
A rough cut is the draft of your edit where the structure is settled and you start refining: you trim clips for pace, shape performance, lay in temporary music and b-roll, and make it feel like the film, even though sound design, color, and final polish are still missing. It is the first version that resembles the finished piece rather than a pile of organized clips.
The rough cut is where editing starts feeling like editing. The assembly was a structural test. The rough cut is where you make it good.
The assembly edit answers 'does the structure work?' The rough cut answers 'does it play well?' By the time you are in a rough cut, you have stopped asking whether a scene belongs and started asking whether it lands. You trim the dead air. You pick the better of two takes. You start covering seams with b-roll and smoothing dialogue with J-cuts and L-cuts.
As Filmsupply describes it in their cut-stages guide, the rough cut 'looks like a movie' but lacks final sound design, effects, and color. That is the right mental image. A viewer could follow it. A client could react to it. It is not done, but it is no longer raw.
The shift in mindset is worth naming because new editors often carry the wrong one into this stage. In the assembly you were a structural engineer asking whether the building stands. In the rough cut you become a storyteller asking whether it moves people. Those are different muscles. You start caring about momentum, about where a viewer leans in and where they check their phone, about the difference between a scene that is technically complete and one that actually plays. If a section bores you on the third watch, it will bore the audience on their first. Trust that reaction and cut accordingly.
Paper edit, stringout, assembly, rough cut, fine cut, picture lock, final. The rough cut is the middle of that chain and usually the longest-lived stage, because this is where most of the actual decisions get made and remade. Frame.io's stages guide maps the same progression if you want it laid out.
Expect more than one rough cut. Editors talk about rough cut v1, v2, v3. Each pass tightens. The first might still be loose and overlong; by v3 it is close to a fine cut. That iteration is normal and expected, not a sign something is wrong.
A useful way to run those passes is to give each one a single job instead of trying to fix everything at once. One pass for pace: trim the fat, nothing else. One pass for coverage: lay in b-roll and fix the visual gaps. One pass for sound and rhythm against temp music. Doing them as separate passes is faster than circling the whole edit looking for any problem, because each pass keeps your attention on one kind of decision. Editors who try to fix pacing, coverage, and rhythm simultaneously tend to make slower progress and miss things, the mind does not multitask well across cutting decisions.
Concretely, a rough cut pass usually means:
Trimming for pace. Cut the breaths, the throat-clears, the slow run-ups to a point. Tighten until it moves. This is also where you remove filler words that survived the assembly.
Choosing performance. Where you have options, pick the take that plays best, not the one that reads best. This is the moment the footage overrules the page.
Temp music and sound. Drop in a temporary track to feel the rhythm. Editing to music changes how you cut, so do it early even though the final track may differ.
Covering and bridging. Add b-roll over jump cuts, use split edits to ease transitions, drop in titles as placeholders.
Removing every um, you-know, and false start is satisfying, but doing it in the assembly is premature, you might cut the whole section. The rough cut is the right moment, because by now you know which material survives. As you trim for pace, strip the verbal clutter from the lines that are staying. Removing filler words cleanly is a big part of what makes a rough cut start to feel professional, a talking-head answer riddled with ums reads amateur even when the content is strong. Tightening that language is half of what separates a rough cut from a raw assembly. On talking-head video especially, this is where the piece stops sounding like a first take.
You have a 12-minute assembly of a brand documentary. The structure is right, you confirmed that watching the assembly. Now the rough cut.
First pass: you trim every interview answer to its core, dropping run-ups and tangents. The 12 minutes becomes nine and immediately breathes better. Second pass: you add b-roll over the talking heads so it is not 9 minutes of a face, and you use an L-cut so the founder's voice carries into the factory footage. Third pass: you lay temp music under the emotional turn and discover the scene runs 15 seconds too long against the music, so you tighten it. Now it is a 7-minute rough cut that a client can watch and respond to. None of it is finished, but the film exists.
The rough cut is where feedback enters, and bad feedback handling is where projects die. When you share a rough cut, frame the stage out loud: this is a rough cut, structure and pace are what I need eyes on, color and final audio are coming. Without that frame, you get notes about a music track you have not licensed and a grade you have not done, and the structural feedback you actually need drowns.
Collect notes against timecode, group them into structural versus cosmetic, and push the cosmetic ones to the fine cut where they belong. A reviewer saying 'this section feels long' is a structural note worth acting on now. The same reviewer saying 'the title font looks off' is a fine-cut or finishing note. Sorting feedback by stage is half of staying sane across multiple rough-cut rounds.
People muddle the rough cut with its neighbors, so to be precise: the assembly before it is untrimmed and only tests structure. The fine cut after it works at the frame level on a settled edit. The rough cut is the broad middle where you do the most reshaping, and it is normal for it to look rough, hence the name. If your rough cut already looks finished, you either spent too long too early or you mislabeled a fine cut. The editor glossary is worth a skim if the terms blur on your team.
Perfecting frames before the cut is settled. Frame-accurate trims and careful color on a rough cut are premature. The shot might get cut. Refine structure and pace first; polish frames in the fine cut.
Falling in love with temp music. Temp tracks help you cut, but cutting too tightly to a song you cannot license means re-editing when the real track arrives. Cut to the rhythm, not to the exact hits.
Showing it raw without context. A rough cut shown to a client with no framing invites notes about color and audio you already know are coming. Tell them what stage it is, so feedback targets structure and pace. Getting client sign-off early avoids this entirely.
The fastest route to a strong rough cut is a strong assembly, and the fastest route to that is a paper edit. If your structure is right before you trim, every rough-cut pass is refinement instead of rescue. In ScriptCut you build the story from the transcript, verify each line by playing it, then export a timeline into your NLE, so you arrive at the rough cut stage with the structure already proven. For the interview-specific flow, see how to edit an interview faster.
You leave the rough cut stage when the structure and pace stop changing in meaningful ways and the remaining notes are about details, a frame here, a transition there, a take swap. That is the signal you have crossed into fine cut territory. A practical test: if your last two rounds of changes were all small and cosmetic, the rough cut is done, lock the structure in your head and move to frame-level work. If you are still cutting whole scenes, you are not done, no matter how many versions you have produced.
Resist the urge to declare it done early to hit a deadline. A rough cut that still has a structural problem will drag that problem through the fine cut, the sound mix, and the grade, and fixing it late costs many times what it would have cost now. The rough cut is the last cheap place to fix story. Spend the rounds it needs.
Here is the pattern across hundreds of edits: the projects with painful, endless rough cuts are almost always the ones where structure was never settled before trimming began. The editor is trying to fix the story and the pacing at the same time, which is two jobs fighting each other. When the structure is locked first, in a paper edit and confirmed in the assembly, the rough cut becomes pure refinement, and refinement converges. You are tightening a thing that works, not rebuilding a thing that does not. That is why front-loading the structural decisions, on the page where they are cheap, pays its biggest dividend right here.
The rough cut is where the film becomes a film. Settle structure in the assembly, then spend your rough cut passes on pace, performance, and rhythm, not on frames you might delete. Iterate, expect several versions, and keep the polish for later. Next stage up is the fine cut. To get here with the structure already locked, try ScriptCut.
A rough cut refines the whole film: pace, performance, temp music, b-roll. A fine cut polishes each moment frame by frame and gets the edit close to final before picture lock.
Several. Editors routinely go through rough cut v1, v2, v3, each tighter than the last. Multiple passes are expected, not a problem.
No. A rough cut looks like a movie but skips final color, sound design, and effects. Those come after picture lock so you do not redo them when the edit changes.
Yes, it helps you feel pacing, but cut to the rhythm rather than exact musical hits so you are not forced to re-edit when the licensed track arrives.