What Is a Paper Edit? A Video Editor's Guide

A paper edit is a written plan for a non-fiction edit: you transcribe your footage, highlight the lines worth keeping, and arrange those quotes into a structure on the page before you ever open a timeline. The phrase sounds quaint, like something from a cutting room full of grease pencils. It is not. The paper edit is the single most reliable way to find a story inside hours of talking, and it is the reason an editor can hand a producer a coherent first cut instead of a shrug.

Here is the core idea, and it is worth tattooing somewhere: edit the words first, and the pictures follow. In scripted film you start with a screenplay and shoot to match it. In documentary, vlogs, podcasts, courses, testimonials, panels, any footage where people just talk, you have the opposite problem. The story is buried in what was actually said. A paper edit is how you dig it out.

What a paper edit actually is

Robert McKee has a useful frame for this. A screenplay is built outside-in: you imagine the whole, then write toward it. A paper edit is built inside-out. You can only promise what the footage contains, so you sculpt the structure from the raw transcript rather than imposing one on top of it. That difference changes everything about how you work. You are not writing a story. You are finding the one that is already there.

Mechanically, a paper edit is a document. Speaker names, timecodes, and the lines you want, in the order you want them. Some editors do it in a Google Doc with colored highlights. Some use index cards. Plenty still print the transcript and attack it with a pen. The format matters less than the discipline: you are making structural decisions in text, where they are cheap to change, instead of in a timeline, where they are expensive.

A good paper edit document has a predictable anatomy. Every line you keep carries three things: who said it, where it lives in the footage (a timecode, ideally word-level), and the words themselves. Around those lines you add structure, section headers for the acts or beats, the occasional note to yourself ('need b-roll here,' 'check if we can use this name'), and placeholders for voiceover or narration you have not written yet. Read top to bottom, that document is your film in miniature. If it reads well on the page, the edit has a real chance. If it does not, no timeline wizardry will save it.

The method was codified by Michael Rabiger in Directing the Documentary, first published in 1987, in a chapter literally titled 'The Paper Edit: Designing a Structure.' It was not invented on a precise date by a single person, it grew out of necessity, but Rabiger is where it got written down and taught. Read that chapter if you cut long-form anything.

Why it exists: cutting negative was a one-way door

In the film era, editing meant physically cutting the negative or the work print. A wrong cut was not an undo away. It was a splice you could not unmake without losing frames. So editors resolved structure on paper, where a mistake cost a pencil eraser, not a roll of film. Transcripts were cut up with scissors and reassembled with tape. Audio was bladed with a razor and a grease pencil. The paper edit was a survival tactic before it was a best practice.

That constraint is gone. Your software has infinite undo. So why keep doing this?

Because reading is faster than watching, and changing your mind on a page is faster than re-cutting a sequence. The 2019 meta-analysis by Marc Brysbaert, covering 190 studies, pegged average adult silent reading at roughly 238 words per minute. People speak in interviews closer to 150 words per minute. Brysbaert's research means you can scan an hour of talking in well under half the time it takes to watch it, and you can re-scan it ten times while you test structures. Try ten structural passes in a timeline and you will lose your afternoon.

There is a second, quieter reason, and it is the one that actually wins arguments with producers. Structure is the most expensive thing to get wrong, and the cheapest thing to fix on paper. If you discover in the rough cut that your whole second act is in the wrong order, you are re-trimming, re-syncing, re-laying b-roll, possibly re-doing music. If you discover the same thing reading a transcript, you drag a paragraph. The paper edit pushes the riskiest decisions to the cheapest stage. That is the whole economic argument, and it holds on a one-person YouTube channel and a network documentary alike.

And there is a review benefit most editors underrate: a paper edit can be read and approved by people who would never sit through a rough cut. A client, a subject, a legal reviewer, a showrunner, they can scan a structured document in minutes and flag a problem before you have built anything. Catching 'we cannot use that quote' on the page instead of in a locked sequence is the kind of save that pays for the method ten times over. If client sign-off is part of your life, getting approval before you edit leans hard on this.

The terminology chain, so you can talk to editors

Documentary post has its own vocabulary, and the terms drift between shops. Here is the chain most people mean:

Selects

Also called pulls or lifts. These are the lines worth keeping, the moments you highlight in the transcript. A selects reel is the video version: the good bits strung together with nothing thrown away yet.

String-out and radio cut

A stringout is your selects laid end to end in rough order. When the project is audio-led, this is often called a radio cut: you build the whole thing on the spoken word, ignoring picture, and judge whether it holds together by ear alone.

Assembly, rough, fine, lock

The assembly edit is the first full sequence in the timeline, everything in order, nothing trimmed for pace. A healthy assembly comes in under roughly 140 percent of your target length. From there you refine into a rough cut, then a fine cut, then picture lock, then the final cut. No Film School keeps a solid glossary if you want the wider list.

The paper edit sits before all of it. It is the blueprint the assembly is built from. People mix up the terms; if you want the fine distinction, paper cut vs paper edit walks through it.

A real example: 12 hours down to a 6-minute story

Say you shot four interviews for a brand documentary, about 12 hours total. Watching all of it once is a day and a half gone. Instead:

You transcribe everything with timecode. You read the four transcripts in an afternoon and highlight maybe 80 strong lines. You copy those into a fresh doc, grouped loosely by theme: origin, the hard year, the turning point, where they are now. You read that document out loud. It is twice too long and the middle sags. You cut 30 lines and reorder two sections so the turning point lands later. Now it reads like a story. Total elapsed time: one day, no timeline yet.

Only then do you build. The assembly comes together in a couple of hours because every decision is already made. You are conforming a plan, not discovering a structure while fighting clip handles. That is the whole pitch. The hard thinking happened where thinking is cheap.

The same shape works across formats, the variables just change. A webinar you are repurposing has a clear agenda you can carve into chapters. A podcast you are clipping is a hunt for self-contained moments. A testimonial needs one emotional arc pulled from a meandering chat. In every case the paper edit is where you decide what the piece is before you spend an hour making it look nice.

The honest criticism: a transcript is not the film

Not everyone buys it, and the best objection comes from one of the great documentary directors. Errol Morris flat out refuses to work this way. In a conversation with Transom he put it bluntly: 'Paper cuts? No, never, never. Paper cuts give you a very false idea.' He went further: 'I don't edit from the transcripts, ever. I edit from the film.' His point is that performance, hesitation, the look on a face, the catch in a voice, none of that survives on the page. A line that reads flat can be electric on camera, and vice versa.

He is right, and you should take it seriously. A paper edit is a hypothesis, not a verdict. The fix is simple discipline: treat your paper structure as a draft, then verify every chosen line against the actual clip before you lock anything. Read for structure, watch for tone. This is exactly why ScriptCut keeps word-level timecodes on every line and lets you play any selected line on the spot, so the paper stage and the watch-it stage are not two separate worlds. You get the speed of reading and the truth of the footage in one pass.

How the pros actually use it

This is not a beginner technique that working editors outgrow. It is how a lot of long-form work gets made. Documentary editors routinely distill dozens of hours of interviews into a few hours of assembly using paper cuts, because there is no other sane way to hold that much material in your head at once. The transcript is the index; the highlights are the memory.

Public radio built an entire culture around it. This American Life and Radiolab structure stories by the spoken word first, logging and arranging text and audio before they finish anything. Jessica Abel documents that listen-first, hand-logging discipline in Out on the Wire. The instinct is the same one Rabiger taught: resolve the story where it is cheap to resolve, then commit.

The BBC has talked openly about paper-editing factual programmes and then conforming the result in DaVinci Resolve, and editors who work this way will tell you it saves days on a long project, not minutes. The savings are not in the typing. They are in not building three different timeline structures to discover which one works.

The paper edit moved into software

The method never died, it migrated. Scissors and tape became transcript tools. Pietro Passarelli's open-source autoEdit grew out of newsroom work to turn transcript selection straight into a rough cut, and a process that took weeks compressed into days. Adobe shipped Text-Based Editing in Premiere Pro at NAB 2023, in the 23.4 release, so you can build a sequence by selecting words, delete a clip by deleting its text, and reorder by cutting and pasting. Descript and Reduct are built entirely around editing the transcript. The paper edit is now the default mental model for a whole category of tools, including the paper edit software editors reach for today.

The reason this matters is not novelty. It is that the gap between the paper stage and the cut stage finally closed. In Rabiger's era you did the paper edit, then rebuilt every decision by hand in the cutting room. Now a highlighted line can carry its own timecode straight into the timeline, so the plan and the cut are the same artifact. That is the genuine advance, not the AI gloss, the removal of the manual conform between deciding and building.

The inside-out discipline, in practice

The McKee idea, that a paper edit is inside-out, has a practical consequence worth sitting with. Because you can only use what the footage contains, the paper edit forces a kind of honesty most planning does not. You cannot write a transition that nobody said. You cannot promise a payoff that was never filmed. Every line in the document is a real thing that exists on a hard drive somewhere.

That constraint is a gift. It is why a paper edit rarely produces the disappointment of a beautiful plan that collapses on contact with the footage, the curse of editing to a script that was written before the shoot. You are not imagining the film, you are auditioning the pieces you actually have. When a structure works on paper, it works because the material genuinely supports it, not because you wished it would. New editors sometimes resist this, they want to impose the story they hoped to get. The footage always wins that fight. The paper edit just lets you lose it cheaply, in an afternoon of reading, instead of expensively, in a week of cutting.

Common mistakes

Three traps catch people who are new to this.

Highlighting too much. If half the transcript is yellow, you have not made decisions, you have just reread it. Be ruthless. A select is a line you would fight to keep, not a line that is fine.

Editing for the words and forgetting the cut. A paper edit that ignores how lines will join, where you will need a J-cut or L-cut to smooth a transition, or where you will need b-roll to cover a jump, will fall apart in the timeline. Note those moments as you go.

Never checking the footage. This is Morris's warning. Skip the verification step and you will lock a structure built on a line that does not actually play. Watch your selects.

When a paper edit is the wrong tool

It is not for everything, and pretending otherwise loses you credibility. If your footage is scripted or visually driven, a narrative short, a music video, a commercial built around shots rather than speech, the transcript tells you almost nothing useful, and you should be cutting picture from the start. The paper edit earns its keep when the story lives in what people say: interviews, documentaries, podcasts repurposed to video, courses, webinars, panels, testimonials, talking-head and UGC content. The more talking, the bigger the payoff.

There is also a judgment call on very short pieces. For a 30-second clip from a single take, a formal paper edit is overkill, just cut it. The method scales with length and with the volume of raw material. Three hours of interview for a six-minute film? Paper edit, always. One clean two-minute answer? Probably not.

How a paper edit connects to the rest of post

It helps to see where this sits. The paper edit produces a plan. That plan becomes a stringout when you drop the selected lines onto a timeline in order. The stringout becomes an assembly edit once you commit to that order as the story. From there it is a rough cut, a fine cut, then lock. Every later stage is faster and calmer when the paper edit was done well, because you are refining a structure instead of searching for one. Skip it and the search leaks into every later stage, which is why some edits never seem to settle.

The takeaway

If you cut anything where people talk for a living, learn to paper edit. It is the cheapest place to be wrong, and being wrong cheaply is most of the job. Build the words, prove them against the footage, then assemble. When you are ready to try it without printing anything, ScriptCut runs the whole flow, transcript to highlighted selects to an exported timeline, in one place. Start with one interview and time yourself. You will not go back.

Sources

• Michael Rabiger, Directing the Documentary (Routledge)
• Errol Morris, conversation at Transom
• Brysbaert (2019), reading rate meta-analysis
• Adobe, Text-Based Editing announcement (NAB 2023)
• autoEdit, Source/OpenNews