Skip to content

PresetForge vs a LUT generator

A LUT generator and a preset extractor are solving the same measurement problem and writing it into two different kinds of file. The difference is what happens after.

Where using a lut generator wins

  • A .cube LUT loads in Premiere, Resolve, Final Cut, OBS, game engines and many cameras — far wider reach than any preset format.
  • Deterministic across applications: the same table produces the same numbers everywhere.
  • Real-time, constant cost, regardless of how complex the look is.
  • Can express arbitrary colour mappings that no set of sliders can describe.

Where measuring wins

  • The preset stays editable — every parameter is a number you can change.
  • Raw-domain white balance, which a LUT cannot perform at all.
  • Grain, vignette, clarity and texture, none of which a LUT can carry.
  • No interpolation error: the transform is evaluated exactly rather than sampled on a 33-point grid.
  • And you still get the .cube, generated from the same measurement, so the two agree.

If the look is only ever going into video, a LUT is all you need. If it is going into photographs — or both — you want the preset, and the LUT derived from it.

Both start from the same measurement

Given a before and an after of the same frame, the transform between them is a fact. What differs is how you write it down. A LUT generator samples that transform on a grid — usually 33 points per axis, 35,937 entries — and stores the results. A preset extractor fits parameters: curve anchors, per-band hue and saturation offsets, split-toning values.

Sampling is more general. Fitting is more useful. A sampled table can describe any mapping at all, including ones no slider set can express; a fitted parameter list can be opened, read, argued with and changed.

What only the LUT can do

Travel. .cube is read by essentially every colour-capable application, and it behaves identically in all of them because there is no interpretation involved — the numbers are the answer. A preset only means anything inside the Adobe raw pipeline, because it relies on Lightroom's specific implementation of each slider.

It is also fast and constant-cost, which is why real-time video and on-set monitoring use LUTs and not parameter lists.

What only the preset can do

Stay editable. Open a preset, see that contrast is +18, change it. Open a LUT and your only control is the mix slider.

Handle raw white balance in the raw domain, before demosaicing, using the camera's colour matrix — a LUT sits far downstream and can only imitate the appearance.

And carry the spatial parts of a look. Clarity, texture, dehaze, sharpening, vignette and grain all depend on where a pixel is or what surrounds it, so they are not colour functions and cannot be baked into a table at all. A look that leans on them will feel flatter as a LUT. What a LUT cannot do that a preset can goes through the full list.

The case for generating both from one source

The usual failure is drift. Someone builds a preset for stills, then separately builds a LUT for video, and six months later the brand's photographs and its video no longer match — because two people made two sets of decisions about the same look.

Deriving both from one measurement removes that. PresetForge evaluates the same look object across the 33³ cube to produce the .cube, and writes the same look's parameters into the .xmp. They cannot disagree, because there is one source. .cube export is a Pro feature; .xmp is available on every plan, including free.

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

Bring back a look you already made.

Upload the original and the edited version of the same photo. The measurement takes a few seconds, and exact before/after mode is on the free plan.