StopMotion

StopMotion is a temporal tool, so a single still frame cannot prove that it is configured correctly. Its purpose is to change when the incoming image stream visibly updates, making real-time playback and frame-by-frame checks around jump points the primary quality-control methods.

The basic operation is frame holding. StopMotion samples one input frame, holds that state for several output frames, then jumps to another sampled state. Continuous motion becomes a stepped cadence reminiscent of stop-motion, animation on twos or threes, and deliberately low-update-rate graphics.

Frame Repeat is the main cadence control. The Resolve 21 reference documents a 1–10 range: low values preserve more temporal detail while larger values hold each sample longer and create increasingly coarse visible steps.

Do not confuse Frame Repeat with retiming. StopMotion does not lengthen or shorten the clip, alter audio speed, or generate intermediate states. Timeline duration remains unchanged while visual states update less often.

At 24 fps, Repeat 2 is a useful teaching comparison because it resembles an on-twos cadence, but it is not a universal creative preset. The same number feels very different on a slow title, a fast hand gesture, a handheld pan, or a long-lens shot.

Advanced Options is not an undefined utility drawer. The Resolve 21 reference explicitly documents Set Reference Frame, Reference Frame, and Sampling Variation inside this block, so it should be understood as phase and regularity control for the temporal sampling grid.

Set Reference Frame captures the current Viewer frame as the temporal anchor. This is useful when the cadence should line up with a gesture, edit point, musical hit, or the start of a graphic movement.

Reference Frame stores the frame number used by the skip pattern. Blackmagic notes that this number is calculated from the start of the entire original media clip, beginning at frame 0, rather than necessarily from the visible Fusion range or current edit point.

Reference Frame does not control effect strength. If the step size feels right but the jumps land at the wrong moments, change the reference phase. If the steps themselves are too large, change Frame Repeat. Diagnose those two problems separately.

Sampling Variation increasingly makes the temporal selection uneven. At zero or a very low value, the cadence is mechanically regular; raising it introduces irregularity that can feel more handmade or less mathematically perfect.

Sampling Variation does not replace Frame Repeat. Establish a readable hold length and a useful phase anchor first, then add variation. Otherwise it becomes difficult to tell whether a timing problem comes from step size, phase, or irregular sampling.

Fast pans, moving hands, thin diagonals, and fine texture expose coarse stepping much more aggressively than slow or static content. Evaluate the chosen repeat value on the fastest meaningful part of the shot.

Motion graphics and titles often tolerate larger steps because their contours and trajectories are predictable. On live action, the same settings can read either as intentional style or as broken playback depending on context.

StopMotion is useful for beat-synced motion. In that workflow, choose Frame Repeat first, anchor Reference Frame to the intended beat or action cue, then decide whether Sampling Variation is desirable. Too much irregularity can destroy precise musical sync.

A convincing stop-motion character does not require random cadence. Traditional frame-by-frame animation often has a very readable rhythm, so Sampling Variation is a character control rather than a mandatory ingredient.

If only a foreground element should be stepped, apply StopMotion to that image branch before Merge and keep the background branch smooth. Do not quantize the whole composite just because one object needs a handmade cadence.

Placement relative to Motion Blur matters. Blur before StopMotion causes already-blurred sampled frames to be held, while blur after StopMotion acts on the stepped result. Those are different looks, not interchangeable orderings.

Placement relative to animation matters too. A Transform before StopMotion creates smooth motion that is then temporally quantized. A smooth Transform after StopMotion can reintroduce continuous downstream movement, so put StopMotion after the animation that must become stepped.

When the source carries alpha, StopMotion holds the complete image state including alpha. Inspect both RGB and the Alpha Viewer because repeated semi-transparent edges can become obvious on fast motion even when opaque footage looks acceptable.

StopMotion does not analyze optical flow and does not synthesize missing positions. Compared with Speed Warp or Optical Flow, it intentionally removes intermediate smoothness instead of reconstructing motion.

It is also not equivalent to TimeSpeed. A retime changes the mapping between output time and source time and can change playback speed, whereas StopMotion preserves overall duration while repeating sampled frames according to its cadence.

The captured Inspector used by this article shows Advanced Options collapsed and Frame Repeat set to 5. That value is evidence of one saved state, not a factory default or a recommendation for every project.

Because Advanced Options is collapsed in the capture, the hidden controls are taken from documented Resolve 21 behavior rather than guessed from pixels: Set Reference Frame, Reference Frame, and Sampling Variation.

The reviewed article topology contains one primary image input and one image output. No separate Effect Mask is added without target-build socket evidence; when spatial localization is required, a separate branch and Merge is the safe teaching pattern.

Use a diagnostic loop with unambiguous motion such as a rotation, linear pan, moving title, or walking subject. A static image says almost nothing about temporal cadence.

For technical comparison, lock Reference Frame and set Sampling Variation to zero, then A/B only Frame Repeat. Once cadence is chosen, hold Frame Repeat constant and compare different reference phases separately.

Evaluate Sampling Variation over a sufficiently long loop. A difference can be hard to see in only two or three frames, while uneven timing becomes obvious across several seconds of repeated motion.

Do not compensate for a bad phase by increasing Frame Repeat. A larger hold only makes a poorly anchored skip pattern more obvious. Synchronize the pattern first, then art-direct the step size.

At shot boundaries, verify that the temporal effect is being applied only where the source stream is logically continuous. A stylized cadence should not hide an unintended edit-boundary problem.

For reproducible templates, record Frame Repeat, Reference Frame, and Sampling Variation explicitly. A note that says only 'stop-motion look' is insufficient because the same repeat with a different phase can land differently against the action.

A practical tuning order is: neutral variation, Frame Repeat, Reference Frame or Set Reference Frame, Sampling Variation, placement relative to animation and blur, then playback quality control.

Perform the final review at the target frame rate. Cadence designed at 24 fps will not necessarily feel identical at 25, 30, 50, or 60 fps even when the numeric Frame Repeat is unchanged.

The key mental model is not clip speed but image update cadence. Hold length, phase anchor, and sampling irregularity are three separate axes and should be tuned in that order.