You shot a wedding last weekend. Ceremony, portraits, reception, speeches, first dance, exits. 3,200 images on the card. You open them in Lightroom or your culling tool of choice and see... a flat timeline. Image 1 of 3,200. Good luck.

You start scrolling. Flag, skip, skip, flag, skip. Thirty minutes in, you're 400 images deep and you've lost track of whether that great candid of the flower girl was better than the one twelve frames back. You scrub backwards, find it, compare, decide, scrub forward again to where you were. Another thirty minutes. You're at image 900.

This is how most photographers cull. It works, technically. But it's exhausting, slow, and makes it genuinely hard to pick the best frame from a burst when the burst is scattered across a timeline mixed with completely different moments.

Scene-based culling is a fundamentally different approach. Instead of one long scroll, your shoot is organised into groups of related images — scenes — and you work through each scene independently. It sounds simple. It changes everything.

What is a scene?

A scene is a group of images that belong together contextually. During a portrait session, one scene might be the couple standing under an oak tree — maybe 15 frames from that setup. Another scene is the close-up detail shots of the rings. Another is the candid walking sequence.

Scenes aren't albums or collections. They're smaller and more specific: a single burst, a single setup, a single moment. Think of them as the natural clusters your shoot already contains. You just haven't had a tool that surfaces them before.

In Selekt, scenes can be created manually (drag a group of similar frames together) or detected automatically using AI that analyses visual similarity, timing gaps, and compositional changes. Either way, the result is the same: your 3,200-image wedding becomes maybe 80-120 scenes, each containing 10-40 related frames.

Why flat timelines fail

The core problem with a flat timeline is context switching. Your brain has to constantly recalibrate as you move from ceremony wide shots to detail close-ups to reception candids. Each transition costs mental energy. Over thousands of images, that adds up to decision fatigue that makes your picks worse, not just slower.

There's a second problem: comparison difficulty. When you're looking at frame 847 of a burst and you want to compare it to frame 839, you have to remember where 839 was, scroll back, hold both images in your head, and decide. Some tools offer a compare mode, but you still have to manually identify which frames to compare — and in a flat timeline of similar-looking images, that's harder than it sounds.

The third problem is progress tracking. In a flat timeline, "I'm 40% done" tells you almost nothing useful. Are you 40% through the ceremony and haven't touched anything else? Are you scattered across the whole shoot? There's no structure to anchor your progress against.

How scene-based culling works

The workflow is straightforward. You open your shoot and see it broken into scenes rather than a single timeline. Each scene shows a thumbnail grid of its contents. You can see at a glance: here are the 22 frames from the first look, here are the 8 detail shots, here are the 35 frames from the speeches.

You pick a scene and dive in. Now you're comparing 22 similar images, not scanning 3,200 mixed ones. The frames are contextually identical — same location, same lighting, same moment. You're just looking for the best expression, the sharpest focus, the ideal timing. This is a comparison task, not a search task. Your brain is much better at comparisons.

You pick your favourites from that scene — maybe 3-5 keepers from 22 frames — and move to the next scene. Each scene takes 30 seconds to 2 minutes depending on size. When you've worked through all scenes, you're done. The whole shoot might take 45 minutes instead of two hours.

The speed difference isn't just about fewer clicks. It's about cognitive load. When every image on screen is from the same moment, you're not context-switching. You're making one type of decision repeatedly within a known context. That's dramatically less tiring than the flat-timeline approach.

The comparison advantage

This is where scene-based culling really shines. Inside a scene, you can pull up two or three frames side by side and directly compare them. Same composition, same lighting — you're looking purely at the differences. Is the smile better in frame A or frame B? Is the focus sharper here or there? Is this the peak moment or was the next frame better?

In a flat timeline, identifying which frames to compare is half the battle. In a scene, the candidates are already grouped. You just pick the top contenders and put them head to head.

For portrait photographers, this is transformative. A 20-minute portrait session might produce 200 frames across 8-10 setups. In a flat timeline, those setups blur together. In scenes, each setup is its own unit. You quickly identify the hero shot from each setup, and you're done. Ten decisions instead of two hundred.

When scene-based culling matters most

Not every shoot benefits equally. Scene-based culling is most powerful when your shoot has natural structure — which, honestly, most professional shoots do.

Weddings and events. The classic case. Dozens of distinct moments across hours of shooting. Ceremony, preparations, portraits, reception, speeches, dancing. Each is its own world. Culling them as separate scenes matches how you actually experienced the day.

Portrait sessions. Multiple setups, multiple outfits, multiple locations. Each setup is a natural scene. Compare within the setup, pick the winners, move on.

Product and food photography. Multiple angles, lighting setups, and arrangements of the same subject. Each setup is a scene. You're comparing nearly identical frames for subtle differences in angle or lighting.

Sports and action. Bursts of 20-30 frames from a single play or moment. Each burst is a scene. Pick the peak action frame from each burst.

Travel and street photography is the exception — shoots without natural structure, where every frame is a different subject and moment. Here, a timeline view might be just as effective. But even travel shoots often have clusters: the morning at the market, the afternoon at the temple, the evening street scenes.

Manual vs automatic scene detection

There are two approaches to creating scenes, and the best workflow uses both.

Manual grouping gives you full control. You select a range of images and group them into a scene. This is ideal when you know your shoot structure — you were there, you remember the setups. It takes a few minutes upfront but gives you exactly the groupings you want.

Automatic detection uses AI to identify scene boundaries based on visual similarity, time gaps between frames, and compositional changes. It's not perfect — sometimes it splits a scene that should be one, or groups two that should be separate — but it gets you 80-90% of the way there in seconds. You can then adjust: merge scenes that were split, or break apart ones that were incorrectly grouped.

The practical workflow: let auto-detection create the initial groupings, then spend 60 seconds adjusting before you start culling. The time investment is minimal and the payoff in culling speed is significant.

What about AI culling?

AI culling tools like Aftershoot take a different approach: instead of organising your images into scenes for human review, they try to make the pick/reject decisions for you. The AI analyses technical quality (sharpness, exposure, eyes open) and learns your preferences over time.

These approaches aren't mutually exclusive, but they solve different problems. AI culling optimises for speed by removing the human from parts of the decision. Scene-based culling optimises for decision quality by giving the human better context.

For photographers who want full creative control over their picks — where the choice between two nearly identical frames comes down to a subtle expression or a feeling — scene-based culling preserves that control while making it dramatically faster. You're still choosing. You're just choosing within context instead of choosing from chaos.

For photographers who are comfortable delegating the initial pass to AI and reviewing its suggestions, AI culling can be faster. The trade-off is trusting an algorithm with subjective creative decisions.

Many photographers will eventually want both: AI to handle the obvious rejects (blurry, eyes closed, test frames), and scene-based organisation for the creative picks from what remains.

Getting started with scene-based culling

If you want to try scene-based culling, here's a practical approach.

Grab a recent shoot with at least 500 images — ideally a wedding, event, or portrait session with natural structure. Open it in a tool that supports scene grouping (Selekt has this built in with both manual and AI-powered scene detection).

Let the tool create initial scene groupings. Spend a minute scanning the scenes — are the groupings roughly right? Merge or split any that need adjustment.

Then start culling scene by scene. Work through each group, compare the candidates, pick your favourites. Notice how different it feels compared to a flat scroll.

Most photographers report two things after their first scene-based culling session: it was faster than expected, and their picks felt more confident. When you're comparing 15 frames from the same moment instead of scanning 3,000 frames from an entire day, the right choice is usually obvious.

That's the whole point. Not a radical new workflow. Just the obvious way to review photos, once you see it.

Scene-based culling explained: why context changes everything