What a LUT cannot do that a preset can

A preset is a recipe: a short list of parameters that an application interprets. A LUT is a result: a dense table of input colours and the output colours they become. Everything below follows from that one difference.

1. A LUT cannot be edited

This is the big one. Open a preset and you can see that the contrast is +18 and change it to +12. Open a LUT and you see 35,937 lines of numbers. The only control you have is the strength or mix slider in whatever applied it, which blends between the original and the fully-LUTted image. You cannot warm only the highlights, or pull back only the greens, without doing that as a separate operation afterwards.

2. A LUT cannot do raw white balance

Temp and Tint on a raw file are applied in the raw domain, before demosaicing, using the camera's colour matrix. A LUT operates on already-rendered RGB, far downstream. It can simulate the appearance of a warmer image, but it cannot perform the operation, and the difference shows up immediately in extreme corrections and in how highlights behave.

3. A LUT cannot do anything spatial

A LUT maps colour to colour. Give it the same input twice and it must produce the same output twice. That rules out every operation whose result depends on where the pixel is or what surrounds it:

• Clarity, Texture and Dehaze — these compare a pixel to its neighbourhood.
• Sharpening and noise reduction — same reason.
• Vignette — depends on frame position.
• Grain — depends on a random field, so it is not even a function.
• Masks, gradients and brushes — per-region by definition.

Video tools have their own versions of most of these, so nothing is lost permanently — but they are separate operations you add on top, not something the LUT brings with it.

4. A LUT has limited precision at the extremes

A 33³ LUT samples each axis at 33 points and interpolates between them. Where the transform changes gradually that is invisible. Where it changes sharply — a hard shoulder, an abrupt hue shift at a narrow range of saturations — interpolation error shows up as banding. Going to 65³ helps at eight times the file size. A preset has no such limit, because it evaluates the actual function at every pixel.

What a LUT does better

The trade runs both ways, and the LUT side is not small.

• It works nearly everywhere. Video editors, colour grading suites, streaming software, game engines, on-set monitors and many cameras all read .cube. No preset format has that reach.
• It is deterministic. The same LUT applied in two applications produces the same numbers. Presets depend on how each application implements each slider, which is why a preset is really only meaningful inside the Adobe raw pipeline.
• It is fast and constant-cost. Applying a LUT is three table lookups and an interpolation, regardless of how complicated the look is. It runs in real time on video.
• It can express transforms no slider set can. A LUT can map any colour to any other colour. A preset can only express what its parameters allow.

Which to use

For photographs, use the preset. Keep the parameters editable, keep the raw white balance behaviour, keep grain and vignette in the same object as the grade.

For video, use the LUT, because it is the native form and nothing else will load. Add clarity, grain and vignette separately in your editor, and expect to want the look slightly stronger than the still version, because motion hides subtlety.

For a look that has to work in both — a brand that shoots stills and video, a photographer who also streams — you want both files generated from the same source measurement, so that they agree. That is why PresetForge derives one look and exports it three ways: .xmp for the Adobe pipeline, a 33³ .cube for everything else, and .json if you want the raw numbers to script against. Export of .cube and .json is a Pro feature; .xmp is on every plan.

For the mechanics of the conversion and how to load the result in each application, see converting a Lightroom preset to a .cube LUT.