The first thing video takes away is the shutter. Choose 25 frames a second and 1/50 arrives with it, because a shutter near half the frame interval is what makes movement look like movement instead of like a series of sharp stills, and the control photography let you spend freely is spoken for before the first shot.
Everything on this page is worked on one small film. Marta and her father sell waffles from a van at a bus terminus in Lublin, and the film is two evenings there plus one afternoon outside the depot while the van is loaded. Four minutes long, finished, and the arithmetic of how it got that short is at the bottom of the page.
What a frame rate costs in light
25 fps means a shutter of 1/50, and 50 fps means 1/100, which is the whole of the rule and also the whole of the problem. Slow motion is not free: doubling the frame rate to get it costs a stop, and there is nowhere obvious to find that stop again.
The afternoon outside the depot makes the cost visible. At two o'clock, sensitivity already at its base, the meter settles on f/22 at 1/50 and ISO 100 in bright sun. That is a working aperture for a document and a useless one for a face: everything from the wing mirror to the depot fence is equally sharp, and Marta's father is a man standing in front of a wall. Opening up is not available, because the shutter is being held at 1/50 and there is no sensitivity left to give back.
So the light gets thrown away on purpose. A six stop ND buys back f/2.8, the fence goes soft, and the shot becomes a shot of a person. This is the piece of equipment the craft genuinely requires, and it is the one nobody photographs their kit bag with.
You cannot cut what you did not cover
The stall in video is not the edit itself. It is what the edit finds: an evening of good-looking material that cannot be assembled, because every shot was taken from the same place. Two shots less than 30 degrees apart read as a jump, and the viewer sees a mistake in the file rather than a change of view.
The repair is mechanical. Change the size of the shot or move 40 degrees round, and preferably both: a wide of the van from the queue, then a close of the iron from the serving hatch. Nobody notices the second cut. Everybody notices the one where the same framing hops six degrees to the left.
This is what makes getting footage that can actually be cut together a different skill from taking good pictures, and it is the reason a first film is usually not fixable in assembling footage into something that holds attention. The material to make the cuts was never recorded, and no editor conjures it.
What survived
Where the four minutes came from
Two evenings and an afternoon, subtracted twice: once for material that could never have been used, and once for material that was good.
Minutes of recorded footage. One film, two evenings at the terminus and one afternoon at the depot.
- 96 minOn the cards
- 44 minNever a candidate
- 37 minCut in the assembly
- 11 minCut in the second pass
- 4 minDelivered
The last fall is the one worth looking at. Everything removed there was usable, and a ratio like this is a property of the shape being made rather than a measure of how badly the shoot went.
Where the cut goes, and where the sound goes
At 25 fps a frame is 40 ms, and cuts are decided at that size. Marta closes the waffle iron in eleven frames, a little under half a second. Cut on the frame her wrist starts to turn, not the frame the lid lands: the eye is already following a movement that has begun, and it will carry that motion across into the next shot without registering that anything was joined.
Waiting for the lid to land does the opposite. The action completes, the shot has nothing left to do, and the cut announces itself in the two frames of stillness before it. This is the single most common repair in a first edit and it moves nothing more than a handful of frames.
Sound is where a cut actually disappears. Bring the next shot's sound in ten frames before its picture and the ear arrives at the new place first, so the picture change is confirming something rather than announcing it. Reverse it and the cut hardens, which is a useful thing to have in the toolkit for the one moment in a film that should be abrupt.
Two evenings, and what came out of them
Two evenings at the terminus and one afternoon at the depot put 96 min on the cards. Of that, 44 min was never a candidate: the iron out of frame, a bus idling across the dialogue, and three takes where somebody looked at the lens. The assembly cut 37 min more, most of it repetition nobody would have defended out loud. The second pass took 11 min out of what remained, and it is the pass that hurts, because everything cut there is good. What went out was 4 min.
That ratio is not a sign of waste, and it is worth knowing before the first shoot rather than after it. It is also the argument for the shapes a video comes in: a four minute portrait, a thirty second thing for a phone and an hour-long interview do not have the same ratio, and choosing the shape before shooting is what decides how much of the two evenings was ever going to survive.
The order of learning
The route goes back past the shoot, and that is the lesson
The first lap teaches nothing anybody can name. The second one is walked by somebody who has sat in front of a timeline with a hole in it.
- 01Shoot somethingOne angle of everything, because it all looks fine on the day
- 02Try to assemble itWhere the missing shots become visible, and only here
- 03Find the holeUsually a wide, a hand, or anything at all from a second position
- 04Shoot for the cutThe same subject, covered on purpose this time
- 05Cut it properlyWhere the frame level decisions start being worth making
- 06Hear the soundThe last discovery, and the one that sends people back to the shoot again
Colour and delivery are not on this route. They sit after it, and putting either one inside the loop is the most reliable way to spend a month grading shots that never reach the film.
The order doubles back, and it is meant to
Shoot, try to cut, discover the shot that does not exist, shoot again knowing what for. That loop is the whole curriculum and the second lap is where the learning is: the first shoot teaches you nothing you can name, and the second one is done by somebody who has sat in front of a timeline with a hole in it.
Colour comes late, and putting it early is the classic wrong order. Making footage look intentional can carry a shot that was exposed and framed with intent, and it cannot rescue a flat one; every hour spent grading before the cut is locked is an hour spent grading shots that will not be in the film. Formats come last and get skipped, which is how a finished film ends up unwatchable on the only screen anyone will watch it on: getting it to the viewer intact is a half day, once.
What to open in 2026
DaVinci Resolve is the answer to the tool question and has been for a few years, but the reason that matters for this page is narrower than the usual one. Its audio page is in the same application as the timeline, so the ten frame sound lead above is a drag rather than a round trip, and it replaced exporting to a separate audio program and back, which is where a four minute film loses an evening and gains a sync error. The free version does all of it.
HandBrake is the second, free and open source, and it is what makes the last half day a half day. It replaced writing ffmpeg command lines from a forum post, and the practical gain is that the settings that decide whether a phone plays your file are in front of you with names on them. The editors, and what each is good at covers the choice properly, including the ones worth using when the machine is old.
The frames themselves, and what it takes to make one that was never in front of a lens, is graphics and characters that move. The afternoon at the depot is the same afternoon as light, a lens, and deciding what to leave out, up to the point where the shutter stopped being a choice. What the two evenings never settled is smaller and harder than either: the terminus sounds better than the waffle iron does, buses and a tannoy and Marta's voice underneath all of it, and a film cut to that sound is a different film from the one cut to these pictures. Which of the two is leading gets decided before a timeline exists, and that argument is in recording sound, shaping it, and making music.