What is a paper edit, and how do you make one?
Before anyone touches a timeline, the film gets written on paper, from the words people actually said. Here is how that works, step by step, and when it is worth the time.
What a paper edit is
A paper edit is the order of your film written out as text before you edit it: which soundbites go where, in what sequence, with notes on the pictures that cover them. Each line points back to a clip and a timecode, so the plan can be built in the edit without guessing.
Documentary and news editors use it because their films are found, not written. Nobody scripted what the interviewee said. The story lives somewhere in hours of transcripts, and it is much faster to move sentences around on a page than clips around on a timeline.
It also helps when someone else signs off. A producer, a client or a subject can read a two-page paper edit in ten minutes and say “move the bit about the factory earlier”. Asking them to watch three rough cuts to say the same thing costs everyone a week.
Step 1: Transcribe everything
You cannot plan from what you remember. Get every interview and every piece of sync sound into text, with timecodes, before you start choosing.
- Transcribe the whole recording, including the questions and the false starts. The best line often comes right after a stumble.
- Keep timecodes in the transcript, at least every paragraph. A soundbite you cannot find again is a soundbite you do not have.
- Name files so the transcript and the clip match: the transcript of A012_C004 is called A012_C004.
- Mark who is speaking. In a two-camera interview, note which camera has the better angle for each answer.
Read the transcripts once without a highlighter. You are listening for the story the material can actually tell, which is often not the one in the shooting plan.
Step 2: Log and highlight the soundbites
On the second read, mark every line that could carry part of the film. Be generous now and strict later. For each one, write down the clip, the in and out timecode, and a few words on what it does.
- Clip and timecode: A012_C004, 00:14:32–00:14:51
- What it says, in your words: “why they nearly gave up”
- How it sounds: clear, emotional, slightly long, starts mid-sentence
- Possible pictures: the empty workshop, her hands on the old machine
That third line is the one people skip. A transcript tells you what was said, not how. Two takes with the same words can be a strong answer and a flat one. Note the delivery while it is fresh, or go back and listen before anything goes into the plan.
Step 3: Write the story order
Now build the film as a list. Start with the A-roll, the soundbites that carry the story, and put them in the order the viewer should hear them. This is the spine. If the spine reads well as plain text, the film usually works.
Then add B-roll notes beside each soundbite: what the viewer sees while the words play. Write it as intention (“show the scale of the warehouse”) and, where you know it, the clip that does it. Mark where the picture should breathe with no talking, and where a line should run under a different shot.
- Open on something that makes the viewer lean in, not on the introduction.
- Give each section one job: the problem, the turn, the result. If a soundbite does not serve its section's job, cut it or move it.
- Read the spine aloud. Your ear catches repetition and missing links that your eye skips.
- Time it roughly. Read the spine aloud with a stopwatch, then add time for the pauses and the B-roll moments with no talking. If it is already over length on paper, it will be over length on the timeline.
A simple paper edit template
You do not need software for this. One row per beat, in film order, with these fields:
- Beat number and section (for example: 3, The turn)
- Speaker
- Clip name and timecode in and out
- The words, copied exactly from the transcript
- B-roll over it, or “on camera”
- Sound notes: music in, natural sound up, a pause to hold
- Rough duration
Keep the words exact. If you paraphrase in the plan, you will plan a line nobody said and find out in the edit.
From paper to assembly
The assembly is the paper edit built literally: every soundbite in order, at the timecodes you wrote, with the B-roll you named. Do it mechanically and fast. It will be too long and it will feel stiff. That is normal.
- Lay the A-roll spine first, with nothing else on the timeline. Watch it. This tells you whether the story works.
- Trim the soundbites to their real starts and ends: breath to breath, not word to word.
- Add B-roll over the joins and over the lines that describe something the viewer should see.
- Only then music, titles and sound design.
Expect the paper edit to change once you watch it. A line that read well may sound flat, and a moment you skipped may be the best thing in the film. The plan is there to get you to a first cut quickly, not to stop you editing.
For a worked example of the whole route, from a long interview to a short film, see cutting a short documentary from a long interview.
A paper edit is not a script
A script is written first and then performed or filmed. It can say anything. A paper edit is written last, from footage that already exists, and it can only use what is there. That limit is the point.
In practice the two meet. A corporate film might start with a script of what it should say, then become a paper edit once the interviews are in and you see which of those points people actually made, and how well. Where the script asks for something nobody said, you need a new line, a voice-over, a title, or a smaller film.
Common paper edit mistakes
- Editing for words, not delivery. The transcript makes every line look equally good. Listen to every soundbite before it goes into the plan.
- Too many short bites. A film made of six-word fragments feels nervous. Let some answers run.
- Writing B-roll you did not shoot. Check the footage before you write “drone over the city” into the plan.
- Ignoring the cut points. A line that starts in the middle of a breath or ends on a rising tone is hard to cut, however good it reads.
- Frankenbites without honesty. Joining half a sentence from minute 12 to half a sentence from minute 47 can make someone say a thing they never said. Tightening a ramble is normal; changing the meaning is not. Ask yourself whether the person would recognise the line as theirs. Mark every join in the paper edit, so whoever signs off knows the line is built from two places.
When to skip the paper edit
Not every film needs one. Skip it, or keep it to a few lines, when:
- The film is short and mostly visual, like an event highlight cut to music.
- There is a script and it was shot as written. You already have the order; you are choosing takes.
- There is very little dialogue to choose from, such as one short interview answer.
It earns its time on interview-led films: documentaries, customer stories, testimonials and anything with more than an hour of talking. The more footage, the more the page saves you.
Doing this in Endroll
Endroll works like a paper edit that assembles itself. You point it at a footage folder and it prepares everything on your computer: it listens to the speech, finds the faces and describes what is in each frame. Then you say what the film should say, as a brief or a screenplay, and Endroll's AI finds those moments in your footage and builds the cut.
- Star into the paper edit: star the moments you like while you browse in Media, and they collect into a screenplay you can draft from. Search everything finds shots by a spoken phrase or by what is in the frame.
- In the Script workspace, Draft a new cut builds a cut from a brief or a screenplay and a target length. ✍ Write the script for me writes one from your brief and what you actually shot.
- Screenplay reading shows a coverage heat on each beat, so you see what footage supports which line.
- The film as text lets you read the cut as its words: drag a section to move it, mark words and press Delete to cut them. ✎ Edit script → Re-cut from this script cuts your edited script exactly as written.
What the AI does not do: it does not invent footage or put words in anyone's mouth when it drafts. Every dialogue line in a script it writes is locked to the exact words of a real take, and a scene with no matching footage is dropped rather than invented. In documentaries, two shots of one camera setup are never cut together. Drafts and re-cuts use credits, with the price shown before anything runs; editing by hand never does.
Endroll runs on Windows 10 and 11 and on Macs with Apple silicon. When the cut is close, finish it in Endroll or open it in Premiere Pro, DaVinci Resolve or Final Cut.