DCTL vs LUT: what is actually different
A LUT stores a grid of colors and works out everything in between. A DCTL stores the math and runs it on every pixel. We measured the difference on real footage, and on a 65 point LUT you cannot see it. Here are the numbers anyway, because most of this category sells the opposite story.
Shooting today? PursuitLuts packs are LUTs, built at 65 points, and they work in any software.
What each one is
A LUT, a lookup table, is a list of colors in and colors out. A 33 point file holds 35,937 of them, a 65 point file holds 274,625. Your footage has millions of colors, so for almost every pixel the software finds the nearest stored points and works out the rest. That working out is called interpolation.
A DCTL, DaVinci Color Transform Language, is a small program instead of a table. Resolve runs it on the graphics card for every pixel. There is no grid, so nothing is worked out.
How big is the difference, really
We can measure this exactly, because we have both halves: the math our files are built from, and the files. We ran three of our own frames through each and compared them in CIEDE2000, the standard measure of color difference, where 1.0 is roughly the smallest difference a person can see side by side.
| File | Typical difference | Worst pixel | Pixels a person could see |
|---|---|---|---|
| 65 point LUT, what we ship | 0.06 | 0.84 | none |
| 33 point LUT, what most packs ship | 0.37 | 1.36 | 6% of them |
So on our own files the honest answer is: the interpolation error is there and you will not see it. Every pixel of a 65 point file lands below the threshold where a person can tell. On a 33 point file a few percent of pixels cross it, which is a real if modest reason to prefer the larger grid.
We shipped 65 point files for exactly this reason, before we had any product that benefited from saying so.
Where the error actually lives
Not where you would guess, and not where we first wrote. Sorted by how colorful each pixel is, on our own frames:
- Near neutral: average difference 0.095
- Mild color: 0.086
- Strong color: 0.048
The error is slightly larger near neutral than in saturated color. An earlier version of this page claimed the opposite, from a measurement in raw code values rather than in what the eye does with them. A big number in code values near white is a small number to a person. The correction stands as a reminder to measure the thing you are claiming.
So why buy a DCTL at all
Not for accuracy you can see. For what it can do that a file of fixed numbers cannot:
- Controls. Look, strength, exposure and warmth, as sliders on the node. Exposure runs in scene light before the look, which is where a cinematographer wants it and where a LUT cannot reach.
- One file, every output. TV, web and cinema from the same node instead of three files.
- Exactness as a guarantee. Our DCTL matches the math to 0.015 of 1023, measured by compiling it and running both. That is not a visible improvement. It is a promise that nothing was approximated.
What a LUT can do that a DCTL cannot
- Run anywhere. A .cube works in Resolve, Premiere, Final Cut, Avid, in a camera and on an on-set monitor. Here is how in Premiere and Final Cut.
- Run in free Resolve. DCTLs need Resolve Studio. In the free version a DCTL will not load at all.
- Be handed to someone who does not use Resolve.
Which one you want
- Free Resolve: the LUT. A DCTL cannot run for you.
- Studio, and you want the controls: the DCTL.
- Studio, but the look also has to reach a monitor, a camera or an editor on Premiere: the LUT, or both.
- Worried about interpolation: on a 65 point file, do not be. We measured it so you do not have to take anyone's word for it.
One honest caveat about fitted looks
Four of our five looks are math end to end, so they port to a DCTL exactly. Golden Hour is different: it was fitted from a hand grade, which makes it a measured grid by nature rather than a formula. In a DCTL it would still carry that grid, so it stays in the LUT pack.
Where this is going
Our DCTLs cover the cameras we have verified against real footage, which today means RED and Sony. Canon and Nikon follow once we have filmed those bodies, not before. The math behind all of it is published and checked against Resolve itself.
Updated 2026-10-06. All guides