Diagnosing a cast

Split the trace into its RGB parade and a suspected color cast becomes something you can test. Start with a surface that should be neutral in the scene: ideally a gray card or verified neutral prop. A white shirt, concrete, or overcast sky can be a candidate, but material color and illumination may tint it. On a balanced reference the three channels relate as expected; on a cast shot one channel may sit apart.

A cast scales each channel, so the gap between the three is widest where the surface is well lit and collapses toward black. Read it over a lit part of the reference, not its shadow side.

The repeatable read is: find a credible neutral candidate, confirm that it should be neutral under this light, then compare the channels over it.

On Resolve's Color page, keep the RGB Parade open while you work in the Primaries palette. Temperature and Tint are global controls there, so every move affects the whole frame. Core keeps this correction global on purpose. If one part of the frame needs a different balance from another, that becomes selective work later in the path.

```sketch { "kind": "parade", "skew": "green", "caption": "One channel sits apart over a verified neutral while the other two line up. Read the gap there." } ```

What qualifies is narrower than it sounds. The surface has to be one you know the truth about before the light hit it. That rules out most of what is bright in a frame. A red jacket is not a gray, and neither is anything that makes its own color. A neon sign is the loudest thing on the scope and tells you nothing about the shot.

Run that read on the washing line below. Production has confirmed that the clean white sheets are neutral under this daylight, so they can serve as the reference. Their brightness alone would not prove that. Sample a lit part of a sheet and compare the three channels there. The gap runs on both balance axes, so temperature and tint both have work to do.

```demo { "plate": "washing-line", "scope": "parade", "controls": ["temp", "tint"], "baseLook": { "temp": -18, "tint": 16 }, "sample": true, "challenge": { "kind": "seat", "chip": "IN LINE", "zones": [ { "read": "ceiling", "band": [87, 97] } ] }, "caption": "The sheets are a verified neutral under this lighting. Sample a lit patch, read the gap between its three channel markers, and work temperature and tint until they align without clipping." } ```

Temperature and tint are the controls that clear the cast, so a cast you can name is a cast you can correct. The parade is the instrument you work on while you do it. Find the gray, watch the three channels there, and move temperature and tint until all three match.

One warning about the surface you pick, because the session below runs straight into it. Temperature and tint are global: they move every pixel in the frame by the same ratio. A room lit by one source can be answered that way, and a room lit by two cannot. Put a window and a tungsten lamp in the same shot and the two halves of that room need opposite corrections. Level the channels perfectly over one surface and the other half of the frame is left leaning the other way. There is no global setting that satisfies both.

So the last part of a mixed-light balance stops being arithmetic off the gray and becomes a judgment about the whole room. You level the surface you trust. Then you look at what that did to everything else, and give a little of it back. A grade that reads slightly off over your reference and right across the room is the better grade, and a bench that scores the whole frame will say so.