in Blender for animation. I've figured out creating markers and tracking them, but I'm stuck at reconstruction. The one video tutorial I've found uses the Camera Solver constraint to project the points in 3D space, but I can't find a way to turn those points into emptys so I can continue with the workflow. Alternatively, the normal reconstruction option in the Movie Clip Editor projects all the points onto a flat Y plane. Would there be some way to effectively combine these 2 effects, getting usable empty tracks in the 3D view that have some depth to them?

I'm currently in v 2.79.

Edit: the motion is from a single camera, and the camera motion is solved. The tutorial I was following used the Camera Solver modifier to calculate 3D depth for some of the points, so I was mostly looking to see if there was some method to get some sort of depth data with just one camera. It seems like that might be impossible, however.

edit2: The video has a stationary camera and the video subject is someone moving their face with control points painted on.

Ray Mairlot
  • 29,192
  • 11
  • 103
  • 125
Synthetic_Thought
  • 1
  • 1
  • Do you mean you are trying to get 3d information out of a single image? –  Jul 03 '19 at 23:30
  • https://blender.stackexchange.com/questions/33625/still-frame-camera-solve/33628#33628 and https://blender.stackexchange.com/questions/19126/motion-tracking-is-oriented-incorrectly-no-depth/19148#19148 and https://blender.stackexchange.com/questions/58098/problem-with-camera-orientation-in-motion-tracking/58120#58120 –  Jul 03 '19 at 23:31
  • Not quite, I'm trying to get depth information from several markers in a single 2D video. – Synthetic_Thought Jul 03 '19 at 23:33
  • Again. What kind of movement are you trying to track? A stationary camera paning and tilting or a camera that is walking or being displaced. Please read the links above and add information on the question, not as comments. –  Jul 03 '19 at 23:36
  • Read: https://blender.stackexchange.com/questions/42329/how-can-i-get-better-results-when-doing-camera-motion-tracking –  Jul 03 '19 at 23:36

1 Answers1

0

Given that you are dealing with a stationary camera you can't determine 3d information by solving the camera. You need to use the trackers to solve an object.

enter image description here

enter image description here

If you can solve the object correctly you should get tracking information in 3d space. Then you can use the link empty to track to get what you need.

Keep in mind that if the object you are tracking doesn't move much and there is no parallax information, blender will not be able to determine 3d information correctly. Currently there are no photogrammetry tools built into blender.

Read:

Object Solve option not avaliable

Old Answer(before the OP explained the kind of scene being tracked) You need to solve the scene first.

Then use the option in the solve menu: Geomety > Link Empty to track.

enter image description here From the Manual:

Link Empty to Track

Creates new empty in 3D viewport and appends contraint which parts it to the active track.

Community
  • 1
  • That does the same thing as the option in the reconstruction tab, it creates emptys in the 3D scene that are all flat on the Y plane. I was wondering if there was anything that could give me depth like the Camera Solver constraint does, but gives me usable emptys as well. – Synthetic_Thought Jul 03 '19 at 23:14
  • You need to add more information. Did you solve the scene?, What kind of motion are you solving? If you are solving tripod information you will never get 3D info, shots with parallax will give you 3D. Please edit your question and add relevant information. –  Jul 03 '19 at 23:16
  • Read edited answer. –  Jul 04 '19 at 03:43