Orelon logoOrelon
Preise

AI Shot Lists for Wedding Cinematography: A Workflow Guide

14. Nov. 2026 · Von Orelon Team

KI-Video-Vorlagen entdecken

Lass dich von ein paar Community-Kreationen inspirieren und öffne dann eine Vorlage, um in Orelon weiterzuerschaffen.

Build repeatable AI shot lists for wedding films, unify the look with style transfer, and streamline post-production without losing directorial control.

Most wedding films are not ruined in the edit. They are lost on the day — in the ten seconds a cinematographer spends deciding whether to chase a reaction or hold the wide, in the coverage nobody wrote down, in the light that shifted while the crew reset a tripod. Shot lists exist to remove that hesitation. What has changed is that a shot list no longer has to be a static document written from memory and guesswork.

AI-assisted planning lets you generate coverage for a venue you have never walked, preview a look before you rent the lens, and hand the edit a structured first assembly instead of a folder of loose clips. The catch is that none of it works without a benchmark — a written, testable definition of what "cinematic" means for this couple, this venue, and this delivery. What follows is a practical system for building that benchmark, translating it into shot lists, applying a consistent look, and automating the parts of post that genuinely deserve automating.

Start With the Story Decision, Not the Tool

Every AI planning tool will happily give you 120 shots. That is the problem. A wedding film needs a spine: three or four story beats that all coverage must serve — anticipation, the exchange, the release, the last light. Before generating anything, write one sentence per beat in plain language. "She is calm until her father sees her." "He forgets the ring and laughs." Those sentences become filters; any generated shot that does not serve one of them is noise.

This is the difference between a shot list and a wish list. A wish list is organized by gear and time of day — drone, gimbal, detail, wide. A shot list is organized by what the edit needs to prove. You get far better output from an AI planner when you feed it beats instead of equipment. Try a brief like: "Two-person documentary crew, outdoor ceremony at five in the afternoon, strong backlight, four beats described below; propose sixty seconds of coverage per beat plus two alternates each, and flag any shot that requires a second camera position."

That single change — beats first, gear second — converts a generic list into something a crew can actually execute under pressure.

What a Production-Ready AI Shot List Actually Contains

A list that only names shots is a suggestion. A list that specifies how each shot is made, why it exists, and what could go wrong is a working document.

Coverage Blocks Instead of Shot Islands

Group shots into blocks of eight to twelve that can be captured without moving the camera more than a few meters. Each block names a location, a lens range, the story beat it serves, and an estimated capture time. Blocks make the day legible: you can see at a glance that the vows block needs eleven minutes and the portraits block needs thirty-five, and you can cut a block without breaking the story if the schedule collapses.

Shot Descriptions Written Like Prompts

"Nice close-up of hands" is useless to a human and worse to a model. Write descriptions as specifications: subject, framing, movement, light direction, duration. "Bride's hands, 85mm equivalent, static, side light from window left, four seconds, ring visible." This phrasing doubles as a previsualization prompt, which means the same document that briefs your operator can also generate a reference frame for the couple to approve.

A Constraints Column Nobody Enjoys Writing

Every venue makes something impossible. A dark stone church with no window light kills the natural close-up during the vows. A rooftop with a west-facing terrace makes the first dance a silhouette unless you plan a bounce. Add a short constraints column and have the AI planner suggest workarounds: reflectors hidden behind the altar flowers, a long lens from the balcony, a deliberate silhouette treated as a stylistic choice rather than an accident. This is where experience usually lives, and it is exactly the knowledge worth encoding into a reusable template.

The Benchmark: Six Metrics That Define Cinematic

"Cinematic" is not a mood, it is a set of measurable choices. Write these six numbers into every project brief and the output stops being a matter of taste.

  1. Coverage ratio. Aim for roughly four times the footage you intend to use. Under-covered weddings force long takes; over-covered weddings drown the editor.
  2. Shot duration curve. Establish an average cut length for each act — longer during vows, shorter during the party — and a maximum for any single shot before it becomes a music video cliché.
  3. Movement budget. Two or three moving shots per minute, no more. Constant gimbal motion reads as stock footage.
  4. Light continuity. Every beat should have at least one frame with a defined light direction, so the grade has something to anchor on.
  5. Color anchor. Pick two reference frames, one skin tone and one environment, and treat them as the truth for the whole grade.
  6. Sound-to-picture balance. Decide where the natural audio carries the scene and where music takes over. Shot lists that ignore audio produce edits that cannot breathe.

Once these are written down, they become instructions. You can literally paste them into a planning prompt, and quality becomes reproducible across crews, venues, and budgets.

Style Transfer Without Losing the Day

Style transfer is the most misunderstood part of AI-assisted cinematography. Used well, it unifies footage from three cameras, two operators, and a phone; used badly, it turns a real wedding into a filter.

Build a Look Bible First

Before generating anything, collect eight to twelve stills that represent the intended look — film stocks, reference films, previous work you are proud of. Reduce them to written rules: contrast curve, shadow tint, highlight roll-off, grain amount, skin saturation. The written rules matter more than the images, because images tempt a transfer model into literal imitation while rules give it direction. A look bible of half a page will do more for consistency than an hour of slider tweaking.

Protect Skin Tones and Candlelight

Two things break first under aggressive transfer: skin and practical light. Apply a skin-tone protection pass, grade faces naturally, then push the environment. Candlelit receptions are the classic failure case, where a warm, cozy scene becomes orange mud. If the transfer flattens candle glow into a single hue, pull it back — the point is atmosphere, not uniformity.

Know When to Refuse the Transfer

Some footage should stay untouched: a grandmother's reaction, a nervous handheld moment, a shot that only works because it is imperfect. Mark these in the timeline as protected. A useful rule is that no more than roughly seventy percent of any scene should receive automated treatment until you have reviewed it at full size on a calibrated display.

Previsualizing a Venue You Have Never Filmed

Destination work is where AI planning pays for itself fastest. You can arrive with a coverage plan built from photographs instead of improvising at the walkthrough.

Start by gathering twenty to thirty stills of the venue from the couple's photographer, Google Maps, and the venue's own social feed. Generate reference frames for each coverage block at the correct time of day, and note where light comes from in each. Generate the same frame at two alternate hours so you know what the backup plan looks like. Then build a short previsualization sequence — thirty to sixty seconds of generated frames cut together — that you can send to the couple for approval.

Two practical cautions. First, generated frames are a planning tool, not a promise; be explicit with clients that these are reference images, not a preview of the final grade. Second, verify the sun path and the room geometry yourself. Models will happily produce a gorgeous window that does not exist, and a plan built on fiction is worse than no plan. Tools that generate quick cinematic frames from text are useful here — the Orelon AI image generator is a straightforward way to produce those reference stills — but the verification step stays human.

From Capture Notes to a First Assembly

Automation earns its place in post when it removes sorting, not judgement. The useful pattern is to let the machine build a rough structure that an editor then destroys intelligently.

Scene Assembly Logic

Feed the edit three things: the story beats, the shot list with its block structure, and the audio transcript. The assembly pass then groups clips by beat rather than by timestamp, marks alternates next to their primaries, and proposes a sequence per beat. Editors typically keep a third of it and rewrite the rest — but starting from a structured timeline rather than a bin of 900 clips saves hours.

The Grading Pass

Apply the look bible as a base grade, then treat it as a first note rather than a final answer. Automated grading is strongest on continuity — matching three cameras across a changing room and a garden — and weakest on faces. Review every shot with a person in it before exporting.

Consistency Checks Across Shots

Use keyframe comparison to flag drift: dresses that change shade mid-sequence, skies that jump between takes, mixed frame rates from a second operator. These checks are boring and therefore perfect for automation. Build the flag list into your delivery checklist and you will catch problems before a client does. If you are still choosing tooling, the Orelon AI video generator is built for turning these planned sequences into motion tests you can review before the shoot.

A Realistic Three-Phase Workflow

Phase one, two weeks out. Collect venue references, write the four story beats, generate the shot list, cut it down to what fits the schedule, and export a one-page brief with the six benchmark metrics at the top. Send the previsualization frames to the couple.

Phase two, on the day. Work in blocks, not shots. Film the block, check it off, move on. Note deviations as you go — a missing shot is fine if you write down why. Keep one operator free for reactions during every transition.

Phase three, seventy-two hours after. Ingest, run the assembly pass, apply the base grade, then edit beats in order. Do a continuity check, a skin-tone check, and an audio check before the first client review. Deliver a short teaser first; it protects the longer edit's timeline and gives you feedback while the day is still fresh.

Mistakes That Quietly Ruin AI-Assisted Wedding Films

  • Generating shots before deciding the story. Volume without structure is just noise.
  • Trusting generated light. Models invent windows; verify the room.
  • Over-applying the look. A grade that swallows candlelight or skin texture reads as amateur.
  • Automating the edit before sorting the audio. Structure follows speech, not the other way around.
  • Letting the list outrank the moment. If something real happens, shoot it and discard the plan.
  • Skipping the protected-shot list. Some frames must survive the pipeline untouched.
  • Never updating the template. A shot list that does not improve after each wedding is a wasted asset.

FAQ

Do AI shot lists replace a cinematographer's judgement?

No. They replace the blank-page problem. A good list gives you a defensible starting structure, and your judgement decides what to keep, what to cut, and what to abandon when the day goes sideways.

How many shots should a wedding shot list contain?

For a documentary-style feature, sixty to ninety listed shots in eight to twelve blocks is workable for a two-person crew. Anything above that tends to collapse by the ceremony. Build alternates into blocks rather than adding new blocks.

Can style transfer match footage from a phone and a cinema camera?

Yes, within limits. Transfer is good at matching contrast, tint, and grain; it struggles when dynamic range and motion blur differ significantly. Shoot with matched frame rates and shutter angles, and protect skin tones in the grade.

Is automated post-production safe for client work?

It is safe for sorting, continuity flagging, and base grading. It is not safe as a final pass without human review. Treat every automated output as a first assembly, not a delivery.

What should I send the couple before the wedding day?

A one-page brief: the story beats, the schedule in blocks, three to five reference frames, and a short note on the intended look. It sets expectations, prevents look disagreements later, and takes under an hour to produce.

Build Your Own Benchmark in Orelon

The gap between a good wedding film and a forgettable one is rarely equipment. It is a written plan, a consistent look, and a post pipeline that respects both. Start with four beats, six metrics, and a look bible of half a page. Then let automation handle the sorting, the continuity checks, and the base grade while you keep your hands on the story.

Orelon is an AI video generator built for cinematic ideas in motion — use it to previsualize venue coverage, test a look before you commit, and turn planned sequences into motion before the shoot. Explore the prompt library for shot-level descriptions you can adapt, browse video templates for repeatable project structures, and build your first benchmark sequence today.