Order in which upper or lower field scans first (upper first vs. lower first) — determines temporal accuracy in deinterlacing and motion. Wrong dominance causes jitter.
When scanning video material, the devil is in the details — and field dominance is one of those details that determines the difference between clean motion and annoying flicker. Interlaced video stores two fields per frame: one contains the even lines, the other the odd lines. The order in which these fields are captured or displayed in time determines the exact temporal correspondence between image and reality. If the upper field is scanned first (upper field first, UFF), it's called upper-field dominance; if the lower field is scanned first (lower field first, LFF), it's lower-field dominance. It seems trivial — but it's not, because a mismatch will cost you time and quality later in deinterlacing or with motion-based effects.
In practice, you'll encounter this problem most strongly when importing archive material or working with mixed-source projects. HDV cameras, PAL material from broadcast archives, NTSC legacies — each source has its peculiarities. Incorrect deinterlacing with the wrong dominance assumption creates jitter in fast pans or detailed motion because the spatial interpolation operates on incorrect timestamp pairs. The flicker appears subtly, but your eye perceives it immediately. Modern NLE software (Premiere, Final Cut, Resolve) forces you to declare the field dominance during import or sequence initialization — don't ignore this dialog.
Field dominance is equally critical in VFX tracking and motion estimation. Tracker software interprets the temporal shift between fields; with incorrect dominance, points drift apart or create frame jitter. Even when time-stretching — for example, when you need to recover 24p material from interlaced NTSC sources (pulldown) — field dominance determines the quality of the reverse telecine. An error here is costly: you'll only see it during color correction or final pixel peeping.
Practical tip: Document the field dominance of your sources from the beginning. Ask the archive, ask the camera operator, or check the material's metadata frame. When in doubt: a quick deinterlace preview with both modes will immediately show you which one is clean. For professional material from broadcast sources, the information is usually embedded; for digitized films, you'll have to test manually. The two minutes of effort will save you hours in post.