Tutorials Video

Culling and Selecting: From 800 Frames to 20

Beginner · ~16 min

Overview

The part of photography that actually consumes time is not shooting. It is deciding which frames to keep. Without a process, people open a folder of eight hundred images, spend two hours on the first fifty, and abandon the session. This guide covers a multi-pass culling method that gets through a large shoot quickly by making each pass answer only one question, and the selection criteria that make the decisions fast rather than agonising.

What You Need

Software with a rating or flagging system: Lightroom, Capture One, darktable, or Photo Mechanic
Raw files already copied to two locations before you start
A consistent screen and a room whose lighting does not change
A defined target number of final images
An hour of uninterrupted time rather than several fragmented ones
The discipline not to start editing during the first pass

Steps

Back up before you delete anything

Copy the full card to two locations and verify both before culling begins. Culling is destructive by intention, and doing it against your only copy means a mistaken keystroke is permanent. This step is boring and it is the reason people who skip it eventually lose a shoot.

First pass: reject only the technically unusable

Go through fast, a second or two per frame, and flag only the ones that are definitively out: badly out of focus, subject blinking, missed the moment, blank frames, accidental shutter presses. Do not compare, do not deliberate, do not think about which of two similar frames is better. This pass typically removes a third of a shoot in a few minutes.

Second pass: pick the best of each cluster

Most shoots contain groups of near-identical frames. Work cluster by cluster and pick one winner from each, marking it. Comparing within a small group is fast. Comparing across the whole shoot is what causes paralysis. Zoom in only when two frames genuinely differ on sharpness.

Third pass: select against the brief

Now look at the winners as a set and choose against what the images are for. An event set needs coverage variety. A portrait session needs a range of expressions. A product shoot needs each required angle. This is where you notice a gap, and where you still have time to fill it from the rejects.

Stop at your target number

Decide the target before you start and hold to it. Delivering everything acceptable is a disservice: a client shown two hundred images will feel the work was uncurated, and a tight set of twenty strong ones reads as better photography even when the same frames are present in both.

Only now start developing

Keep culling and editing strictly separate. Editing during selection is how an hour becomes an afternoon, and it biases the choice. A frame you have spent ten minutes on feels more valuable than one you have not, regardless of which is actually better.

Pro Tips

Use one rating pass per question. Multi-criteria judgement per frame is what makes culling slow.
Trust your first reaction on the initial pass. Frames you hesitate over are almost never the ones you finally deliver.
Cull on a consistent screen in consistent light. Judging brightness in a room that changes leads to inconsistent selections.
Never cull tired. Fatigue makes everything look mediocre and you will discard usable work.
Keep the rejects until delivery is accepted. The one frame you dismissed is occasionally the one the client asks for.

Knowledge Base

What You'll Learn

Culling is a decision-making problem more than a photographic one, and the method exists to reduce the number of decisions per frame. Below: why multiple fast passes beat one careful one, and how to choose between technically similar images.

Why Several Fast Passes Beat One Careful One

The instinct is to look at each frame once and decide its fate completely. This is slower and produces worse results, for a reason worth understanding.

Deciding "is this a keeper" requires holding several criteria at once: technical quality, expression, composition, whether a better version exists elsewhere in the shoot, and whether it serves the brief. That is a large cognitive load per frame, and it degrades quickly. By frame two hundred, judgement is measurably worse.

Splitting into passes means each frame is judged against one simple question at a time. "Is this technically unusable" is nearly instant and requires no comparison. "Which of these five similar frames is best" is a small local comparison. "Does this set cover the brief" is a single decision about the whole group rather than hundreds of individual ones.

The result is both faster and more consistent, because no single decision is carrying the full weight.

Choosing Between Technically Similar Frames

The hard cases are the ones where two frames are equally sharp and well exposed. The tiebreakers in rough order of importance:

Eyes and expression. In any photograph containing a person, this decides it. A slightly softer frame with a genuine expression beats a critically sharp one with a held or awkward face, every time.

What the hands are doing. Hands caught mid-gesture in an odd position ruin otherwise good frames and are easy to overlook while focused on the face.

The background. A pole behind a head, a stranger looking at camera, a bright distraction at the edge. These decide between otherwise equal frames.

Whether it adds to the set. Between two good but near-identical images, keep the one that differs more from what you have already selected. Variety within a delivered set matters more than the marginal quality of any single frame.

Where This Fits

This guide covers one specific part of photography. The wider picture, exposure, focus, light, and composition for stills, including what does and does not transfer from shooting video, is in Still Photography Fundamentals: Composition, Exposure, and Light, which frames the discipline as a whole and links out to the detailed guides underneath it, including this one. If you are starting from scratch rather than solving a specific problem, read that first and come back here.

FAQ

Q: How many images should I deliver from a shoot?
A: Fewer than feels generous. The right number depends on the job (a portrait session might be fifteen to thirty, an event fifty to a hundred and fifty) but in all cases a tight edit reads as better work than a large one containing the same highlights. Agree the number in advance so expectations match.

Q: Should I delete the rejects?
A: Not immediately. Keep them until the client has accepted delivery, since requests for a specific moment you dismissed do happen. After that, deleting obvious technical failures is reasonable housekeeping, but keep the raw files of anything you delivered indefinitely. Your editing taste will change.

Q: How long should culling take?
A: Much less than most people spend. With a multi-pass method, several hundred frames should take well under an hour to get to a shortlist. If it is taking substantially longer, the usual cause is editing during selection or trying to judge every criterion on every frame at once.

Q: Does culling software actually matter?
A: Somewhat. Dedicated tools load previews faster, which matters a great deal at volume, waiting two seconds per frame across eight hundred frames is most of an hour of pure waiting. Any software with flags or star ratings and quick preview loading is sufficient. The method matters more than the tool.