How to color match multicam in Premiere Pro

The full color workflow for a multicam edit in Premiere Pro, from a rough cut to a matched timeline. Two parallel paths, the manual one editors learn the hard way and the MOVON one that collapses it to a single Analyze pass, with every Lumetri step named.

Most multicam color tutorials online stop at the point the editor actually has the problem. They cover one clip at a time, treat color matching as a creative grading exercise, and assume only two cameras of the same brand. The footage on a real timeline is rarely that polite: three or five or seven cameras, different sensor families, mixed LOG profiles, an operator who took the B-cam off auto white balance once and forgot, and a deadline.

This is the end-to-end workflow for that timeline. From a fresh rough cut to a matched, gradable sequence inside Premiere Pro. Two paths run side by side throughout: the manual workflow every editor learns by losing a weekend to it, and the MOVON Match workflow that does the matching pass on its own. Both paths share the same upstream and downstream steps, so the comparison stays honest. Pick the lane that fits the project; sometimes you want the practice, sometimes you want the cut done.

Before the workflow begins

Three things should already be true on the timeline. If any are missing, fix them first.

Bins by camera. Every multicam shoot benefits from a one-bin-per-camera organisation in the Project panel. Drag everything from Cam A into a bin called *Cam A*, same for B, C, drone. Premiere can colour-tag the bins so the timeline shows which clip is which at a glance. This is not strictly required for MOVON (the model groups by reading the pixels, not the bin), but it makes the human side of the work tractable.

A rough cut. The story works, the pacing works, the audio is locked. The colour pass goes faster on a cut that is not still moving underneath, and re-running the match after a re-cut is cheap, so the order is: cut first, match second, grade third.

Display LUTs applied via Lumetri Color → Basic Correction → Input LUT. This is the most important upstream step and the one most often missed. Sony S-Log3 footage needs the *S-Log3 to Rec.709* conversion LUT in Lumetri's Input LUT slot on the source clip. Canon C-Log3 needs *C-Log3 to Rec.709*. DJI D-Log needs the D-Log conversion. Without this, the timeline is showing flat, low-contrast LOG signal that no match tool, manual or automated, can do anything useful with. The deeper story on why this matters lives at LOG to display: what actually happens.

A common mistake: trying to set the LOG profile in *Modify → Interpret Footage*. That dialog handles framerate, alpha, fields, and pixel aspect ratio. It does not set the colour space for a misread LOG clip. The fix lives in Lumetri's Basic Correction → Input LUT dropdown on the source clip itself.

Step 1. Identify what you've got

Before any match runs, the editor needs to know what's actually on the timeline. Two questions to answer.

Which LOG profile is on each clip? Right-click the clip in the Project panel → *Properties*. Premiere reports the codec, frame rate, colour space, and (when available) the LOG profile tag. Confirm each clip is what you expect it to be. Mixed-format ingest, especially through proxy workflows, sometimes flips a clip's reported profile. Catch it now.

Which camera is the timeline's hero? The hero is the camera that nailed the look on this shoot. Not the most expensive camera. Not the one with the most clips. The one whose whites are cleanest, whose exposure was honest, whose skin tones the project wants to keep. Sometimes it's the A-cam, often it's the B-cam in better light, occasionally it's the drone in golden hour. Pick it on look, not on price tag. The full argument for how to pick a hero lives at Picking a good hero camera.

Once both are answered, the rest of the workflow is either *make the others look like the hero* (manual) or *tell MOVON which is the hero and let it produce the matching LUTs* (automated).

Step 2. The manual path

The five-step workflow every multicam editor figures out the hard way. It produces a good match. It also takes most of an afternoon.

2.1. Confirm the Input LUT on every clip

Walk through the Project panel one camera at a time. Open the first clip, check Lumetri Color → Basic Correction → Input LUT. Confirm the right conversion LUT is loaded. If multiple clips from the same camera share a profile, select them all in the Project panel and apply the Input LUT once for the whole bin. Repeat for every camera group.

2.2. Scalar corrections in Lumetri

On a representative clip from each non-hero camera, fix the obvious things first.

• Temperature in Basic Correction nudges the whole frame warm or cool. Match the white balance of the hero. If the hero is locked to 5600K and the B-cam drifted to auto, this is where it comes back.
• Tint corrects the green/magenta axis, the one drone footage almost always fights.
• Exposure matches the brightness. Use the Lumetri scope (Window → Lumetri Scopes → Waveform) to compare hero and non-hero side by side. If the non-hero is half a stop hot, pull Exposure down accordingly.

These are scalar corrections: one number per axis. They get most cameras into the same neighbourhood. They do not, by themselves, get cameras to agree on every colour in the volume.

2.3. Match the mids with curves

In Lumetri Color → Curves → RGB Curves, work on each colour channel separately. The goal is to make the *midtones* of the non-hero clip agree with the hero's midtones on the parade scope.

• Pull the red curve's midpoint until the red parade matches.
• Same for green.
• Same for blue.

This is where the hours go. The midtones never quite settle in one pass. You'll pull red, find that the skin now reads slightly cool, pull blue, find the highlights got muddy, go back to red. Two-camera projects converge in twenty minutes. Three-camera projects take an hour. Four-camera projects with a drone take an afternoon, and the drone always loses the longest. Why the drone is the worst version of this problem lives at The drone clip always loses.

2.4. HSL Secondary for skin tones

After the parade matches, the skin sometimes still reads off. The eye is much more sensitive to skin than to walls or skies, so this is where the audience picks up a mismatch even when the scopes look agreed.

In Lumetri Color → HSL Secondary, pick the face with the colour picker, refine the mask until just the skin is selected, and nudge Hue and Saturation in tiny increments. Five points or less. Larger moves are the creative grade, not the match.

2.5. Comparison View to verify

The eye adapts. Looking at one clip for ninety seconds tells the brain that this is the new normal. The fix: Premiere's Comparison View in the Program Monitor (toggle from the wrench icon menu, then split the view between hero and non-hero). Comparing two frames side by side defeats the eye's adaptation and surfaces the gap that's still there.

A separate tutorial on scope-based verification lives at Scopes, not eyeballs: how to verify a match.

When the side-by-side looks agreed and the scopes agree, the manual match is done.

Step 3. The MOVON Match path

The same upstream prep (Input LUTs in Lumetri, rough cut locked, bins organised by camera). The downstream creative grade is the same. Only the matching step itself collapses.

3.1. Open the MOVON panel

After install, MOVON docks as a UXP panel in Premiere. *Window → Extensions → MOVON Match*. Click to open. The panel shows the open project's timeline and a single Analyze button.

3.2. Hit Analyze

MOVON samples a few frames from each clip on the timeline and runs them through the two on-device models: the backbone that produces a per-camera colour embedding, and the predictor that turns those embeddings into matched LUTs. Everything runs locally on the Mac. Nothing uploads.

When analysis finishes, the panel shows the plan: every clip on the timeline grouped by camera (by reading the actual pixel signature, not the metadata), and one group marked as the proposed hero.

3.3. Confirm or change the hero

MOVON's auto-pick is usually right but not always. If the project wants a different camera as the hero, click that group to promote it. MOVON re-runs the match against the new target. The other LUTs change to point at the new hero.

A common case: the A-cam carried the most clips so MOVON proposed it, but the B-cam in better light was the look the project wanted. Promote the B-cam. The match tightens.

3.4. Confirm the plan

When the hero is right, confirm. MOVON generates one standard 33³ .cube LUT per non-hero camera group and exports the files from the panel. The default location is the project folder under a *MOVON LUTs/* directory; naming convention is *<project>_<camera>.cube*.

3.5. Apply the .cube per camera

For each non-hero camera, select all its clips in the bin or on the timeline and apply the matching .cube via Lumetri Color → Creative → Look → Browse → pick the camera's .cube file. Repeat for every non-hero camera. The hero's clips need nothing applied; they are already the target everything else was matched toward.

The longer Premiere-specific walk-through is at Exporting and applying the .cube.

Step 4. Verify

Same verification step for both paths.

Play the timeline at speed. The cut that took an afternoon to colour-match feels different at speed than at single-frame inspection. Skin should read as the same skin across cuts. Walls should read as the same walls. The scene should feel like it was shot on one rig.

Use the scopes. Open Lumetri Scopes (Window → Lumetri Scopes). Set to Waveform RGB. Step through a few cuts. The parade traces of hero and non-hero should sit on top of each other at the midtones. They probably won't be identical at the highlights and shadows (that's the creative grade's job), but the midtones agree on a finished match.

Trust the side-by-side over single-frame inspection. If the cut plays clean and the comparison view holds up, the match is good. If a single frame looks slightly off but the cut plays clean, leave it alone. The eye reads skin from neighbouring frames; an isolated still is a worse test than a played sequence.

Step 5. The creative grade goes on top

This is where the project's actual look comes in: teal-and-orange, day-for-night, the editor's taste. The match step exists so that *one* creative grade can be applied uniformly across the cut, without each camera responding differently.

The cleanest pattern for the creative grade on a multicam project: an adjustment layer above the entire timeline carrying the Lumetri Creative settings, on top of the matched per-clip Lumetri instances. The match lives on the clips. The grade lives on one layer. Tweaking either does not disturb the other.

A grade applied *before* the match (or interleaved with it) means the match runs against the graded signal, not the matched signal it was calibrated on. The math goes wrong and the cameras start disagreeing again. Match first, grade second, every time. The full chain explanation lives at Where MOVON sits in your grading chain.

Common pitfalls

Putting the matched .cube in the Input LUT slot. The Input LUT is for the LOG-to-display conversion. The match LUT belongs in Creative → Look, applied *after* the Input LUT has done its job. Dropping it into the wrong slot stacks the conversions and produces a picture that looks fundamentally wrong, no matter how careful the rest of the workflow was.

Picking the most expensive camera as the hero by reflex. The A-cam is not automatically the hero. If the B-cam was in better light, the B-cam is the hero. Pick by look, not by price tag.

Trying to match in the creative grade. Matching and grading are two different passes. A creative grade on top of unmatched cameras produces a different result for each camera. Get the cameras agreeing first, then put the look on top of all of them at once. The longer argument for keeping the two separated is at Matching cameras is not grading them.

Forgetting the drone. Drone footage almost always needs more correction than the cinema cameras because the sensor is smaller and the colour science is different. Plan for the drone first, not last. If the drone resists every match attempt, it probably needs its own scalar correction pass before the broader match runs.

Letting the eye decide. The eye adapts. The vectorscope does not. When the cameras *look* matched but the scopes disagree, the scopes are right. Step through the cut on scopes, not on the program monitor alone.

Common questions

Does the manual workflow take the same time on every shoot? Roughly proportional to the number of camera groups. Two cameras, twenty minutes. Three, an hour. Four with a drone, an afternoon. The drone always extends the curve.

Can MOVON do the manual workflow's steps in any order? MOVON does the matching pass only. The Input LUT step (Step 1 manual) and the creative grade step (Step 5) stay manual; they are different jobs. MOVON sits exactly where Step 2 lives in the manual path, and collapses 2.1 through 2.5 into a single Analyze pass.

What if MOVON's match is too aggressive? Pull the Lumetri Look intensity back to 70-80% on the applied .cube. MOVON is calibrated to err toward over-matching on purpose; dialing back is faster than chasing missing colour. The full troubleshooting tree lives at Match looks too aggressive or off.

What if a camera ends up in the wrong group? Open the MOVON panel after Analyze, move the clip to the correct group manually, re-analyze. The auto-grouping handles 95% of cases; the override handles the rest.

Does MOVON work with proxies? It analyses whatever pixels Premiere is currently showing on the timeline. Proxies that preserve the original LOG colour space behave like full-res footage for the match. Proxies rendered to Rec.709 will skew the result; toggle off proxies before running Analyze, or generate proxies that keep the original colour space.

Will the matched .cube work in Resolve or Final Cut? Yes. It is a plain 33³ .cube file in standard format. A project matched in Premiere can finish in Resolve or Final Cut without re-running MOVON, just by carrying the .cube files into the other NLE's LUT folder. The cross-tool workflow is at Exporting and applying the .cube.

The shorter version

A multicam colour workflow in Premiere has the same shape every time. Input LUTs go on first via Lumetri Basic Correction. Cameras get matched to a chosen hero, either by hand through curves and HSL or by MOVON in a single Analyze pass. The creative grade goes on top, ideally on an adjustment layer over the matched per-clip Lumetri instances. The verification happens on scopes, not on the eye.

The manual path takes an afternoon and teaches the editor what's actually going on under the hood. The MOVON path takes a minute and frees the afternoon for the creative grade. Both produce a matched timeline; both leave the creative decisions to the editor.

For the technical background on *why* multicam footage never matches out of the box, the colour-science pillar is at Multicam color matching: the complete guide. For the question of static profile lookup versus adaptive analysis (and why MOVON does the second), there is a separate doc at How MOVON compares to other color-matching plugins.

The private beta is free through January 2027 and the download is here. The blog tracks the build status, honestly, in the meantime.

— The MOVON Labs team