Why the order is the whole thing
Most guides to a craft are a list of techniques. This one is mostly about sequence, because clipping has a property that makes sequence unusually important, and it is a property nobody mentions.
Almost every fault in a clip is discovered in the same place – the last watch-through, just before posting. A caption with a name spelt wrong, a source too small to crop, a moment that turns out to need context it cannot carry: all of them feel identical at the instant you notice them. But they were introduced at completely different points, and what a mistake costs has far more to do with when you made it than with what it was.
That is the argument of this page, and it has a practical shape. Some decisions are worth genuine deliberation because they cannot be revisited. Others can be changed in a minute, and agonising over them is simply wasted care. Knowing which is which is most of what separates somebody who finds this quick from somebody who finds it exhausting.
Everything below links out rather than restating: what clipping is covers the concept, how to clip a video is the two-minute practical version, and this page is the map they both sit inside.
The five stages
Choose the source, find the moment, set the in and out points, crop and caption, check and post. The diagram at the top of the page is those five against what each costs to change afterwards, and the interesting column is the second one.
Choosing the source is barely a stage at all – usually it is a file you already have, chosen by whoever pressed record. It nevertheless caps everything downstream, because a vertical crop keeps only about a third of the width and no export setting invents detail that was never recorded.
Finding the moment is the one judgement no tool takes off you, and the one that is genuinely irreplaceable. If the passage you chose needs thirty seconds of setup to make sense, no amount of care at the later stages rescues it.
Setting the in and out points is where a clip starts feeling made rather than torn out of something. It is a skill, it takes practice, and it is also trivially revisable – which, as the next-but-one section argues, is an unusually good combination.
Cropping and captioning is the part that looks like editing, and is mostly cheap right up until you export. After that, changes cost quality rather than time, which is the argument for doing the whole job in one pass.
Checking costs nothing and catches more than any export setting. It is also the stage people skip, precisely because by then the clip feels finished – which is the moment it is least trustworthy.
What was decided before you started
Here is the uncomfortable version of the argument. Three facts about your footage – how big it is, where the file came from, and what the audio sounds like – between them settle most of what is available to you later.
What are you cutting from?
Where did that file come from?
What does the audio sound like?
- A vertical crop that still looks sharp—
- Punching in on one speaker in a two-shot—
- Captions you correct rather than write—
- Letting something else find the moments—
- Changing your mind after you have exported—
—
Not one of these five doors is opened or shut by the editor you use. They are all decided by three choices made before any editing starts, two of which are usually made by whoever pressed record. This is the whole argument of the page in one panel: the craft is mostly upstream of the part that looks like the craft.
Change the answers and watch the doors move. Then notice what does not move: nothing in that panel responds to which editor you use. The software question, which is the one almost everybody asks first, has no effect on any of it.
This is not an argument for despair, and it is not an argument that footage has to be perfect. Two of the three inputs are usually fixed before you arrive, and a clip made from an imperfect source can be entirely good. It is an argument for knowing which constraints are real before you spend an hour fighting one – and, if you are ever the person pressing record, for understanding that the most consequential clipping decision you make happens then.
One pairing in that panel is worth singling out, because it is the least obvious thing on this page. The audio that gives you correctable captions can still be too poor for something to make choices from. Displaying words and judging which words mattered are different demands on the same transcript, and the second one fails first – quietly, and without telling you which lines it was unsure of.
The two worth slowing down for
Two decisions account for nearly all of the irreversible cost, and they share a shape: both are made before any editing starts, and both are discovered at the very end.
The source. Use the original rather than a copy, and the largest version you have. A re-download or a screen recording is already a generation down, and each generation discards detail the encoder judged you would not miss – judgements that compound and cannot be reversed. If you have a 4K master, that is the file, even if the clip is going out small. Why clips look blurry covers how to tell what you are actually looking at, and the diagnosis is simpler than it sounds: pause on a still frame.
The moment. This is the irreplaceable one. A clip has to make sense to somebody who arrives with no context, in the middle of scrolling, with the sound off as often as not. Most passages that feel strong inside a long recording fail that test, and they fail it for a reason you cannot edit away: they were strong because of what came before them. Choosing the best moments is the judgement itself, and clipping content is the earlier question of whether a given recording contains any such moment at all – which is worth asking first, because a lot of perfectly good recordings do not.
Everything else on this page is recoverable. These two are not, which is the entire reason to give them disproportionate attention.
The ones to make fast
The corollary matters just as much, and it is the part that saves people actual hours. Caption wording, the exact in-point, the title card, the crop framing before you export, the thumbnail, which platform to post to first – all cheap. Make the decision, look at the result, change it if it is wrong. Deliberating in advance buys nothing that looking would not have told you faster.
But this needs one clarification, because otherwise it contradicts the advice everywhere else on this site. Cheap to change is not the same as unimportant.
The in-point is the clearest case. Where a clip starts is among the most consequential few frames in it – a cut two frames late removes the first consonant, which every viewer hears and no caption repairs. It is also completely trivial to move. High stakes and low cost is the best combination a decision can have, and the right response to it is not caution but iteration: try it, watch it, move it, watch it again. How to clip a video has a demo that makes the size of that window obvious in about ten seconds of clicking.
The mistake people make is the reverse of both rules: agonising over the export preset, which barely matters and is easily changed, while accepting the first moment that seemed good, which matters enormously and cannot be changed later at all.
Which page for which stage
The rest of this site is organised by topic, which is the right way to find something when you know what you are looking for and the wrong way when you do not. So here it is arranged by stage instead.
- 1. Choose the sourcebefore an editor is openwhy clips look blurry, what a vertical crop costs you
- 2. Find the momentthe irreplaceable judgementwhether there is a clip in it at all, choosing the best moments, letting something else choose
- 3. Set the in and out pointswhere a clean clip is decidedhow to clip a video, writing better hooks
- 4. Crop and captioncheap, until you exporthorizontal to vertical, captions, and correcting them
- 5. Check, then postthe only free stageYouTube Shorts, Reels, TikTok, trimming the YouTube original
Two cross-cutting questions sit outside the stages. Which tool, which is genuinely a question about how much footage you have rather than about features – best video clipping tools and free tools both sort it that way. And doing it at volume, where the hard part stops being cutting and becomes finding: long video to short clips.
Start at stage three
Mark the in and out points – nothing to install
FAQ
What is the best order to learn video clipping in?
Backwards from how it is usually taught. Most guides start with software, which is the stage that matters least. Start instead with judging whether a recording contains anything worth cutting, because that skill survives every change of tool – clipping content is that skill. Then learn in and out points, then captions and crop, and treat the software as the last and least interesting question.
What is the single most common mistake?
Cutting on the word rather than a beat before it, by a wide margin. It is also one of the cheapest to fix, which is a useful pairing. The most expensive common mistake is different: choosing a moment that cannot stand on its own, which is not discovered until the clip is finished and cannot be repaired at that point. How to clip a video covers the first and choosing the best moments covers the second.
Does the tool I use actually matter?
Less than the category of tool, and far less than the footage. None of the five constraints in the panel above moves when you change editor – they are set by the recording and by where the file came from. Tools differ mostly in how much footage they are built to handle, which is why the tools page sorts them by volume rather than by features.
Can I fix a source that was recorded too small?
Not really, and this is worth being blunt about because a great deal of software implies otherwise. Detail that was never captured is not hidden somewhere in the file waiting to be sharpened out. You can make a small source look as good as it can look, and that is the whole of what is available. Why video clips look blurry covers how to tell what kind of softness you are looking at, which decides whether anything can be done.
How long should one clip take?
From a video you have already watched, with the moment already in mind, a few minutes. The number is unhelpful on its own though, because the time is not evenly spread – almost all of it goes on finding the moment and on the handful of frames at the start. If a clip is taking much longer than that, the cause is usually that the moment was not really chosen before the editing began.
Do I need an AI clipper to do this properly?
Only if finding the moments is genuinely the bottleneck, and for one video it is not. What these tools do is read a transcript and propose candidates, which means they are strong on recordings that are mostly talking and weak on everything else, and they leave you the checking. AI clipping against manual editing sets out where the saving is real and where it is imaginary.
I have a backlog of long recordings. Where do I start?
Not at the beginning of the backlog. Start by working out which recordings contain clips at all, because a backlog is usually mostly empty and the cost of finding that out late is enormous. Clipping content is the triage, turning long videos into short clips is the route once something has passed it, and generating several clips from one source is the batch.
Where does this go if I want to do it for other people?
The craft above is the same; what changes is that somebody else is judging the result and paying for it. Building a clipping portfolio covers showing the work, packaging clipping services covers scoping it, and earning money clipping covers the routes and what each one actually guarantees.
Related: what video clipping is, how to clip a video and clipping content.