Smear

Blackmagic classifies Smear under Resolve FX Temporal and marks it Studio-only. Its core purpose is to simulate motion blur by directly blending multiple frames rather than applying a spatial blur inside one frame or estimating optical-flow vectors.

Frames Either Side is the structural control. A value N extends the temporal window up to N frames before and N frames after the current frame, so increasing it lengthens the visible motion trail and expands the time span being sampled.

The captured Resolve 21 panel used by this article shows Frames Either Side = 5. That is a UI reference state, not a recommended preset: five frames on each side can produce a very dense trail on fast motion.

Smear should not be described as a repair algorithm that understands objects or reconstructs a correct motion path. It blends temporal samples; complex movement, occlusion and fast camera motion can therefore create ghosting or repeated object positions.

Luma Threshold is not an inter-frame difference threshold and not a generic sensitivity control. Blackmagic describes it as a brightness gate: low values allow only darker pixels to contribute, while higher values progressively include brighter pixels.

The captured Luma Threshold is 100.0. At that setting the luminance gate is broadly open, so most of the tonal range can contribute. If lamps, windows or specular highlights produce overly heavy trails, lower the threshold rather than raising it.

Chroma Threshold follows the same concept in the saturation domain. Low values keep contribution biased toward low-saturation pixels, while raising the threshold admits increasingly saturated colors.

The captured Chroma Threshold is 100.0. On neon, RGB lighting, vivid clothing or saturated graphics this can create strong colored trails; lower Chroma Threshold when color trails become heavier than the desired luminance smear.

Luma and Chroma Threshold are independent. They do not need to match: you can preserve a broad tonal smear while restricting saturated colors more aggressively, or do the reverse.

In the captured build, the Global Blend section contains Blend = 1.000. This is the final mix between the fully calculated Smear and the source. At 1.000 the effect is fully visible; lower values restore more source sharpness and reduce the visible trail contribution.

Blend is best adjusted last. If the temporal window is too long or thresholds admit the wrong image regions, reducing Blend only hides the setup problem instead of correcting it.

A useful diagnostic order is Frames Either Side first, Luma Threshold second, Chroma Threshold third and Blend last. This separates temporal length, pixel eligibility and final effect dosage.

On the Fusion page Smear remains a regular 2D image processor: its main image input receives an RGBA branch and its output remains RGBA image data. It is not Shape, Particle, 3D or vector-flow data.

An optional Effect Mask limits spatial coverage after the effect is calculated. It does not supply temporal samples and does not change Frames Either Side; it only determines where the processed Smear is allowed to contribute.

If the mask follows a moving subject, track it separately. A static tight mask can clip the trail or reveal Smear in the wrong region even when the temporal controls themselves are correct.

Never judge Smear from a single still frame. Frames Either Side only makes sense over time: a still can show ghost geometry but cannot reveal how stable or distracting the trail feels during playback.

Use both playback and frame-by-frame review. Playback reveals motion character; frame stepping exposes duplicated silhouettes, edge echoes and problems around occlusions.

Smear differs from Motion Blur. Motion Blur can use motion estimation to build blur along estimated movement, while Smear directly blends neighboring frames and can therefore reveal discrete positions of a fast-moving subject.

Smear also differs from Motion Trails. Motion Trails is designed around ghost-like copies and trail accumulation/transform controls; Smear is a more direct multi-frame blend with luminance and saturation contribution gates.

Ordinary Blur and Directional Blur operate spatially within one frame and do not know where an object was on adjacent frames. Smear uses a temporal neighborhood and therefore changes the appearance of motion even with no spatial blur controls.

On a static shot Smear may appear weak because neighboring frames are nearly identical. That is expected: temporal blending becomes visible mainly where image content changes position or appearance over time.

Hard cuts are a special hazard for temporal effects. If the sampling window crosses an edit, neighboring frames belong to a different shot and can contaminate the smear. Keep temporal sampling inside the intended shot boundary.

For titles and graphics with transparency, inspect the alpha edge separately. Multi-frame blending can create soft or repeated transparent boundaries that read differently over the real background than over a checkerboard or black viewer.

For performance planning, remember that Smear requests multiple frames around the current one. Wider Frames Either Side means more temporal reads and potentially heavier preview, render and caching cost.

For natural-looking softening, a short temporal window, selective thresholds and partial Blend are usually easier to control. For stylized dance, light trails or abstract motion, broader gates and a longer window may be appropriate when ghosting is part of the design.

Finish production QC with three A/B tests: node bypass versus active, Blend 1.0 versus a partial mix, and short versus long Frames Either Side with identical thresholds. Those comparisons reveal where an unwanted trail is actually coming from.