Skip to content

Look development II — film emulation

IntermediateDuration ~13 min video + 45 min hands-onTools DaVinci Resolve (free)

“Film look” is the most requested and least understood thing in grading. Most people chase it by dropping a film LUT on a shot, and it either looks amazing on that one frame or falls apart on the next — and they have no idea why, because a LUT is a sealed box. In 2.9 you learned to analyze a look. Here you learn to build the specific one everyone actually wants: the look of photochemical film. The key that unlocks it is realizing a film look is not one effect but a chain of understandable steps — a negative, a print, a glow around the highlights, a layer of texture — each of which you can build and adjust. Once you see the chain, the one-click LUT stops being magic and becomes just one link you can place, tame, or rebuild at will.

Watch for: Melara takes the Kodak 2383 print-film-emulation LUT that ships with Resolve and rebuilds it, node by node, as a PowerGrade you can open up and edit. Watch the order he reconstructs it in: first the main contrast curve (he desaturates to see it as pure grayscale), then the color tint — where he shows the red, green and blue curves converging at the midpoint, meaning there is no tint in the shadows, a blue cast building below and warmth building in the highlights. Near the end (around 9:00) watch how he *uses* it: a color space transform to get from log into Rec.709 — the correct way, rather than pushing saturation to 100 — a balance node placed before the look, and the point that because the contrast curve is its own node, he can simply swap it out when a film LUT feels 'too contrasty.' That editability is the whole argument for understanding a look instead of trusting a LUT.

Start with what film actually is, because the whole look follows from it. Film emulation is reproducing photochemical film on a digital image, and real film is a two-stage system. There is a camera negative stock, exposed on set, and there is a separate print stock — like Kodak 2383 — that the negative is printed onto to make the positive you watch. The character we call “the film look” comes mostly from that print step and from how the two stages interact. Understanding this is what tells you a film look has structure: it is not a filter, it is a pipeline.

On top of that base sit the two artifacts that actually sell the look to a viewer’s eye. Halation is the soft red-orange glow that blooms around bright highlights — light passes through the emulsion, scatters off the back of the film base, and re-exposes the area around a hot window or a practical light. Grain is the fine random texture of the emulsion. Grain does double duty: it reads as filmic, and its randomness breaks up digital banding in smooth gradients like skies, hiding the stair-stepping that clean digital shows. Neither of these comes from a print LUT — you add them yourself, on their own nodes, which is exactly why LUT-only “film looks” feel flat.

Now the LUT itself. A print-film-emulation (PFE) LUT — 2383, or the Fuji 3513/3510 — is the print-stock step, packaged so you can apply it. The word “apply” hides three disciplines. First, normalize and expose correctly first: get the shot from log into your working space with a color space transform (the correct move — not shoving saturation to 100), and balance it before the look, so the LUT receives a neutral, properly exposed image rather than fixing exposure for you. Second, place it correctly in the tree — the look belongs late, after your correction and balance and before the final output, so you are grading in a scene-like state, not on top of the LUT’s baked-in curve. Third, add halation and grain separately rather than expecting the LUT to do everything.

The reason to understand all this rather than collect LUTs is what Kelly calls the Look User trap. In the 5-levels talk he auditions the 2383 against a Fuji print emulation and gets a genuine “wow” — then travels to another shot and the same LUT falls apart. His two rules answer it. Be choosy about sourcing: treat your LUT folder like your fridge, and don’t put junk in it, because if it’s in there you’ll use it. And accept that a sealed LUT gives you nowhere to go — when a client asks for the shadows a touch greener, you can’t reach inside it. That is precisely the limitation Melara’s rebuild removes: with the 2383 broken into nodes, the overly contrasty curve everyone complains about is just a node, and you can swap it for a gentler log-to-Rec.709 transform, or dial its strength down, without throwing away the color. A LUT applied with understanding is a reliable creative look; a LUT applied blind is a coin flip.

  1. Take a normalized, exposed log shot (use a color space transform to get from your camera’s log into your working space — don’t just crank saturation).
  2. On a node before the look, balance the shot: warm or cool it with offset, nudge saturation, settle the highlights. The look should receive a neutral image.
  3. On a new node, apply a PFE LUT — Resolve ships a Kodak 2383 in the LUT browser. Place this node after your normalize and balance, before any output transform.
  4. Audition it across shots. Drop the same look on three or four different clips and find the one where it breaks — too contrasty, a color that goes wrong. That failure is the lesson in LUT discipline: a look must serve every shot, not just the hero frame.
  5. If it’s too contrasty, tame it — because it’s on its own node you can lower that node’s key output / gain, or feed it a gentler contrast curve, instead of fighting it downstream.
  6. Add a halation node: qualify the brightest highlights, blur that selection, tint it red-orange, and add it back softly.
  7. Add grain as the last step, on top of the look (Resolve’s Film Grain effect if you have it, or a grain overlay clip), at a strength that reads as texture, not noise.
  8. Toggle the whole chain off and on. Note how much of the “film” feeling came from halation and grain, not the LUT.
Level 2 workbook — every Do it exercise, 2.1–2.12, plus the capstone (printable)level-2-workbook.pdf799 KBOriginal course material — free to use

Check yourself

  1. Why is real film described as a two-stage system?

  2. What causes halation?

  3. What does a print-film-emulation (PFE) LUT like the Kodak 2383 actually model?

  4. What is the right way to apply a PFE LUT?

You can move on when you can… explain a film look as a chain — negative-to-print base, then halation, then grain — and apply a print-film-emulation LUT the right way: normalized and balanced first, placed correctly in the tree, auditioned across shots, with halation and grain added separately.

Cullen Kelly — Level 2, “Look User” (auditioning print LUTs): the segment where he throws the Kodak 2383 and a Fuji print emulation onto a shot, hits a genuine “wow,” then watches a poorly built LUT collapse on the next frame. His fridge analogy for sourcing LUTs, and the point that a sealed LUT leaves you no way to make a refinement, are the discipline behind Steps 3–5 of your exercise.

Segment: 23:46–31:56 — Level 2 'Look User': auditioning 2383/Fuji PowerGrade LUTswatch full video

Watch for: The 2383-versus-Fuji audition and the 'wow moment,' the poorly built LUT that works on one shot and breaks on the next, and his prescription: be choosy about where LUTs come from, and recognize when you're compensating in your grade for a look you can't edit.

Next up: 2.11 · Repair & refinement — cleaning up the shots a look exposes, from noise to blemishes to distractions.