Two editors are handed the identical footage — the same interview, the same B-roll, the same double-system audio from a documentary shoot — and told to have a first cut by Monday. One of them opens the timeline on Saturday morning and is cutting...
Prerequisites
- 26
Learning Objectives
- Offload and ingest camera and audio media safely, verifying every file before a card is reused.
- Design a project folder structure and a set of bins that make any clip findable in seconds.
- Explain what a proxy is and generate proxies to keep an edit responsive on modest hardware.
- Sync double-system audio to picture using a clap, matching waveforms, or timecode.
- Cull raw footage into selects and lay them into a string-out that becomes a first cut.
- Run a repeatable edit-prep checklist that takes a shoot from card to string-out.
In This Chapter
- Overview
- Learning Paths
- 27.1 Ingest and offload: get the media in safely
- 27.2 Folder structure and bins
- 27.3 Proxies and why they save your edit
- 27.4 Syncing double-system audio
- 27.5 Selects and the string-out
- 27.6 A repeatable edit-prep workflow
- Production Checkpoint
- Summary
- Spaced Review
- What's Next
Chapter 27: The Editing Workflow and Media Management
Overview
Two editors are handed the identical footage — the same interview, the same B-roll, the same double-system audio from a documentary shoot — and told to have a first cut by Monday. One of them opens the timeline on Saturday morning and is cutting scenes by lunch: their clips are named and binned, their audio is already married to picture, their footage plays back without a stutter, and the good takes are flagged and waiting. The other spends Saturday hunting. Which folder had the recorder files? Was the good answer in take 3 or take 7? Why does the 4K footage lurch and drop frames every time they scrub? By Sunday night they have organized nothing, cut nothing, and quietly started to hate the project.
Same footage. Same software. Same deadline. The entire difference happened before either of them made a single cut — in the unglamorous, invisible work of getting the media in, getting it organized, getting it synced, and getting it to play. That work is the subject of this chapter, and professional editors will tell you a quiet truth about it: it is where the edit is actually won or lost. The cut you are proud of is only possible because, at 2 p.m. on a deadline, you could find the shot.
Chapter 26 put you at the timeline and taught you the cut itself — the fundamental unit, transitions, pacing, and the assembly. This chapter is the half-step that comes logically before all of that and that nobody teaches beginners, because it looks like housekeeping. It is not housekeeping. It is the difference between an edit that flows and an edit that fights you the whole way. We will take a real pile of media — cards from a camera, files from an audio recorder, some B-roll, a music track — and walk it through the exact pipeline that turns "stuff on cards" into "a project you can cut."
In this chapter you will learn to:
- Ingest your media safely — offload it, verify it, and never lose a shoot to a wiped card.
- Build a folder structure and matching bins so any clip is findable in seconds, not minutes.
- Generate and use proxies so a heavy 4K project plays smoothly on the computer you actually own.
- Sync your double-system audio to picture — the editorial payoff of everything you did in Chapter 15.
- Cull your footage into selects and build a string-out that becomes your rough cut.
- Run all of it as a single repeatable checklist you will use on every project for the rest of your life.
Learning Paths
This is the most workflow-heavy chapter in the book, and it rewards the 🎓 Student and 💼 Pro-track reader most — but everyone who wants to finish an edit needs it. Weight your attention like this:
- 📱 Phone-first: §27.1 (offload safely) and §27.2 (bins) are your essentials — even single-system phone footage needs organizing, and phone codecs stutter more than you'd expect, so skim §27.3 (proxies) too. You can lightly skip the timecode parts of §27.4.
- 🎥 Creator: §27.2 (a fast, repeatable bin structure) and §27.3 (proxies) are what keep your weekly edits from bogging down. Learn the round-trip cold.
- 💼 Pro-track: all of it. This chapter is your professional backbone — the habits here are what a client is really paying for when they hire someone who "won't lose the footage." §27.4 (sync) and §27.6 (the repeatable workflow) especially.
- 🎓 Student: the whole pipeline, in order. This is the most examinable chapter in Part VI, and the Production Checkpoint applies every step to your documentary short.
27.1 Ingest and offload: get the media in safely
The shoot is over. The camera cards are in your pocket, the audio recorder has its own card, and somewhere there is a phone with a few extra clips on it. Everything you spent a day capturing exists, right now, in exactly one place each — and that is the most dangerous moment in the entire life of a project. A dropped card, a corrupted file, a card reused by accident before it was copied, and the shoot is simply gone. You cannot re-shoot a wedding, a one-time interview, or yesterday's light.
So the first job in post is not editing. It is getting the media in, safely. We call the whole of that job ingest: getting your footage and audio off the cards and into your editing project — copied, verified, and imported — so it is organized and ready to cut. Ingest has two halves, and beginners collapse them into one and get burned. The first half is the offload: the physical act of copying the files from a card to a drive. The second half is bringing that copied media into your editing software so the project can see it. Offload protects the footage; the import makes it editable. Do the offload first, always, and never let the import be the only copy.
Here is the single rule that matters most in this entire chapter, the one that separates people who lose footage from people who don't: never work off the card, and never reuse a card until its contents live in at least two other places. A card is transport, not storage. It is the least reliable link in your whole chain — small, easy to lose, easy to corrupt, easy to hand back to the camera and format without thinking. The moment you have a shoot on a card, your only job is to get it off the card, twice, before that card is ever at risk again.
FIGURE 27.1 — The offload: from a single fragile card to two safe copies (left to right)
[ CAM card ] ──► verified copy ──► DRIVE A (your working / edit drive)
│ └─► DRIVE B (a second, separate backup drive)
│
[ REC card ] ──► verified copy ──► DRIVE A / AUDIO + DRIVE B / AUDIO
│
▼
ONLY NOW: format the card, in-camera, for reuse.
Two good copies must exist BEFORE any card is wiped. The card is the only
place a shoot can die — get it off, twice, before you risk it.
Notice the word verified in that diagram, because it is doing real work. When you drag files from a card to a drive, the operating system copies them and, most of the time, says nothing about whether every byte arrived intact. Occasionally a file copies with a silent error — it looks like it's there, the size looks right, and then in the edit it won't play, or it plays with a green smear across a critical frame. A verified copy (also called a checksum copy) re-reads every byte it wrote and confirms it matches the source, so a bad copy is caught now, while the card still exists, instead of three weeks later when the card is long since reused. Dedicated offload tools do this automatically; many editing suites do it too. In DaVinci Resolve — the free, professional edit and color suite this book anchors its hands-on steps to — the Media page has a "Clone" tool built exactly for verified card offloads, and the exact menu path in other software is translated in Appendix E. Whatever tool you use, the principle is frozen: copy with verification, not with a naked drag.
⚠️ Common Mistake: formatting a card before it's backed up (or editing straight off it). The two classic, unrecoverable ways to lose a shoot are the same two mistakes every beginner makes once. The first: you're low on cards on the next shoot, you glance at a card, think "I copied that... probably," and format it. You didn't, or you copied it to one drive that then failed. The second: you skip the offload entirely and import straight from the card into your editor to "save time," then pull the card out to use it — and now your project points at files that are gone. The fix is a discipline, not a tool: cards are never wiped and never removed from the workflow until two verified copies exist elsewhere. Make it a ritual you do not think your way out of when you're tired.
This is also the chapter's first payoff of a promise the whole book has been making. On set in Chapter 15 you recorded double-system sound — picture on the camera, clean audio on a separate recorder — precisely because it gives you better sound and a safety copy. All of that discipline lands here, at the offload, as a second set of files from a second device. They are gold, and they are also a second thing that can be lost. Offload the recorder's card with the same paranoia you give the camera's.
✂️ In the Edit. Everything you did "for the editor" on the shoot becomes real at the offload, and the editor is usually you. The labeled, slated takes from Chapter 18; the double-system audio from Chapter 15; the extra coverage you grabbed in Chapter 20 because you were "shooting for the edit" — none of it helps until it survives the trip from card to drive. The most senior habit in this book is boring on purpose: the offload comes before the edit, every single time, no exceptions when you're tired. An editor who has never lost footage is not lucky. They are disciplined at exactly this moment.
A short, sobering illustration of the stakes, before we make the habit routine. Pixar has publicly recounted that during the production of Toy Story 2 in the late 1990s, a stray command began deleting the film's files from the studio's servers — and the backups, it turned out, were not working. Much of the film was recovered only because one supervising technical director, working from home after having a baby, happened to have a recent copy of the files on a personal computer. A billion-dollar studio nearly lost a film to exactly the mistake this section is about: one fragile copy and an unverified backup. If it can nearly happen to Pixar, it can happen to your documentary. (We build the full professional backup discipline — the 3-2-1 rule — in Chapter 37; for now, "two verified copies before you wipe a card" is the piece you need today.)
🔗 Connection. This section is deliberately the ingest-safety slice of a bigger topic. The complete professional system — the 3-2-1 backup rule (three copies, on two kinds of media, one of them off-site), consistent file naming, drive speed, and long-term archive — is the whole subject of Chapter 37 (§37.1–37.3). Here we do only what the edit needs: get the media in, safely, in two places. Don't build your career-long backup fortress today; do build the two-copy reflex.
One more practical layer, because ingest is also where the media first meets the project. Once your files are safely on your drives, you import them — you point your editing software at the folders and let it read the clips in. In every editor this brings the media into the project as clips that reference the files on disk; the editor does not usually copy the media again, it just links to it. That means two things you must respect from this moment on. First: the folders you offloaded into are now load-bearing. If you move or rename those files later, the project will lose track of them (the dreaded "media offline"). Second: this is why organization has to happen now, at ingest — because where a file lands on disk today is where the project will look for it for the rest of the edit. Which is exactly the next section.
🔄 Check Your Eye. From memory: 1. What are the two halves of ingest, and which one must come first? 2. What does a verified (checksum) copy protect you from that a plain drag-and-drop does not? 3. Complete the rule: a camera card may be reused only when ______.
Check yourself
- The offload (copying files off the card to drives) and the import (bringing the copied media into the editing project). The offload comes first — it protects the footage; the import only makes it editable.
- A silent copy error — a file that looks present and the right size but is actually corrupt. Verification re-reads what it wrote and catches the bad copy while the card still exists.
- ...its contents exist in at least two other, verified places. (The card is transport, never storage.)
27.2 Folder structure and bins
You have your media on two drives. Now the question is where on those drives, and how your project will make sense of it — because a project with a thousand clips and no structure is a project you cannot cut. This is the domain of media management: the discipline of keeping every asset in a project — footage, audio, music, graphics, project files, and exports — organized, findable, and safe, from the first offload all the way to the final archive. Media management is not a personality trait some editors are born with. It is a system you set up once, at ingest, and then never fight again.
The system has two mirrored layers. On the disk, you have folders — the actual directories your files live in. Inside the editing project, you have bins: the folders inside your editing software that hold and organize your clips. (The word is a wonderful holdover from film editing, when the trimmed strips of film literally hung on hooks over a cloth-lined bin so an assistant could find and re-thread them. The name survived the move to digital; the job never changed — a bin is where you find your pieces.) The crucial idea is that these two layers should mirror each other: the way your files are arranged on disk should match the way your bins are arranged in the project, so that a clip's home is obvious whether you're looking at your drive or your timeline.
FIGURE 27.2 — A project folder structure (on disk) mirrored by bins (in the NLE)
ON THE DISK (where the files live) IN THE EDIT PROJECT (how the project sees them)
─────────────────────────────────── ──────────────────────────────────────────────
PROJECT-NAME/ 📁 PROJECT
├── 01_FOOTAGE/ ├── 01 FOOTAGE
│ ├── CAM-A/ │ ├── Cam A
│ ├── CAM-B/ │ ├── Cam B
│ └── B-ROLL/ │ └── B-roll
├── 02_AUDIO/ ├── 02 AUDIO
│ ├── RECORDER/ │ ├── Recorder
│ └── ROOM-TONE/ │ └── Room tone
├── 03_MUSIC-SFX/ ├── 03 Music & SFX
├── 04_GRAPHICS-STILLS/ ├── 04 Graphics & stills
├── 05_PROJECT-FILES/ ├── 05 SELECTS ◄─ project-only bin
├── 06_PROXIES/ ├── 06 TIMELINES ◄─ project-only bin
├── 07_EXPORTS/ │ ├── String-out
└── 08_DOCS/ (brief, script, releases) │ ├── Rough cut
│ └── Fine cut
Numbered so they always sort in workflow Some bins (Selects, Timelines) have NO disk
order, not alphabetically. folder — they just point at clips already
filed above. Disk = where; bins = how you find.
Read the figure as a claim about findability. Every kind of asset has exactly one obvious home, and the homes are numbered so they sort in the order you use them rather than alphabetically (footage before audio before music before exports). When, three weeks from now, you need "the room tone from the interview location," you do not search — you know, because room tone has a folder and a bin and it has had them since the day you ingested. Notice too that a couple of bins — Selects and Timelines — have no matching folder on disk. That's correct and worth understanding: those bins don't hold new files, they hold pointers to clips already filed under Footage and Audio, gathered together for the edit. The disk stores media once; the project can point at it from as many bins as are useful.
A quick, honest note on file naming, because it is tempting to solve it all here. Good names matter enormously — a clip called A017_C022 tells you nothing, while DOC_INT-A_Maria_T03 tells you the project, the setup, the subject, and the take. But naming is a deep enough discipline that it has its own home: Chapter 37 (§37.2) builds a naming scheme that survives across projects and years. For this chapter, adopt one light rule and move on: names should be consistent and sortable, and they should never rely on the camera's default gibberish alone. Rename or label as you ingest, enough that a clip's name tells you what it is.
🔗 Connection. File naming that survives is Chapter 37's territory (§37.2), alongside the 3-2-1 backup rule and archiving. Don't over-engineer names now. The folder/bin structure in FIGURE 27.2 is the piece the edit needs; the naming convention is the piece the archive needs, and we'll build it properly when we get there.
What goes in the bins, and how finely you divide them, depends on the shoot. A one-person talking-head has almost nothing to organize — one footage bin, one audio bin, done. A multi-day documentary or a multicam event needs more structure: bins by scene or day, then by camera, so you can find "day two, camera B, the market sequence" without scrolling. The discipline that pays off most is a dedicated Selects bin (or a set of them, one per scene): as you review footage in §27.5, the keepers get gathered here, so your best material is never buried among the outtakes. A useful rule of thumb: organize so that the clip you'll want next is the clip you can find fastest.
✂️ In the Edit. This is where the coverage discipline of Parts II–IV cashes out — or doesn't. In Chapter 7 you shot a wide, a medium, and a close-up so the scene would cut; in Chapter 9 you kept continuity and screen direction so the pieces would fit; in Chapter 20 you grabbed cutaways and full sequences so the edit had somewhere to go. All of that is only usable if, in the project, you can find the matching pieces. Binning your coverage by scene — with the wide, medium, and close-up sitting together, the cutaways grouped, the room tone parked and labeled — is the moment "shoot for the edit" becomes "edit from what you shot." A shoot with great coverage and no organization is a warehouse with no shelves.
🚪 Threshold Concept: you cannot cut what you cannot find. Beginners think organization is clerical — a chore you do instead of the creative work. It is the opposite. Every hour of an edit is a stream of tiny questions: do I have a better take of this line? is there a cutaway that hides this jump? where's the wide? An organized project answers those questions in seconds, so you stay in the creative flow. A disorganized project answers them in minutes of scrolling, and every one of those minutes is a small tax on your attention and your momentum. Findability is not the opposite of creativity; on a real deadline, findability is what protects your creativity. Set up the shelves, and the warehouse becomes a workshop.
🎬 On Set (at the desk): ingest and organize one real shoot. Take any real footage you have — a shoot of your own, or the material you gathered for Project 2 — and run §27.1 and §27.2 on it end to end. Offload it (verified) to two drives, build the FIGURE 27.2 folder structure, import it into a fresh project, and rebuild that structure as bins. Constraint: when you're done, close the project, reopen it, and time yourself finding three specific clips ("the widest B-roll shot," "the room tone," "take 3 of the main answer"). Self-review: if any of the three took longer than ten seconds, your structure isn't fine enough yet — fix it now, while the project is small.
🔄 Check Your Eye. 1. What is the difference between a folder and a bin? 2. Why are the top-level folders in FIGURE 27.2 numbered instead of just named? 3. Why can a Selects bin exist with no matching folder on disk?
Check yourself
- A folder is a real directory on your disk where files live; a bin is a folder inside the editing project that organizes clips. They mirror each other, but the folder holds the media and the bin holds the project's pointers to it.
- So they sort in workflow order (footage → audio → music → exports) rather than alphabetically, putting the things you use first at the top.
- Because a bin holds pointers to clips, not new files. The Selects bin gathers clips that already live in the Footage folder on disk; nothing is copied.
27.3 Proxies and why they save your edit
Your media is in and organized. You drop a 4K clip on the timeline, hit play — and it stutters, drops frames, and lags a half-second behind every button you press. Nothing is broken. You have just met the reason proxies exist. A proxy is a lightweight, easy-to-play stand-in copy of a heavy original clip: a smaller, gently compressed version that your computer can play back smoothly, that you edit with in place of the demanding original, and that you swap back out for the full-quality original only when you export. You cut on the proxy; you deliver from the original. The audience never sees the proxy, and your computer never has to sweat.
To understand why footage gets heavy, recall what you learned about recording formats back in Chapter 3. A modern clip can be punishing to play for three stacking reasons: resolution (4K is four times the pixels of 1080p, 6K more again), bitrate (a high-bitrate file is a firehose of data per second), and, most sneakily, the codec it's stored in. Many camera and phone codecs are Long-GOP — they save space by storing only the changes between frames, which means that to show you any single frame the computer must decode a whole group of surrounding frames on the fly. That is brutal work when you're scrubbing back and forth. A proxy sidesteps all of it: it's lower-resolution, lower-bitrate, and stored in an easy, all-frames-independent codec, so the computer barely notices it.
FIGURE 27.3 — The proxy round-trip: edit light, deliver heavy
FULL-RES ORIGINALS ─┐ (heavy: 4K/6K, high bitrate, Long-GOP — hard to play)
(the master copy) │
│ generate proxies (once, up front; the computer grinds so you don't)
▼
PROXY COPIES ◄──────────── edit on THESE ── smooth scrubbing, instant response
(light: e.g. half-res, │
low-bitrate, easy codec) │ ...cut, trim, build the whole story on the proxies...
│ │
│ at export: relink / switch back to the ORIGINALS
▼ ▼
FULL-RES ORIGINALS ──────► final render at full quality ──► delivery
GOLDEN RULE: proxies are temporary stand-ins. The ORIGINALS are the master.
Never delete the originals. Always confirm you're on originals before you export.
The workflow is a round-trip, and the round-trip is the whole point. You generate proxies once, up front — the computer takes its time making the light copies while you get coffee. Then you edit entirely on the proxies, in comfort. When it's time to deliver, you flip a single switch that tells the project "stop using the stand-ins, go back to the originals," and it renders the finished cut from the full-quality masters. Every editor implements this a little differently — some call them proxies, some optimized media, some let you toggle proxy-on/proxy-off with one button — and Appendix E maps the terms across the common software. The concept underneath is identical everywhere: cut on a copy your machine can handle; finish on the original the audience deserves.
⚠️ Common Mistake: delivering the proxy (and its evil twin, deleting the originals). Two failures haunt the proxy workflow, and both are avoidable. The first: you edit on proxies, forget to switch back, and export — shipping a soft, low-bitrate version of your beautiful 4K film. Always confirm the project is on originals before the final render; make it the first item on your export checklist. The second, worse one: proxies are small and tidy, so at some point you "clean up" by deleting the originals — and now the small, soft proxies are the only copy you have, and the master is gone forever. Proxies are disposable; originals are sacred. If you delete anything, delete the proxies (you can always regenerate them from the originals — never the reverse).
🎒 Gear Note: do you need proxies? (Yes, probably, even on a phone.) Proxies aren't a "big camera" luxury. Whether you need them is a simple question about the gap between your footage and your computer. If 4K footage plays back smoothly and scrubs instantly on your machine, you may not need them — edit on the originals and enjoy it. If playback stutters, lags, or drops frames, proxies (or your editor's "optimized media") will transform the experience for free. And here's the surprise for the 📱 phone-first reader: modern phones shoot highly compressed Long-GOP HEVC/H.265, which is space-efficient but genuinely hard to decode — so phone footage often stutters more on a laptop than footage from a "real" camera in an easier codec. If your phone edit feels sluggish, proxies are the fix, and they cost nothing but a little disk space and one up-front render.
⚙️ Settings Box: a sane starting point for proxies. A place to begin and adjust — not a universal recipe. The right proxy is "as light as possible while still letting you judge framing and focus."
Setting Typical starting choice Why Proxy resolution Half of original (e.g. 4K → 1080p) Big playback win; still sharp enough to edit by Proxy codec An "easy," all-intra editing codec (e.g. a low-data ProRes/DNxHR flavor, or your editor's built-in proxy format) Every frame independent = effortless scrubbing Proxy bitrate Low — proxies are for cutting, not for looking good Small files, fast drives, smooth playback Where they live The 06_PROXIES/folder from FIGURE 27.2Disposable and regenerable; kept apart from masters Relink for export ALWAYS back to originals The audience gets full quality; proxies never ship 🔬 The Tech: why one 4K clip plays and another chokes (skippable). If you skip every Tech box you'll still finish the book — but this one demystifies a daily frustration. Playability is roughly resolution × bitrate × codec difficulty, divided by your computer's decoding muscle. The codec term is the sleeper. Long-GOP (Long Group of Pictures) codecs like most H.264/H.265 store a full "key" frame only occasionally and, in between, store just the differences — so displaying frame 37 can require decoding frames 30 through 37 together. Great for card space (Chapter 3's whole trade-off), miserable for scrubbing. All-intra (also called intraframe or All-I) codecs like ProRes and DNxHR store every frame whole: bigger files, but any frame is instantly available, which is exactly what an editor's constant back-and-forth needs. A proxy is, in essence, a small all-intra copy — you're trading disk space for playback ease, the opposite of the trade your camera made. Some editors also lean on the GPU to decode; a capable graphics card can make proxies unnecessary for footage that would otherwise stutter. But when in doubt, proxies are the universal, hardware-independent answer.
🔄 Check Your Eye. 1. In one sentence, what is a proxy and what do you do with it? 2. Name the three things that make a clip hard to play back. 3. At export time, must you be on proxies or originals — and why?
Check yourself
- A proxy is a light, easy-to-play stand-in copy of a heavy clip; you edit on it for smooth performance and switch back to the original to export.
- High resolution (lots of pixels), high bitrate (lots of data per second), and a hard-to-decode codec (especially Long-GOP, which must decode groups of frames to show one).
- Originals — proxies are soft, low-bitrate stand-ins meant only for editing comfort; the finished render must come from the full-quality master so the audience sees the real image.
27.4 Syncing double-system audio
Now the payoff you've been building toward since Part III. In Chapter 15 you recorded double-system sound — picture on the camera, clean dialogue on a separate recorder — and you were promised that the small extra effort would be worth it in the edit. Chapter 15 even defined sync in the general sense (aligning separate audio to its picture so lips and sound match again) and taught you to make it easy on set with a clap. This section is the other half of that promise: syncing in the edit — the actual editorial act of aligning your separately recorded audio to its picture on the timeline, so the two play back as one. You did the setup on set; now you do the marriage in post.
There are three ways to sync, from most manual to most automatic, and it's worth knowing all three because you'll meet all three.
1. By the clap (the slate). This is why you clapped. A clap — hands, a clapperboard, the slate from Chapter 18 — makes a single, sharp transient: a spike that appears in both the camera's own audio (the scratch track you kept rolling) and the recorder's clean audio. To sync, you find that spike in each file and slide one until the two spikes stack. From that frame on, the whole take is locked. It takes seconds, and it works in any editor ever made, including a phone app.
FIGURE 27.4 — Syncing at the clap: line up one spike and the take is locked
BEFORE (out of sync — the two recordings started at different moments)
camera scratch ────╮________________/\________________ ← clap spike, here
recorder audio ──────────╮________________/\____________ ← SAME clap spike, offset →
AFTER (slide the recorder audio until the spikes stack)
V picture (cam) ______________________/\________________ the moment of the clap
A recorder audio______________________/\________________ spike now aligned under it
^
one shared transient, stacked = the whole take is in sync
No clap? Line up any loud sound present in BOTH files (a door, a hand-slap), or match
the camera's scratch audio to the recorder audio by eye and ear. Slower, but it works.
2. By matching waveforms (auto-sync). Most modern editors can do the clap-alignment for you. You feed the software the camera clips (with their scratch audio) and the recorder clips, and it compares the audio waveforms, finds where they match, and aligns them automatically — often creating a single "synced clip" or "merged clip" that behaves like one piece of media with good picture and good sound. This is why you keep the camera's scratch track even though you'll never use it in the mix: it is the reference the software matches against. Point-and-click, dozens of clips at once. When it works, it feels like magic; when it fails (usually because the scratch audio is too different or too quiet), you fall back to the clap by hand.
3. By timecode (the professional method). On bigger shoots, the camera and the recorder are set to share the exact same running clock — timecode, the HH:MM:SS:FF stamp — either by cabling them together or by "jam-syncing" them to a common source at the start of the day. Then the editor simply aligns clips by their timecode and everything snaps into place with frame accuracy and no clapping at all. It's the fastest and most reliable method, and it's why you'll see productions clap and run timecode (belt and suspenders). For most readers of this book it's optional gear; know it exists so you recognize it when you meet it.
✂️ In the Edit. Here is the exact moment the on-set clap earns its keep — and the exact moment its absence hurts. If you slated every take in Chapter 18's discipline, or even just clapped your hands once on camera, syncing is a few seconds per take and you'll barely think about it. If you didn't — no clap, no scratch track, no timecode — you are now sliding audio back and forth by ear, guessing at the frame where the lips match, for every single take. It is the most tedious job in post, and it is entirely preventable on set with one clap. This is "fix it in pre, not in post" and "you shoot for the edit" fused into a single hand-slap. Next time you're on set and tempted to skip the clap because "we'll figure it out," remember this paragraph.
⚠️ Common Mistake: no reference, and long-take drift. Two sync failures show up again and again. The first: recording double-system with no clap and no scratch audio — the camera mic switched off entirely — leaving nothing to sync against but blind guesswork. Always keep the camera's own mic running as a scratch/reference track, even when the "real" sound is on the recorder. The second, subtler one: on very long takes, cheap devices whose clocks run at slightly different speeds can drift — perfectly synced at the clap, but visibly off by the end of a ten-minute interview. If you see drift, sync in shorter chunks (re-clap every few minutes on long takes), or use a recorder and camera whose sample rates you trust.
♿ Accessibility & Inclusion. Sync is also the doorway to a habit that makes your work reachable and your edit faster: transcripts. Once dialogue is synced, running the interview through a transcription tool (many editors now do this in-app) gives you a searchable text of everything said — which is the backbone of the paper edit you'll learn in Chapter 30, and the raw material for the captions every finished video should carry (a subject we return to at delivery). Transcribing early is a rare two-for-one: it speeds your edit and builds accessibility in from the start rather than bolting it on at the end. Sound is half the picture; a caption is how the other half of your audience hears it.
🔄 Check Your Eye. 1. Name the three ways to sync double-system audio, from most manual to most automatic. 2. Why do you keep the camera's scratch track even though it never appears in the final mix? 3. A ten-minute interview is perfectly synced at the start and visibly out of sync at the end. What's happening, and one fix?
Check yourself
- By the clap/slate (align the shared spike by hand), by matching waveforms (the software auto-aligns), and by timecode (clips share a clock and snap into place).
- It carries the clap spike and is the reference that waveform auto-sync matches against — its job is sync and backup, not the mix.
- Drift — the camera and recorder clocks run at slightly different speeds over a long take. Fix by re-clapping/syncing in shorter chunks, or using gear whose clocks you trust (or timecode).
27.5 Selects and the string-out
Your media is in, organized, synced, and playing smoothly. You are, at last, ready to make decisions about the footage itself — and the first decision is not "what's my first cut?" It's "what's good?" You shot far more than you'll use (if you followed Chapter 20's coverage mindset, gloriously more), and the job now is to pan that river of footage for the gold. This is culling, and it produces your selects.
A select is simply a clip — or, more often, the good part of a clip — that you have marked as a keeper: the take where the answer landed, the B-roll shot that's actually in focus, the three seconds of a longer clip worth using. You review everything, and as you go you flag the keepers: a star, a colored marker, a favorite, or an in-and-out range that carves the good ten seconds out of a two-minute take. Then you gather those keepers into the Selects bin you built in §27.2. The outtakes don't get deleted — they stay filed, in case — but they get out of your way, so that from here on you're working with your best material, not wading through the misses. Editors have culled this way since the film days, when the good takes were literally "circled" on the camera report (the origin of "circle takes" — a callback to the slates and take logs of Chapter 18).
🔗 Connection. Selects and the string-out get their full, formal treatment in Chapter 30 (§30.2), where you'll build an entire interview-driven documentary from them using a paper edit — deciding the story on paper, from the transcript, before you touch a clip. This section gives you the working version you need right now to prep any edit; Chapter 30 turns it into a complete method for the talking-head piece. Treat what follows as the tool; Chapter 30 is the craft built on the tool.
Once your selects are pulled, you lay them out. A string-out is those selects placed end to end on the timeline in a rough, first-pass order — not cut to the frame, not polished, just in a line and in roughly the right sequence. It's the moment your organized bins become an actual sequence you can watch start to finish. A string-out of a documentary might be: the best answer to question one, then question two, then the strongest B-roll for each, then the ending soundbite — long, loose, and probably three times its final length. That's fine. Its only job is to get every good piece into one watchable strip so you can start shaping.
And shaping it is the next step, which brings us to the term this chapter owns and Chapter 26 set up. When you take that long, loose string-out and begin to order it deliberately — putting the beats in story sequence (Chapter 17's arc), cutting the obvious dead weight, watching it through — you arrive at your rough cut, also called the first cut: the first complete, watchable version of your edit, with every beat present and in order and roughly the right length, but not yet trimmed to the frame or polished. It's rough on purpose. It's allowed to be too long, to have clumsy cuts, to lack music. What makes it a first cut rather than a string-out is that it now has a shape — a beginning, middle, and end you could show someone (a brave someone) and they'd follow the story.
FIGURE 27.5 — The editorial ladder: where a "first cut" sits (and where you are now)
raw clips
│ (Chapter 26) assemble — pull the good pieces onto the timeline
▼
STRING-OUT ........ selects laid end to end, long and loose, rough order
│ order into story shape; cut the obvious fat; watch it through
▼
ROUGH CUT / FIRST CUT .... every beat in place, watchable, ~right length ◄ Ch.27 ends here
│ (Chapters 28–29) trim to the frame, J-/L-cuts, shape the rhythm
▼
FINE CUT ......... polished, precise, every cut deliberate (Chapter 29)
│ (Chapter 30) approve and freeze
▼
PICTURE LOCK ..... no more picture changes; hand off to color and sound
The ladder is worth memorizing because it tells you what "done" means at each rung and stops you from polishing too early. A classic beginner mistake is to fall in love with the first cut of the first scene — trimming it to perfection, adding music — while the rest of the film doesn't exist yet. Don't. Get the whole thing to a rough cut first, top to bottom, however clumsy, so you can see the shape of the entire piece before you sweat any single cut. You cannot judge the pace of a scene until you've felt the scene before and after it. The rough cut is where the film first exists as a whole; the fine cut (Chapter 29) is where you make it good.
✂️ In the Edit. The quality of your selects is set entirely by the quality of your coverage, which was set entirely on the shoot. If Chapter 20's coverage mindset was real — if you got the wide and the medium and the detail, multiple takes, the cutaway that hides a jump — then your Selects bin is deep and the string-out has options. If you shot thin (one take, one angle, no cutaways), your selects are whatever you happened to capture, and the edit will feel cornered no matter how skilled you are. Selecting is where you finally see, with total clarity, whether you shot for the edit. Note it honestly for next time; it's the most useful feedback a shoot can get.
🎬 On Set (at the desk): cull to a string-out. Take the organized, synced project from the §27.2 assignment and pull selects: watch every clip, flag the keepers, and range-select the good part of long takes. Gather them into a Selects bin, then drop them onto a timeline in rough story order — your string-out. Constraint: do not trim to the frame or add music; the string-out is allowed to be long and loose. Self-review: watch it through once, start to finish, and write a single sentence naming the story it's telling. If you can't, you either need better selects or a different order — which is exactly the work Chapters 29 and 30 will teach.
🔄 Check Your Eye. 1. What is a select, and where do your selects live in the project? 2. What separates a string-out from a rough cut? 3. Why should you get the whole piece to a rough cut before perfecting any one scene?
Check yourself
- A select is a clip (or the good portion of one) you've flagged as a keeper; selects are gathered into the Selects bin, out of the way of the outtakes.
- A string-out is selects laid end to end in rough order — no shape yet. A rough/first cut has been ordered into a story and trimmed of obvious fat, so it's watchable start to finish with a beginning, middle, and end.
- Because you can't judge a scene's pace or length until you've felt the scenes around it — the rough cut is the first time the whole piece exists, and only then can you see its overall shape.
27.6 A repeatable edit-prep workflow
Everything in this chapter is one pipeline, and the professional secret is that it's the same pipeline every time. You don't reinvent it per project; you run it as a checklist, in order, until it's muscle memory. Here is the whole thing on one card.
FIGURE 27.6 — The edit-prep pipeline: from card to string-out (run it in order, every time)
1 OFFLOAD cards ─► verified copy ─► TWO drives (never work off the card)
2 BACK UP confirm two good copies ─► only then wipe cards (full 3-2-1: Chapter 37)
3 INGEST import into the project; link to the offloaded files (don't move them again)
4 ORGANIZE build folders; mirror them as bins; label; by scene / day / camera
5 SYNC marry double-system audio to picture (clap ► waveform ► timecode)
6 PROXY if playback stutters, generate proxies; edit on them, deliver on originals
7 SELECTS watch EVERYTHING; flag the keepers into a Selects bin
8 STRING-OUT lay selects end to end in rough order ─► shape toward a first cut
───────────────────────────────────────────────────────────────────────────────────►
...and now, and only now, you are ready to EDIT for real (Chapters 28–30).
A few notes on the order, because the order encodes hard-won lessons. Offload and back up come first and are non-negotiable — nothing else matters if the footage dies. Organize before you sync, so the pieces you're syncing already have homes. Sync before selects, so that when you flag a keeper it comes with its good sound attached. Proxies can slot in wherever your playback demands — if the footage won't even play well enough to review, generate proxies right after ingest (step 3.5); if it plays fine for reviewing but bogs down once the timeline gets busy, make them later. And selects before string-out, always — you can't lay out keepers you haven't chosen.
There's one honest exception worth naming: if your footage is so heavy it stutters during the review in step 7, do the proxies earlier, right after organizing. The pipeline is a default, not a straitjacket; the principles (safe first, organized before synced, chosen before strung out) are what's fixed. Internalize those and you can adapt the order to any shoot. And note where this pipeline sits in the bigger picture: it gets you into an edit safely, while Chapter 37 (§37.1, §37.5) closes the loop on the way out — archiving a finished project so you can reopen it cleanly in a year. Together they are the full life-cycle of a project's media, and a pro runs both without thinking.
💡 Why It Works: the checklist beats the memory. Airline pilots, surgeons, and senior editors have the same insight: for a repeated, high-stakes, easy-to-fumble sequence, a written checklist you run every time beats relying on a smart person's memory most times. The failures in this chapter — the wiped card, the unverified copy, the exported proxy, the missing scratch track — are not failures of skill. They are failures of a tired person skipping a step they usually remember. A checklist removes "usually." Print FIGURE 27.6, tape it above your desk, and run it top to bottom on your next ingest. The goal is that safe, organized, synced, and strung-out becomes something you do without deciding to — which frees your whole mind for the only part that's actually art: the cut.
🔄 Check Your Eye. 1. Why must "organize" come before "sync" in the pipeline? 2. When does it make sense to move the proxy step earlier? 3. What's the argument for running this as a written checklist rather than from memory?
Check yourself
- So the clips you're syncing already have homes (bins); you're not creating synced clips into chaos, and the synced result files itself where it belongs.
- When the footage is so heavy it stutters even during review — generate proxies right after ingest so you can watch and select in comfort.
- The pipeline's failures come from tired people skipping a usually-remembered step, not from lack of skill; a checklist you run every time removes the "usually" and protects the footage automatically.
Production Checkpoint
Project 2 — set up your documentary edit. You shot your documentary short across Chapters 19–21: an interview (with double-system audio, per Chapter 15's discipline), a B-roll list, maybe a voiceover option and some archival. It's time to turn that pile of cards and files into an edit you can actually cut — by running this entire chapter's pipeline on it, once, in order.
Do all six steps:
- Ingest and back up. Offload every card — camera, audio recorder, phone — with verified copies to two separate drives. Confirm both copies before you consider a card reusable. (Preview only of Chapter 37's full 3-2-1 rule — two verified copies is today's floor.)
- Organize. Build the FIGURE 27.2 folder structure on disk, import the media into a fresh project, and mirror the structure as bins — footage, audio, room tone, music, all with a home.
- Sync. Marry your double-system audio to the interview picture — by the clap if you slated it, by waveform auto-sync if your editor offers it, or by hand if you must. Confirm lips and sound match.
- Proxy. If your interview and B-roll stutter on playback, generate proxies and switch the project to edit on them (you'll relink to originals at export in Chapter 36).
- Selects. Watch everything. Flag the keepers — the answers that land, the B-roll that's sharp, the moments with life — into a Selects bin.
- String-out. Lay your selects end to end in rough story order. Don't trim to the frame; don't add music. Just get the whole story into one watchable strip.
Why this matters: you now have a synced, organized, proxied project with a string-out — which is exactly the launch pad Chapters 28–30 need. Every hour you spent here buys back three in the edit to come. You have not made a single real cut yet, and you are already ahead of most people who have.
Summary
- Ingest = getting media off the cards and into the project, safely. It has two halves: the offload (the physical copy) and the import (linking media into the editor). The offload comes first, always.
- The cardinal safety rule: never work off a card, and never wipe a card until two verified copies exist elsewhere. A card is transport, not storage. (Full backup discipline — the 3-2-1 rule — is Chapter 37.)
- Media management is the discipline of keeping every asset organized, findable, and safe. It has two mirrored layers: folders on disk and bins in the project.
- Proxies are light stand-in copies you edit on for smooth playback and swap back to originals to export. Cut on the copy; deliver from the master. Never delete originals.
- Sync in the edit marries double-system audio to picture. Three methods, and when to use them:
| Method | How | Best when |
|---|---|---|
| Clap / slate | Align the shared spike in scratch + recorder audio by hand | Always available; any editor; the on-set clap makes it seconds |
| Waveform auto-sync | The software matches waveforms and aligns automatically | You have scratch audio; many clips to sync at once |
| Timecode | Camera + recorder share a clock; clips snap by TC | Bigger shoots with TC gear; fastest and most reliable |
- Selects (keepers, flagged and binned) → string-out (selects end to end, rough order) → rough cut / first cut (ordered into a watchable story shape) → fine cut (Chapter 29) → lock. Get to a rough cut whole before perfecting any scene.
- The edit-prep pipeline, in order: offload → back up → ingest → organize → sync → proxy → selects → string-out. Run it as a checklist every time; the failures it prevents are failures of memory, not skill.
- The throughline of the whole chapter: you cannot cut what you cannot find. Organization is not clerical — on a deadline, it is what protects your creativity.
Spaced Review
Retrieval from earlier chapters — answer before you check, then look back if you're unsure.
- (Chapter 26) What is the assembly, and how does it relate to the string-out and rough cut you built in this chapter?
- (Chapter 26) Why is a hard cut the default edit, and a dissolve or transition the exception you have to justify?
- (Chapter 3) A clip is stored in a Long-GOP H.265 codec. Explain, using Chapter 3's ideas, why it's harder for your computer to play than an all-intra codec — and connect that to why you'd make a proxy.
- (Chapter 3) What three recording choices from Chapter 3 most affect how "heavy" (hard to play, big on disk) a clip is?
Check yourself
1. The *assembly* (Chapter 26) is the raw first pass of dropping selected clips onto the timeline. The *string-out* is essentially that — selects laid end to end, loose. It becomes a *rough/first cut* once you order it into a deliberate story shape and cut the obvious fat. 2. A hard cut is invisible and instantaneous — the natural grammar of the edit, mimicking how attention jumps. A dissolve/transition draws attention to *itself* and implies a passage of time or a change of place, so it needs a reason; used by default it reads as amateurish (Chapter 26, §26.3). 3. Long-GOP stores mostly the *differences* between frames, so displaying one frame means decoding a whole group of surrounding frames — heavy CPU work, especially when scrubbing. An all-intra codec stores every frame whole (easy to fetch any single frame). A proxy is a small all-intra copy, trading disk space for effortless playback. 4. **Resolution** (pixel count — 1080p vs 4K vs 6K), **bitrate** (data per second), and the **codec / compression** (Long-GOP vs all-intra, and how aggressively it compresses). (Chapter 3, §3.1–§3.2.)What's Next
Your media is safe, organized, synced, light enough to play, and strung out into a rough cut. In other words: the hard part nobody talks about is done, and the fun part everybody wants is finally clear ahead of you. Chapter 28 takes you into the grammar of the edit itself — the J-cuts and L-cuts that let sound lead the picture, cutting on action, the match cut, and the montage — the moves that turn a serviceable string of clips into editing that's invisible and powerful. You've been preparing to cut for a whole chapter. Now, with a project that finally gets out of your way, you get to actually do it. Bring your string-out; we're about to make it sing.