How to Cull 2,000 Wedding Photos in Under an Hour

Why a 2,000-frame cull takes three hours

Take a typical wedding day: two shooters, a long ceremony, a reception with a dance floor. Two thousand frames is ordinary, and if you spend 5 seconds on each, that is 2,000 × 5 s = 10,000 s, or about 2 hours 47 minutes of staring at your own work.

The five seconds is not spent well. Most of it goes on frames that were never going to be kept: the blink, the half-turned head, the third identical frame of a first kiss. A decision to reject takes a fraction of a second when you only have to answer one question. It takes much longer when you are also asking whether this is the best of the set, whether it needs a crop, and whether the client will like it.

Splitting those questions across two passes is the whole technique. Each pass asks exactly one thing, so each decision is fast.

The two-pass budget

Here is the budget. The per-frame seconds are assumptions you should replace with your own after the first wedding; the totals are straight multiplication.

The 800 in the second row assumes pass one removes 60 percent of the frames (1,200 of 2,000). That is a plausible figure for a day shot with bursts, but it depends entirely on how you shoot. If you shoot sparsely and reject only 30 percent, pass two starts with 1,400 frames and takes 1,400 × 1.5 s = 2,100 s, or 35 minutes, which pushes the total to about 68 minutes. The method still works; the hour does not.

That honesty matters, because the point of a budget is that you can see which number is carrying the weight. For a burst-heavy shooter it is the reject rate in pass one.

Pass one: reject without looking twice

Work in a full-screen grid or loupe view and press a single key. You are only hunting for frames that fail for a reason you can name in a word:

• Focus: the eyes are soft and nothing else is sharp.
• Eyes: a mid-blink, or a face caught between expressions.
• Exposure: blown skin on the veil or a silhouette that is not a deliberate one.
• Obstruction: a guest's phone, an arm, a pillar through the aisle shot.
• Duplicates you are certain about: the frames either side of a clearly better one.

Do not compare in this pass. If you are not sure, leave the frame alone; pass two exists to compare. A one-second budget means roughly one glance, and hesitating is the signal to skip ahead.

One practical trick is to zoom to 100 percent only on a frame that looks sharp but suspicious, and nowhere else. Checking focus on every frame at full magnification is how a pass that should take half an hour takes two.

Pass two: pick one winner per moment

Now the survivors are grouped by moment: the first look, the ring exchange, the walk back down the aisle. Show yourself a burst or a near-duplicate group side by side and choose the single best frame. Compare expression first, composition second, light third. Most groups settle in under two seconds, which is why 1.5 is a fair planning number.

Keep a second frame only when it shows something the first does not: a wider view that sets the scene, or a reaction shot from the other side of the aisle. If two frames say the same thing, one of them goes.

If you already work with a fixed rating system, use it here rather than inventing a second one. Our comparison of star ratings, colour labels and flags covers why one scheme beats three, and the short version applies directly: use flags for reject versus keep, and reserve stars for the finer pick in pass two.

Set up the session so it stays fast

The passes only take the time above if nothing slows the screen. Three things make the biggest difference.

• Previews first. Build standard-size previews on import so you are never waiting on RAW decoding while you flick through the grid. Culling from full RAW rendering is a common cause of "culling is slow".
• One scorer per pass. If a second shooter's frames are mixed in, cull your own set first and theirs second, so you are not switching style or camera in the middle of a pass.
• A keyboard, not a mouse. At one second a frame there is no time to reach for a toolbar. Learn the reject, keep and advance keys, and keep a hand on them.

Sort by capture time, not filename. Moments cluster in time, and pass two depends on seeing a burst together. If two shooters were on different clocks, sync them first; a ten-minute offset scatters one moment across the timeline.

What to do with the 300 you kept

At the end of pass two you have a working set, not a delivery. The next step is to put it in front of the couple and let them make their own selects, especially if an album or print order is part of the package. Your approval workflow for weddings should say who sees what and by when, because a cull that takes under an hour is wasted if the gallery then waits two weeks for an answer.

Send the kept frames as one review link where the couple can mark favourites and comment per photo, so their selects come back attached to the images rather than as a list of filenames in an email. That is the part of the job a client-review tool such as Appruva handles; the cull itself stays in your own software.

Keep the rejects. Do not delete them until the gallery is approved and delivered, in case a couple asks for the frame you thought was too soft, and so that your backup holds the whole day. See the backup strategy post for how to make sure it does.

A worked example: one ceremony, start to finish

Say the ceremony produced 420 of your 2,000 frames, because you shot the processional, vows, ring exchange and recessional in bursts. Apply the same budget to just that slice: pass one at 1 second a frame is 420 s, or 7 minutes. If 60 percent go, 168 remain, and pass two at 1.5 seconds is 252 s, a little over 4 minutes. The whole ceremony is done in about 11 minutes, and you have roughly 168 candidates to pull a final 40 to 60 from.

The useful part of doing it per segment is that you can check your pace against a small number. If the ceremony took 25 minutes instead of 11, you know early that the hour is at risk and which habit to change, before the reception's 900 frames make the problem expensive.

Work the segments in the order they happened. Context carries over: you remember which expression the groom had in the last frame, so you notice faster when the next one is worse.

Mistakes that quietly double the time

Four habits account for most overruns, and each is easy to catch once named.

• Re-judging in pass two. If you reopen a frame you already kept in pass one and ask whether it is good enough, you have merged the passes again. Pass one only removes; it never ranks.
• Editing while culling. Cropping, straightening or tweaking exposure on a frame you may not keep is time spent on a rejected image. Edit only after the count is final.
• Culling tired. Standards drift late in a long session. If the reception is the last 900 frames, take a break before it, or cull it first thing the next morning with fresh eyes.
• Keeping everything "just in case". A 900-frame delivery from a 2,000-frame day does not give the couple more; it gives them a harder choice and delays the sign-off.

The last point connects to a bigger question: how much of your cull should the client see at all? That is a decision with real trade-offs, covered in whether clients should see every frame or only your selects. A tight cull makes that question much easier to answer.