I have a scene where I walk out from between two trees. I stop, turn back, speak to someone off camera and then walk out of frame left, laughing. I've attached three screen shots which I hope to illustrate my intent.
My intent is along the lines of having my shadow be left behind once I move on from the position I appear in Gnome-1 attachment. As I get to the position in Gnome-2 I want the shadow to begin to follow the Gnome. Gnome-3 shows how I move out of left frame.
I plan to represent my shadow by duplicating the original clip and using Colour Correct Mask to lower the Brightness and Saturation, perhaps with some blur or something to represent a somewhat mysterious figure that is the shadow.
I watched this video - which demonstrates the author's son walk into a freeze frame of himself a couple of times. I worked through this example with my footage, but now I want to walk out of a past freeze frame. I think the same technique will work in reverse, but have yet to try it out.
My apologies, I forgot to click Insert. I believe, given they are above, you'll be able to see them.
The Gnome is walking, he leaves himself frozen for a second or two as a shadow and then the shadow begins to follow the gnome. You are correct, it does do more, but given this was my first attempt at this, the youtube clip I posted seemed a good place to start.
Having attempted to do it using the example, I'm thinking there must be a simpler way, e.g. to follow your suggestion.
kindomalay wrote: … The Gnome is walking, he leaves himself frozen for a second or two as a shadow and then the shadow begins to follow the gnome. You are correct, it does do more, but given this was my first attempt at this, the youtube clip I posted seemed a good place to start.
ahh, gotcha, Colin!
kinda sort of …:
For mimicking the outcome I used Pixelmators Magnetic Lasso to mask you out of the first shot - and that is really difficult! The recording is blurry, colors of foreground and background match… I hardly doubt you get a showable result.
In that respect, the usage of coremelts TrackX is correct - it is intended to 'key' (=mask) non-keyable moving objects…
At the school project, I support as a volunteer, we do a lot of 'silly' v/fx too… we won a prize for a 14min Fairy Tale with a magic dwarf, doing all nonsense.
It is essential to 'plan' such v/fx way before production .
As an example: your camera is moving = parts from a different TC will not match! Or, when I do know, I will key some material, the scene has to be made key-able as possible. In your case: two recordings, rock solid fixed cam, gnome running thru scenery; 2nd take: two assistants hold a green blanket into scenery (for same light conditions) , Gnome runs again.
=> you get an easy to key clip, adjust colors/blur/opacity to make it 'ghostly'…