You drop the LUT and the groom's sherwani turns into a black hole. The bride's hair fuses with the mandap drape behind her. The dark corner of the venue, where the family was sitting, is now a single flat slab of nothing. The LUT is not the villain here — the exposure handed to it is.
This is the most common email we get from editors who buy a LUT pack: "Your LUT crushes the blacks." And almost every time, when we ask them to send a frame and the node tree, the cause is the same. The LUT is being asked to grade a clip that was never lined up to the input it expects. It does its job — faithfully — on a signal that's already 1.5 stops too low. The shadows were already living near the floor before the LUT touched them. The LUT just sealed the lid.
This post is about the node tree we run on every Positiva grade, and why exposure normalization belongs before the LUT, not after. Once you wire it up this way, the same LUT that "crushed your blacks" yesterday will hold detail in the sherwani, the hair, and the dark corner today — without you touching the LUT itself.
The diagnosis: every LUT assumes an input
A look-up table is a fixed mathematical map. Input value X always becomes output value Y. There is no intelligence inside a .cube file — it doesn't know your scene was lit at f/2.0 in a candlelit haldi or at f/8 in noon Goa sun. It only sees code values.
Every well-built LUT is authored against a specific input exposure: usually S-Log3 at the camera-correct exposure (Sony's spec puts middle grey at 41% IRE on the wave-form for S-Log3). The LUT's curve is sculpted around that anchor. When your dailies arrive at, say, 26 IRE for middle grey — because your DOP underexposed by 1.5 stops to keep the FX3 quiet in a low-light reception — the LUT bends an already-low value further down. Shadows that should have landed at 8 IRE instead land at 0. They don't crush. They fall off the cliff.
This is why the same LUT that looks gorgeous on a friend's correctly-exposed test footage looks like a coal mine on yours. The LUT didn't change. The signal you fed it did.
The pre-LUT node tree
In DaVinci Resolve, this is the order we use on every shot. Four nodes, no exceptions:
Node 1 → Color Space Transform (log → Rec.709)
Node 2 → Exposure normalization (Offset wheel → mid grey ~40 IRE)
Node 3 → LUT (the creative look)
Node 4 → Creative trim (taste, per shot)Node 1 — Color Space Transform
The CST does the heavy lift of converting log to display gamma in a mathematically clean way. It is not a creative move — it is the unflattening. If you skip it and apply the LUT directly to log footage, you'll get a different flavor of broken (we wrote about that in a companion piece — see Why Applying a LUT Directly to S-Log3 Looks Flat for the long version). Set input as S-Log3 / S-Gamut3.Cine, output as Rec.709 Gamma 2.4, tone-mapping on Luminance Mapping, saturation mapping on. Done.
Node 2 — Exposure normalization
This is the node nobody talks about and the one that fixes 80% of "your LUT is too dark" complaints. Open the Primaries wheels. Find the Offset wheel (it's the fourth one, alone on the right). Pull up your Waveform on the scope panel. Look at where middle grey sits — usually a forehead, a grey wall, the bride's sari border that's neither black nor highlight. For S-Log3 graded to Rec.709, you want that value sitting at around 40 IRE on the waveform before the LUT is applied.
Drag the Offset wheel's master slider until that mid-grey reading climbs to ~40. That's the entire move. No curves, no lift, no gain. Just a translation of the whole tonal range up the waveform until the LUT receives the input it was authored against.
Node 3 — LUT
Now drop the LUT here. With the input exposure in the right neighborhood, the LUT's contrast curve lands the shadows where the colorist who built it intended — not at zero. The sherwani holds texture. The hair separates from the drape. The dark corner of the mandap shows you the family.
Node 4 — Creative trim
Final taste — a touch of contrast, a slight warm push in highlights, whatever the shot wants. Keep this node small. If you're doing big moves here, the LUT was probably the wrong choice for the look, not a fix-it.
Why offset, not lift or gain
This is the part that trips up people who've been grading for a while. Why use Offset when Lift would also raise the shadows?
Because Lift, Gamma and Gain each pivot the curve around an anchor point. Lift raises shadows but barely touches highlights — which means it squashes the relationship between the two. Gamma pulls midtones independently. Gain stretches highlights. All three change the shape of the tonal response.
Your LUT is a fixed shape. Feed it a different shape and the LUT compounds the distortion. Feed it the original shape, just moved, and the LUT does its job. Offset is the only tool in the Primaries that performs a uniform translation across the entire range — shadows, mids, highlights all rise together by the same amount. The relationships between tones stay intact. You're not editorializing the image. You're handing the LUT what it expects.
Use Lift after the LUT, in node 4, if a specific shot needs its own shadow lift. Never before.
The 3-frame test
Before you set this up across an entire timeline, sanity-check it on three frames pulled from your dailies. Open them as stills in the Color page gallery.
- One mid-grey frame. A wide of the venue at neutral light, or a clean medium of the bride at a steady exposure. Build your node tree on this one. Get mid grey to ~40 IRE before the LUT.
- One bright frame. Outdoor portico, a sun-flooded entry shot, the haldi yard at noon. Apply the same node tree. Look at your scope. If highlights are clipping at 100 IRE and clamping flat — dial the
Offsetback. The bright shot doesn't need as much push. - One dark frame. The sangeet stage during a low-light dance, the candlelit phera. Apply the same tree. If the shadows still crush after the offset move, you have an exposure floor problem at capture, not a grading problem. No node tree saves underexposed footage when there's literally no information in the file. That's a conversation for the DOP, not the colorist.
This test takes three minutes and tells you everything. If all three frames hold up, your node tree is sound and you can apply it as a group across the timeline (or save it as a PowerGrade for the next project).
The Positiva Bundle
Every LUT in our library ships with the exposure-normalization node pre-built into the PowerGrade. You load the grade, the Offset is already wired up — you adjust per shot instead of rebuilding the tree from scratch every time.
Premiere users: do the same thing in Lumetri
If you're not in Resolve, the principle is identical — the implementation is just a stack instead of a node graph. In Lumetri, the effect order on the clip determines the math order. Drag a Curves effect onto the clip above the Lumetri Color instance that contains the LUT. Or open Lumetri and put your exposure move in the Basic Correction tab while the LUT lives in the Creative tab — Basic runs first.
Use the master curve to lift midtones until the waveform shows mid grey near 40 IRE. Then let the LUT do its thing in the Creative tab. Same principle: normalize the input, then apply the look. If you ever drop a LUT in Premiere and the blacks slam shut, this is your fix.
What this changes about your workflow
Once you internalize the four-node order, "the LUT crushes my blacks" stops being a sentence you say. It becomes a diagnostic question: what is mid grey reading on the waveform before node 3? If the answer isn't around 40 IRE, you know exactly where to fix it. You stop blaming the LUT and you stop hunting for a "less crushy" LUT — which is a doomed search, because every LUT will crush footage that arrives 1.5 stops low.
You also stop fighting the grade in node 4 with desperate Lift moves trying to rescue shadow detail that, in many cases, is actually still in the file — just sitting in the wrong place on the waveform when the LUT runs. Move it first. Then run the LUT. Then taste it.
If you want to see how this fits into the larger LUT library and the rest of the post-production decisions we make on a wedding cut, our pack hub is at positivafilms.com/luts. Each pack ships with the exact PowerGrade described above — CST in, Offset wired, LUT loaded, trim node empty and waiting for you.