Free · no account · nothing is uploaded
Preset and LUT converters
Every direction between Lightroom presets and 3D LUTs — and, for each one, what it actually keeps.
Convert a format
The two formats name their develop settings identically — this changes the syntax, not the meaning. Nothing is estimated.
The global grade converts exactly. Masks, vignettes and grain cannot: a LUT maps colour to colour and knows nothing about where in the frame a pixel sits.
Same as above, written in Autodesk's format at 12-bit with the entry order transposed to blue-fastest.
Same grid, same values, reordered. The only loss is integer quantisation — about 0.012% at 12-bit, well below anything visible.
Lossless: integers become decimals with room to spare. The bit depth is inferred from the file and displayed, because .3dl does not record it.
A 33³ LUT holds 35,937 independent values; a preset holds roughly sixty parameters. Tone curves are extracted exactly, colour is estimated, and the tool renders both so you can see the residual.
Change a LUT without changing its format
Blend a LUT toward the identity transform and bake the result, so a reduced look survives into software with no opacity control.
Change the grid sizeResample between 17³ and 65³, with the worst-case channel shift measured on your file rather than assumed.
Inspect what it doesPlot a LUT's tone response and hue shifts before trusting it — 35,937 numbers tell you nothing by inspection.
Convert a whole folderAny number of files in one pass, returned as a single archive. No cap and no account.
Why the fidelity column is the useful part
Plenty of sites list which conversions are possible. Almost none say which ones cost you something, which is the only question that matters once you are choosing between them.
Two of the rows above are not lossless, and neither is fixable by writing better code. A preset carries instructions that apply to parts of a frame — a graduated filter over the sky, a brushed adjustment on a face. A LUT is handed one colour at a time and told nothing about where it came from, so there is nowhere in the file for that information to live. And converting back is worse: a 33-point LUT holds 35,937 independent colour triplets against a preset’s sixty-odd parameters, so most LUTs are simply not expressible as a preset at all.
That is why the LUT-to-preset tool renders the same frame twice — once through your LUT, once through the generated preset — and reports the measured error. Showing the residual is the only defensible way to ship an approximation.
Frequently asked questions
Which conversions lose something?
Two. A preset to a LUT keeps the whole global grade but cannot carry masks, vignettes or grain, because a LUT maps colour to colour and has no idea where in the frame a pixel is. A LUT back to a preset is a genuine approximation: 35,937 independent values do not fit into roughly sixty parameters. Everything else on this page is exact apart from integer rounding.
Do I need to install anything?
No. Every converter here runs in the browser using the File API and your own GPU. Nothing is uploaded, there is no account, and the pages keep working with the network disconnected once loaded.
Can I convert a whole preset pack at once?
Yes — the batch converter takes any number of files and returns them as a single archive, with no cap. The single-file tools are the better choice when you want to inspect one conversion rather than move a library.
Why does my converted LUT look different from the preset?
Usually input, not conversion. A look built on photographs assumes photograph-like contrast; applied to log footage it comes out flat because the LUT is being fed numbers it was not designed for. Convert the footage to the space the look expects first. The other common cause is a preset that leaned on masks, which cannot survive into a LUT at all.
You have the preset. Do you know what it does?
A preset is sixty numbers, and most of them barely matter. Drop the photo and its sidecar into the Studio and the edit replays in six stages — light, curves, colour, masks, finish — so you can see which decision actually carried the look rather than guessing from the finished frame.