Color-Matching A7S III to FX3 on a Wedding Multicam Edit (Resolve, 2026)

You cut the bidaai sequence together, A-cam to B-cam to A-cam, and the bride's lehenga shifts hue every time the edit returns to the FX3. The cameras are supposed to be twins. They are not behaving like twins.

This is the most common color complaint we hear from Indian wedding shooters in 2026, and it has a clean technical answer. The A7S III and the FX3 share the same 12.1MP back-illuminated sensor and most of their image pipeline. On paper, they should match out of the camera. In a real edit, they don't — not catastrophically, but enough that a viewer who never thinks about color will still feel the cut.

What follows is the exact workflow our color desk runs on every Sony two-camera wedding job. It assumes you are in DaVinci Resolve, you are shooting S-Log3/S-Gamut3.Cine on both bodies, and you have a 200+ clip timeline you cannot afford to grade per-clip.

Why two "identical" cameras drift

The sensor is the same. The colour science branding is "the same." But the firmware paths are not. The FX3 is positioned as a video-first body, and Sony's video-pipeline gamma processing — especially the way Cinetone influences the in-camera image even when you think you've bypassed it — nudges things slightly differently from the A7S III's stills-leaning pipeline.

Even with both cameras locked to S-Log3/S-Gamut3.Cine, identical white balance, identical exposure index, identical lens, you will see two consistent drifts on a parade scope:

• FX3 sits a touch warmer and slightly magenta in the midtones — bride's skin reads pinker.
• A7S III sits a touch cooler with a green lean in the same midtones — the same skin reads more olive.
• Luma reads marginally darker on the FX3 at the same metered exposure — not a stop, but enough for the scope to register half to one and a half IRE.

None of this is broken. It's the cost of two different firmware teams shipping two different products. Your job in post is to neutralize the gap before the look LUT goes on, not after.

The five-minute capture step that saves you hours in post

Before the haldi starts, before the makeup tent gets crowded, get both cameras together for ninety seconds. We shoot a grey card and a Macbeth chart with both bodies, identical exposure, identical white balance target, lit by the same source. Three seconds of clean chart on each is enough.

This is your ground truth. You will use it twice — once to derive the offset, and once to validate the offset is dialled in correctly. If you skip this and try to match by eye on real footage later, you will spend an hour chasing a number you could have measured in two minutes.

Sync timecode while you're at it. Multicam without genlock is bad enough; multicam where you can't even auto-sync is worse.

The Resolve match workflow

The whole point of this pipeline is that it survives a 200-clip timeline. Group grades, not clip grades. We will say this twice because everyone learns it the hard way.

Step 1 — Build the A-cam base grade

Select every A7S III clip, group them as "A-CAM." On the group's pre-clip stage, build:

• Node 1: Color Space Transform from S-Log3/S-Gamut3.Cine to DaVinci WG/Intermediate (or to Rec.709 if you're going straight out, but we prefer working in a wide intermediate). If this part feels handwavy, our walkthrough on why your S-Log3 LUT looks flat without a CST node covers exactly why the order matters.
• Node 2: Exposure normalization — lift or pull until your skin tone hits the IRE range you want before the look goes on. Our pre-LUT exposure workflow piece goes deep on the trim values we use for Indian skin under tungsten and LED.
• Node 3: Look LUT or PowerGrade.

Step 2 — Apply the same group grade to B-cam

Select every FX3 clip. Group them as "B-CAM." On the B-CAM group's pre-clip stage, paste the identical three-node tree from A-CAM. Do not modify it.

Now park your playhead on a B-cam clip that cuts directly against an A-cam clip — ideally a wide of the mandap that has both cameras' angles in roughly the same light. Open your parade scope. You will see the drift clearly: the red trace on FX3 sits a hair higher, the green trace sits lower, the whole luma sits half an IRE south.

Step 3 — Add the offset node, B-cam only

On the B-CAM group's pre-clip stage, add a fourth node before the look LUT. This is your match node and its job is to make the FX3 pretend to be an A7S III.

Typical correction values, ballpark, for two bodies on default firmware:

• Hue offset: rotate the global hue wheel toward green, between +1 and +3 degrees. Counters the FX3 magenta lean.
• Tint: -1 to -3 on the temperature/tint pair, biased on tint. Same job, finer control.
• Gain or offset: lift luma by +0.5 to +1.5 IRE — whatever brings the FX3 trace level with A7S III on the parade scope.
• Optional: a tiny saturation pull, -2 to -3, if your FX3's midtones feel slightly more aggressive in chroma.

Those numbers are starting points, not gospel. Your two specific bodies will have their own offset, and that is what the chart you shot at the start of the day is for.

Step 4 — Validate against the chart

Drop the chart clips from both cameras onto the timeline, side by side, in the same group context. Use the picker on the same neutral grey chip on each. The RGB values you read out of the picker tell you exactly how far apart the bodies are, and you can dial your match node until the two readings collapse into the same number. This is the only objectively correct way to do this. Eyeballing it on a hero shot will get you 80% there; the chart gets you to 100%.

Group grades, not clip grades

The whole point of putting the match on the group's pre-clip stage is that you set it once and 200 clips inherit it. If you fix the magenta lean per-clip on a wedding edit, you will be fixing it for the rest of your life, and the next round of revisions will undo half of it.

Per-clip nodes exist for problems that are actually per-clip — a window light blew out one shot, a guest stepped in front of the lamp during the pheras. Camera-to-camera drift is not a per-clip problem. It's a per-camera problem. Solve it where it lives.

If you re-cut and the timeline grows from two cameras to three, your grouping survives. You add a third group, paste the base tree, derive its offset against the chart, done.

What about the A7 IV in the same edit?

This comes up on bigger weddings: the second shooter rolls in with an A7 IV for stills-and-clip duty and you end up with twenty B-roll clips on the timeline. Don't drag those into the FX3 group. The A7 IV is a different sensor with a meaningfully different colour pipeline, and Sony's Cine profiles bake more aggressively on that body. Build it its own group, derive its own offset against your chart, accept that it may need a slightly heavier hue rotation than the FX3 does. Trying to wedge it into the FX3 match will cost you more time than just giving it its own three-line group.

Where this leaves you

The match itself is small — one node, four to six values. The discipline is in doing it once at the group level instead of fighting it across 200 clips. Shoot the chart. Build the A-cam tree. Mirror it onto B-cam. Add the offset node. Validate on the parade scope using the chart. Move on to the actual creative grade, which is the part of the job worth your attention.

If you want a head start, the offset preset and the full base node tree are both inside our LUT library — built specifically against the two-Sony wedding setup that runs most Indian weddings in 2026.