-
Difference between Greenscreen Shot and Mo Cap Data
Hi,
After shooting a greenscreen foreground using a 30′ crane equipped to capture motion data. I’ve found that the encoders on the crane do not account for twising or bouncing motions of the arm. So, that CG generated backgrounds based on the mo cap data show a kind of idealized crane motion without twisting, bending and bouncing motions that are visibile in the greenscreen footage.
The problem: the two layers don’t line up.
I’m able to track the forground (on some shots) with FP Track and alternatively track some shots with a 2D tracker. Is there some expression or way of creating an expression that would compare the key frames created by the mo cap data with the FP Track keyframes or the 2D key frames (forgetting in this case about the z axis data) – with purpose being to find the difference between the tracks and try to isolate the bumps and twists in the greenscreen footage in order to add them to the mo cap data?
That was fairly complex, I hope I’ve stated that clearly. Please let me know if you have any questions. I should also state that I use expressions, but generally and very basic level.
Thanks you in advance for any help.
Eric
-
11 Replies
-
I think I understand what’s happening, but without the actual data to work with I’m not sure if I can offer a solution.
It’s one of those situations I think I could resolive if I play around with it in practice, but I can’t just offer a pure mathematical solution from theory.
But let me think about it… 😎If you could upload a snippet, that might help though.
-
I tried simulating your situation by creating a virtual camera move in a virtual room.
Then duplicate the comp, made it a green room with a character, but with some wiggle added to the camera.I then added both plates to a new comp, keyed out the green, and now I have a FG that’s slightly off from the perfect camera move of my background plate.
I tracked a common point in both comps (the corner of the room). if you don’t have a common point in your plates, then find a point that tracks well in your green screen footage, and create a new 3D background where you create a virtual tracking marker in your 3D scene that matches that position.
when I had the two tracks I pickwhipped this simple expression on the position of my (more jittery) foreground:
diff=motionTracker(“Tracker 2”)(“Track Point 1”).featureCenter-thisComp.layer(“CG background”).motionTracker(“Tracker 1”)(“Track Point 1”).featureCenter;
value-diffThis worked reasonably well unless I start adding enormous amounts of wiggle.
I’m not sure if this fully mimics what’s happening in your real world situation, bu it might set you on your way.
-
Hi Filip,
Thanks for your help. I’m going to give your suggestion a try and I’ll get back with the results.
Best,
Eric
Eric Goldstein
Giraffe Film Company
Los Angeles
eric@giraffefilm.com -
Hi Filip,
I’m using two layers,CG background and greenscreen. I tracked each. I didn’t apply the tracks to either layer. I placed the expression in the “more jittery” greenscreen layer. I changed the formula to accomodate the tracking in two different layers. I got the error I included as a jpg., which also shows the expression and layer set up.
Where did I go wrong?
Thanks,
I think you just have to change the name/numbers of the trackers. Switch 1 and 2
Hi Filip,
I switched the 1 and the 2 as you suggested – although I don’t understand that part since tracker 1 is in the .png sequence not it the other. And I got this new error message.
Thanks,
Eric
Eric Goldstein
Giraffe Film Company
Los Angeles
eric@giraffefilm.com
OK I see,
after you subtract the two trackers from each other there’s a bit of extra code that shouldn’t be there:
after the 2ns “featureCenter” delete that final “-thisComp.layer'”200..”)”; from the the first line and it should work.
You probably pickwhipped that in by mistake.
The first line should just be: “my Tracker – the BG tracker”
Hi Filip,
I deleted the portion of the expression you indicated. But, I’m confused. Exactly what do I pickwhip, from where to where? And where is the expression placed – in which layer and what property?
Thanks,
Eric
Eric Goldstein
Giraffe Film Company
Los Angeles
eric@giraffefilm.com
Hi Eric.
You add an expression to the position of your Greenscreen Footage (more jitters):
– type in “diff=”
– pickwhip to the Trackpoint’s Feature center of the Greenscreen Footage
– tyoe a minus
– pickwhip to the Trackpoint’s Feature center of the 3D Footage
– type a semicolon
– 2nd line: “value-diff;”
The result should be that your Greenscreen is pinned to your idealized move.
If you want it the other way around where the end result retains the slightly jittered move you need to reverse the process:
You add an expression to the position of the Background Footage:
– type in “diff=”
– pickwhip to the Trackpoint’s Feature center of the 3D Footage
– tyoe a minus
– pickwhip to the Trackpoint’s Feature center of the Greenscreen Footage
– type a semicolon
– 2nd line: “value-diff;”
Reply to this Discussion! Login or Sign Up