Cull in or cull out: which is faster

Every culling workflow answers one question before anything else: do you mark what stays, or mark what goes? Most of us picked an answer years ago and never revisited it. That is a mistake, because the faster method flips depending on the size and the hit rate of the job, and using the wrong one is how a shoot turns into a giant, undecided maybe pile.

The two systems, defined properly

Cull in means every frame starts neutral. You flag or rate the ones you want to carry forward, then filter down to those. In Lightroom Classic that is P for pick. In Photo Mechanic it is a color class or a rating. Unmarked files stay in the folder and the catalog, they simply do not enter the next pass.

Cull out means every frame starts as a presumed keeper. You mark the blinks, the focus misses, the accidental frames and the repeats with X or a reject label, and whatever is still unmarked at the end is your working set.

On paper these are mirror images of each other. In practice they are not, because they ask you a different question. Cull in asks whether a frame has earned a place. Cull out asks whether you can prove a frame should go. The first sets a positive bar that a mediocre frame has to clear. The second gives every technically acceptable near-duplicate the benefit of the doubt, and near-duplicates are exactly what a busy shoot produces most of.

Worth noting that the machines have picked a side. Adobe's Assisted Culling in Lightroom, labeled early access, scores subject focus, eye focus and eyes open, and flags exposure problems and misfires as rejects. That is automated cull out, and Adobe states it is optimized for portraits and headshots, which is a clue about where negative selection works best.

One rule decides it: mark the smaller pile

Strip away the philosophy and the speed question is arithmetic. Every mark costs you a keystroke and, more importantly, a decision you have to be sure enough about to commit. So the faster method is whichever one requires fewer marks, and that is decided by your keep rate, not by your temperament.

If a third of the folder survives, marking keepers means committing on a third of the frames and skimming past the rest. If nine tenths survives, marking keepers means committing on nine tenths, and marking failures means touching a tenth. Same shoot, same you, wildly different amounts of work. The crossover sits at roughly half: below it, cull in wins, above it, cull out wins.

Now put real volumes against that. Commonly cited industry ranges put a full wedding day at 2,000 to 4,000 RAW frames with 400 to 800 images delivered. Even allowing for a generous first pass that carries several times the final gallery forward, a wedding sits nowhere near half. It sits far below it, which is why cull in is the right default on almost every event job. Where your own keep rate actually lands, and why it is worth knowing, is covered in the photo keep rate explained.

The useful consequence is that this is not a personality test. Measure the keep rate on the kind of work you do most, and the method chooses itself.

Why volume pushes you toward cull in

The reason cull out collapses on a big shoot is not the number of failures. It is the number of frames that are neither good nor bad. A 3,500-frame wedding does not contain 2,800 obvious failures. It contains a few hundred clear misses, a few hundred strong frames, and a vast middle of perfectly exposed, perfectly sharp, slightly worse versions of a frame you already have.

Cull out forces you to make a reject decision on that entire middle, one frame at a time, with no context to decide against. Cull in never asks. The middle is simply not flagged, which costs nothing and decides nothing, and the frames stay exactly where they are in case you want them later.

This is also the stage where the clock actually goes. Comparison work inside bursts is what makes a large shoot cost more than its frame count suggests, an effect worked through in culling time by shoot size. Cull in shrinks the set you have to compare before you start comparing. Cull out hands the whole middle to the comparison stage, which is the most expensive part of the evening.

There is a story benefit too. Flagging a reaction, a scene setter, a clean portrait, a key action or a useful detail means you are building a gallery from the first pass rather than tidying a hard drive and hoping a gallery falls out of it.

Where cull out genuinely wins

Cull out is not the amateur option. It is the correct option whenever the hit rate is genuinely high, and there is plenty of work like that.

Tethered studio headshots, where exposure and pose were reviewed on the screen as you shot. Product photography with deliberate variations, where most frames are meant to be usable. Real estate brackets, where the decision is which sequence to blend rather than which frame is best. A short session you already reviewed with the client standing next to you. In all of those, most of the folder is potentially deliverable and the failures are few, obvious and quick to mark. A headshot day, where the whole point is one strong frame per person out of a small controlled set, is a good example, and that workflow is in a headshot day for 200 employees.

Cull out also earns its place as a second pass inside a cull-in workflow. Once you have a candidate set of a couple of hundred frames, marking the technical misses within that set is fast and precise. The difference is scale: you are rejecting a few dozen frames out of 200, not trying to pass judgment on 2,000.

The warning sign is when good enough is common. A first kiss burst can hold a dozen acceptable frames and one exceptional one. Mark only the bad ones and you have deferred eleven decisions rather than made any.

The two methods side by side

Neither is a permanent identity. The most common working setup uses both, in a fixed order, and the only question that matters at each stage is which decision you can make confidently right now.

The hybrid most event photographers end up with

Positive selection first, negative selection later, once the set is small enough for negative selection to be cheap. For a 3,000-frame wedding or event, the sequence looks like this.

1. Fast cull in. Full screen, one flag, one question: does this frame have a readable expression, a key action, a clean portrait, useful context or a real reaction? Nothing else gets asked.
2. Mark the hard failures, inside the flagged set only. Focus misses, blinks, a blocked face, an obvious accident.
3. Compare by sequence. One hero per burst, and an alternate only when it does a different job: a different expression, a wider crop, another person's reaction. Two frames of the same smile is not a different job. The judgment behind that is in choosing the best frame from a burst.
4. Check coverage by chapter. Getting ready, ceremony, portraits, reception. Missing people, missing moments, missing details, missing transitions.
5. A final negative pass. Frames that repeat something the gallery already says, marked down out of the delivery set.

The design keeps the first decision generous enough to protect a quiet moment and the last decision strict enough to keep the gallery tight. It works in Lightroom Classic with flags, rejects and collections, and in Photo Mechanic with color classes and a filtered view.

One rule holds it together: one criterion per pass. If you are asking whether a frame is your favorite, whether it needs a crop and whether the color works, all at the same time, the pass will take three times as long and the standard will drift. Set a speed limit as well. Longer than about two seconds on a frame in the first pass and you flag it rather than deciding it, because you will have far better information once you can see the rest of the sequence.

Which method, by genre

Weddings and documentary events. Cull in first, always. Too many technically fine repeats, and story coverage matters more than any single frame. Cull out later, inside the candidate set.

Family and newborn sessions. Cull in for expression and connection, then cull out within each family grouping for blinks and awkward hands. The gallery needs the safe group shot and the personality frames both.

Corporate events. Cull in against the required list first: the speakers, the sponsors, the people who must appear. Then cull out repeated podium and handshake frames once the checklist is satisfied.

Studio headshots and product. Cull out is often faster because the hit rate is high, but still flag one clear best per subject or per setup before a proof gallery grows past what a client can decide on.

Sport, pets and concerts. Cull in, burst by burst. The technical hit rate can be low and the peak moment matters more than the number of acceptably sharp frames around it.

The pattern underneath all of these: the more the shoot was controlled, the better cull out works. The more the shoot happened to you, the more you want cull in.

The mistakes that make either method slow

Mixing marks without a rule. The worst outcome is a folder where an unmarked frame means three different things: maybe, not reviewed yet, and not good enough. Write the meanings down and keep them fixed. Flag is a candidate, reject is a definite no, no mark is not carried forward. If you would rather use stars or colors, the trade-offs are laid out in ratings, flags and labels compared.

Deleting rejects on the spot. A reject label is a filter, not a bin. Lightroom hides rejected frames without touching the files, and permanent deletion is a separate deliberate step for a reason. Keep everything until the gallery is delivered and backed up, because a client asking for one more frame from the ceremony is a normal Tuesday. The longer view on that is in should you keep rejected RAW files.

Treating a big flag count as failure. The first pass is supposed to be inclusive. Selectivity is the job of the comparison stage, when you can finally see frames next to each other. Flagging generously and then cutting hard is faster than agonizing early, and it is easier on your confidence, which is its own subject in rejecting photos without second-guessing.

This is also the part of the workflow Kepla is being built around. The picking app for iPhone, iPad and Mac is still in development, and the idea is a first pass that hands you a shortlist to approve rather than a card dump to judge, with every frame it did not pick still sitting untouched, one tap away. It never deletes, moves or renames a file. The booking page is live and free today while the rest gets built.

FAQ