PFTrack Support Community

It was my comment. They changed the way trackers add up, so you can properly group different trackers in the same tree downstream for different purposes. Before you would have to manually select which trackers you would use for Camera solve against the ones you would use for object solving, but now you can just assign an entire User Track node for it. In your solver, you can choose which tracker groups you want to use I deleted it because I have already asked and they already answered it here :p https://www.thepixelfarm.co.uk/group/pftrack-support-community/discussion/0abfc7ec-037f-4200-9b19-e3fb89a1f284

Hi Adam, below is a link to download the LiDAR data.

If it helps with the diagnosis, I resent the file to one student while troubleshooting and it didn't change the outcome. Thanks again for looking into this! https://we.tl/t-jC2PJOoBUa

Hello naayagam, welcome to the group!

Yes, you certainly can import a camera FBX file into PFTrack and then refine that to get a better quality track. Here's a basic workflow to get started with, although the exact steps you'd need to take are going to depend on the type and accuracy of your camera motion.

First of all, you need to setup your shot:

1. Set up your camera and lens in the 'Clip Input' node, including any lens distortion

2. Attach an 'Edit Camera' node and import your FBX file (make sure your frame rate is correct otherwise your camera motion will drift away from your background clip)

3. Create some tracking points. You can attach either a 'User Track' or 'Auto Track' node beneath the Edit Camera depending on your requirements.

Now you've got your approximate camera motion and some 2D tracking points you can start to refine your camera as follows:

4. Attach a 'Camera Solver' node and click the "Solve Trackers" button. This will solve for the 3D coordinates of your tracking points using your imported camera motion. Provided your imported motion is fairly accurate, this should give you a good set of tracking points to refine

5. Click the 'Refine All' button. This will refine your 3D tracking points and imported camera motion to give you a better result.

You can then make any adjustments to your trackers or camera and continue refining until you're happy with the solve.

If you want to use the imported FBX as a starting position to track further, you would need to set the clip in/out points in step 1 to match the motion you've got, and then after you've refined that you can reset them to your full clip and use the Extend buttons in the Camera Solver node to extend the solve outwards into other frames.

PFTrack's 'Survey Solver' can use 'LiDAR' scans to help solve the camera position. Could you share a bit more about your intended workflow so the group can provide more targeted advice?

Hello Naayagam,

The in-app 'AI assistant' in the ‘Help’ menu is the fastest way to guide you through the process. Simply ask your question, and it will provide step-by-step guidance on which nodes to use, the actions to take, and the buttons to press. If you run into any issues, a quick follow-up question can often give you exactly what you need, much faster than searching through videos.

If there’s enough interest and feedback, I’d be happy to create a set of combined learning resources for the website, but I’d need to understand where the AI tools we’ve recently added are falling short first.