Colour Correction vs Grading: Do the Boring One First

Most colour problems in finished video come from doing one step. The editor drops a look onto the timeline, likes it on the shot they were watching, and moves on — and three shots later the same look has gone green, or crushed the shadows, or turned skin an odd shade, because it was added to footage that started somewhere else.

There are two separate jobs here and they are not interchangeable. Correction brings every shot to a neutral, accurate, consistent baseline. Grading takes that baseline and makes a choice about how the piece should feel. This guide covers why the order is not negotiable, what to read while you work, the sequence of correction operations, and where the whole thing tends to break.

Two Jobs, One With a Right Answer

Correction is technical, and it can be checked. Whites should be white, skin should sit in a plausible range, exposure should use the available scale without clipping, and two shots of the same scene should look like the same scene. When correction is done, nobody notices it — that is the point.

Grading is authored. Warm or cool, contrasty or flat, saturated or restrained, faithful or stylised. There is no correct answer, only a decision that suits the material, and it is the part that carries a piece's identity.

The order follows from the difference. A grade is a transformation applied on top of what is already there. Apply it to twelve shots that each start from a different white balance and a different exposure, and you get twelve different results from one operation. Correct first and the grade becomes a single decision applied to a consistent input, which means it can be copied across the whole timeline and will behave the same way everywhere.

Read the Scopes

The most important habit in colour work is distrusting your own eyes, and the reason is physiological rather than a matter of skill.

Human vision adapts. Sit in front of a slightly green image for twenty seconds and it stops looking green — your visual system normalises it, and from then on your eyes will confirm that the shot is fine. Worse, correcting a second shot afterwards will push it green to match what now looks neutral.

Scopes do not adapt.

Two readings do most of the work. On the parade, find something in the frame that should be neutral — a white wall, a grey surface, the whites of an eye — and check whether the three channels sit level there. On the vectorscope, check that the skin tone cluster lands near the line the scope provides, which is a reliable indicator across a wide range of complexions.

The Correction Order

Each step changes what the next one is looking at, so doing them out of sequence means repeating work.

The last step is the one that produces the biggest visible improvement and the one most often skipped. A sequence of individually well-corrected shots that do not match each other still reads as amateur, because the change at every cut announces the edit. This matters most with footage from a second camera or a different day, which is the common case with cutaways — the point at which a shot gets rejected for "not fitting" when it simply had not been matched yet, as noted in choosing B-roll that earns its place.

Then Grade

With a consistent baseline, the grade becomes a single decision rather than twelve.

Start from what the piece is doing rather than from a look you liked elsewhere. A product demonstration wants accuracy and cleanliness, because the viewer is judging the object. A personal story can afford warmth and lifted shadows. A technical explainer usually wants to get out of the way entirely, which is a legitimate grade and not an absence of one.

Apply the grade to a group or an adjustment layer covering the whole timeline rather than shot by shot. That is the practical payoff of correcting first: one node, one set of values, and a change to the look is a change in one place.

Where a LUT is involved, treat it as a starting position rather than a finished result. A LUT expects a particular input — a specific log profile, correctly exposed — and produces its intended result only when it gets one. Feeding it uncorrected footage is where most of the complaints about LUTs actually originate.

Skin Is the Shot Everyone Judges

If a video contains a person, their skin is what the audience uses to decide whether the colour is right, whether or not they could articulate that. Nothing else in the frame is checked so automatically.

That has a practical implication for where you spend effort. A background that is slightly off will pass unnoticed; skin that is slightly off will not, even by a viewer with no interest in colour. When correction and grading pull in different directions — a look that wants to be cool, a subject who then reads greyish — the subject wins and the look bends around them.

The vectorscope makes this checkable rather than a matter of opinion. Skin tones cluster along a consistent diagonal on the scope across a wide range of complexions, because what varies between people is largely how far out along that line they sit rather than the direction. A cluster that has drifted off the line is telling you there is a cast, and it will keep telling you that after your eyes have stopped being able to see it.

Secondary correction is the tool when the subject and the room disagree — isolate the skin, bring it back to the line, and leave the rest of the frame where the grade put it. Used sparingly this is invisible; used heavily it produces the detached, cut-out look of a subject who does not appear to be lit by the room they are standing in.

Where It Breaks

Three failures account for most of it.

Not enough information to push. Heavily compressed footage carries limited colour data, and an aggressive grade exposes that as banding in skies and smooth gradients. The amount of headroom you have is a function of how the footage was captured and encoded, which is the subject of how bitrate, resolution and quality relate.

Grading on an unreliable display. A monitor with a strong factory profile will lead you to compensate for something that only exists on that screen. You do not need an expensive display, but you do need scopes, and you need to check the result somewhere else before delivering.

Losing the grade at export. A piece that looked right in the timeline and arrives washed out or over-saturated on the platform has usually hit a colour-space mismatch on the way out. The per-destination settings that avoid this belong with the rest of your delivery spec, in export settings for each major platform.

Next Steps

Open a finished project and put the parade scope up on three shots from the same scene. If the channels sit at noticeably different levels between them, that sequence was never matched, and matching it will do more for it than any look you could apply.

Then separate your correction from your grade in the node or effect stack, so the two can be adjusted independently. That single structural change is what makes the next look you try take five minutes rather than an afternoon. The rest of the finishing craft sits under editing craft and workflow.