-
Avoiding field mush when stabilizing interlaced footage
I’m stabilizing two clips in order to composite them and haven’t gotten any satisfactory results with any of my experimentation so far… which has resulted in various flavors of field mush. In addition to interpreting the footage to separate the fields and rendering fields on output, I’ve told the motion tracker to “track fields”, and have even written an expression on the resulting anchorPoint keyframes to round them to the nearest multiple of 2. I’ve also tried preprocessing the original footage with the FieldsKit DeInterlacer. What I haven’t yet tried is doubling the comp framerate to 59.94 so each field gets its own frame.
Are the following assumptions correct?:
1. If you do not separate fields and the comp framerate is 29.97fps, a single progressive frame will consist of two interlaced fields, 1/59.94th of a second apart in time.
2. If you do separate fields and the comp framerate is 29.97fps, each frame will appear to contain only its first field– the second will be interpolated. In this case, the second field of the footage will be used for the second field of the render when rendering fields. If you do not render fields, will the second field of the source footage be discarded?
3. If you do not separate fields and the comp framerate is 59.94fps, the frame will consist of the two interlaced fields (as in “1.” above), and that frame will be duplicated.
4. If you do separate fields and the comp framerate is 59.94fps, each field of the source footage will get its own frame.
Do any of you have experience doing this? Any suggestions?
Any pointers are much appreciated!