In my many attempts in PFTrack, the biggest problem that I can say occurs the most is getting an accurate focal length estimation for a zooming free motion shot. Now my process may be completely wrong from what is supposed to be done, which is why I wanted to ask the community if they have a working method they use which could benefit us all.
My process is as follows:
Step 1.) After getting my manual and autotracks setup and solid I use the estimate focal length box on the starting frame to get the FL to just about where it should be.
Step 2.) If the box does not (which is normally the case) give accurate results, I use Scene Orientation lines on 3-4 frames ranging throughout the shot. After lines are set and (camera>estimate focal length) and that value is stored I solve again.
Thats my process and most of the time does not get anywhere near what I need. For instance an aerial shot orbiting a building with a zoom. Start frame is a solid 25 mm FL and end frame is stored as around 105. After the solve however the start frame is set at 3.—- mm or even .0000143 This number varies from time to time depending on the orientation lines setup.
The other issue that occurs is: If I do get an accurate solve and FL estimation, on the frames where the lines were drawn there is a dramatic change in FL resulting in a pop.
I apologize for the length of this post, but im sure if im having these issues others will eventually as well and I might as well start it.
Thanks in advance guys! :niceone:
Post a Comment