Last week I ventured out to capture some footage with a mind to incorporate it into the final piece.
One of the things holding me back from making progress on this film has been the planning stage. Planning is in no way one of my strong points. Arguably I should be pushing myself into the process, however at the point where it's holding back a project, my best bet is to play to my strengths and improvise and be inspired as I work.
The following footage is the result of the shoot:
It may be very raw and poorly edited, but the main purpose of this was to work out which scenes would require match-moving and which would need reshooting.
The next stage is to take the parts that will require match-moving, then take them into Matchmover/Boujou in order to test the effectiveness of the tracking.
(Currently, I am testing the footage in Boujou, as it handles motion blur far more effectively than Matchmover. Also, it's masking and manual tracking tools are far more intuitive to use.)
Once loaded into boujou, the first step was to mask out Andrew in order to prevent boujou from tracking him.
A few attempts to solely use automatic tracking were met with limited success. There was a lot of slipping around the frames where the camera movement was particularly harsh.
Due to this I incorporated a number of manual tracks, picking points that I could make out when the motion was particularly blurred.
Using the manual data I then ran another automatic track, which resulted in a successful 3D solve:
This is the main shot that will contain matchmoving, a couple of others will too. Particularly the post-roll shot. Then preceeding rolls are so short, that matchmoving is possibly not necessary, and I could simply represent blurred bricks hurtling across the screen. If that method is successful, then I will do so. If not then matchmoving it is.
Next post I'll go into more detail about the film itself and the considerations behind the shots and editing.
This comment has been removed by a blog administrator.
ReplyDelete