[Aconover]

I’ve been having this same problem as well for a long time. All I’ve been able to figure out is that earlier dot-versions of Final Cut don’t have the problem, and that it’s definitely a bug, because if you check the clip settings it will say the clip is NDF as well. I believe that if you were to log and capture 50 clips, as the tape goes on the in and out points of the captured clips become more and more inaccurate — I’ve worked around it by using Capture Now and breaking the clip up into subclips. However, this does make recapturing later (if you want to up-rez, say) a tricky business — sometimes your clips will be improperly aligned (because of in-out drift) and you’ll have to go readjust them. Unfortunately, I have not found a real fix — after hours of Googling for the past few months, I’ve given up and am just waiting for the next revision to hopefully fix the problem.

[Aconover]

Well, I can’t afford Automatic Duck. I remember a year or two ago (before I started using AE in earnest) they had a free version of their FCP-to-AE export tool — I even still have a copy of it — but it doesn’t work with FCP5, and they seem to have discontinued it, frustratingly enough. What’s a poor, enthusiastic hobbyist to do

[Aconover]

Thanks, muiisal — this advice was excellent, and really helped me get through the project after the first few atrocious starts that prompted this posting. There are still some details of roto that I need to work out — specifically, how keyframes relate to control points (does each set of control points constitute one keyframe position, or is each point keyframed individually, etc?), but those are things I can pick up myself — your advice definitely got me started on the right path. Thanks!

[Aconover]

Thanks, that was enormously helpful and gives me a great starting point. One place I’m getting tripped up, though, is exactly how to keyframe / move the mask points. For instance, through most of the shot, the hand is positioned more or less the same way — all five fingers are more or less extended. Towards the end of the overlapping portion (see note below), however, the hand dips down, forming more of a fist shape (i.e., no fingers). Should I create a new mask / new layer at this point, or should I keyframe the hand-shaped mask so that it forms a fist shape? To make matters more complicated, the character then stands up, again requiring an entirely different mask shape. What’s the best way to divide up these masks?

Note: The general shape of the shot is of one version of the actor sitting in a chair, and another version of the same actor squatting on the floor to the side / slightly behind him. At one point, the foreground character moves his hand in front of the other’s face — later, he stands up and begins to move across the second character. So I started with a simple “curved line down the center of the frame” mask, then added another version of the topmost layer (with the seated character), in which I masked the hand separately, starting when the hand first moves over the BG character’s face. Should I then create a third mask for the segment where the first character moves bodily across the second, or am I approaching this in the wrong way? Let me know if this is difficult to visualize and I will attach some screenshots. And thanks again for the advice!