Appendix F: Editing & Post Reference

The screenshot-and-scan card for the edit — Chapters 26 through 30 compressed into tables, ladders, and diagrams you can keep open on a second monitor while you cut. It assumes those chapters taught you the why; this is the what, the order, and the fix. The single idea beneath all of it: the cut is a thought, not a transition (Chapter 26) — you are not gluing clips together, you are directing a stranger's attention and assembling a sequence of feelings in their chest (Chapters 28–29). Everything below serves that.

The whole appendix in one line. Get the media in safely and organized — offload twice, bin, sync, proxy — then pull selects, string them out, and shape rough cut → fine cut → lock. Find the story in the words or the shape before you polish. Default to the hard cut; make a cut invisible by carrying something continuous (sound, motion, form, or idea) across it; and when priorities collide, protect emotion and sacrifice continuity (Murch's Rule of Six). Cut the boring parts ruthlessly, then hold the one that matters.

Cross-references point at the chapters that teach each move in full: the timeline and the cut (26), media management (27), the grammar of cutting (28), story and emotion (29), the interview edit (30). Finishing — color, sound, titles, captions, delivery — lives in Chapters 31–36 and in Appendices D (audio) and E (the NLE Translation Guide, for where each button lives in your software).

F.1 The cut-types glossary

Every cut in the book, in one table: what it is, and when to reach for it. The first column is the name; the last is where it is taught in full. The hard cut is your default — you will use it more than all the others combined (Chapter 26).

The split edit, in one line. A J-cut and an L-cut are the same move — a staggered audio edit — pointed in opposite directions. Both are split edits: cuts where picture and sound do not change on the same frame. Unlink the audio from the video, trim one past the other, and you have one. Beginners cut every seam as a straight cut, and it clicks like a slideshow; staggering the audio at most seams is the single most valuable upgrade in the edit (Chapter 28), and it costs nothing but attention.

The invisible-cut principle (Chapter 28). A cut disappears when something continuous carries the eye or ear across it, and it announces itself when nothing does. Ask at every seam: what carries the viewer across? Sound (J-/L-cut, a bed), motion (cut on action), form (graphic match), idea (match cut), or a held rhythm (montage, cross-cut). If nothing bridges it, either find a bridge or expose the cut on purpose. The amateur's error was never that their cuts showed — it was that their cuts showed by accident.

F.2 The edit-timeline legend

Every timeline diagram in this book uses the same notation. Learn it once here and you can read any figure in Chapters 26–30 at a glance.

LEGEND — how to read an edit-timeline diagram
  V2, V1    video tracks; a HIGHER track (V2) sits ON TOP of / covers the lower (V1)
  A1, A2    audio tracks, BELOW the video; A1 usually sync dialogue, A2 a bed
  [ ..... ] a clip  (its width ≈ its screen duration)
  #####     a clip's picture or sync sound, running
  .......   a CONTINUOUS bed (room tone / ambience / music) that never cuts
  |         a CUT — one frame to the next;  ][  = two clips meeting on a cut
  ^         an annotation pointer to the exact moment above it
  time →    left = earlier, right = later

The two most useful cuts in the book, drawn. In a J-cut, the incoming sound reaches back under the outgoing picture; in an L-cut, the outgoing sound extends forward under the incoming picture.

A J-CUT — the NEXT shot's sound arrives BEFORE its picture
  V1  [ street: person walking ############## ][ café interior: they enter ###### ]
  A1  [ street ambience ############ ]
  A2  [ café clatter + espresso hiss begins here ............................... ]
                                     ^ SOUND of the café leads    ^ PICTURE cuts to café
   We hear the café before we see it; the picture change lands as a confirmation, not a surprise.

An L-CUT — THIS shot's sound continues after its picture is gone (the engine of B-roll)
  V2  .................[ B-roll: hands in flour ############################# ]
  V1  [ interview: face #######]|(picture cuts up to the B-roll on V2)
  A1  [ interview sync voice — UNBROKEN ....................................... ]
  A2  [ room tone / soft music bed ............................................ ]
                              ^ PICTURE cuts to the hands        ^ voice finally ends
   The voice keeps running while the picture roams. Hide the cut under the continuing voice
   and it disappears — which is the entire reason B-roll exists (Chapter 28).

WHAT CARRIES THE VIEWER ACROSS A CUT? (the invisible-cut checklist, Chapter 28)
  Want it HIDDEN — find a bridge:
   ├─ SOUND continues?   → J-cut / L-cut / a music or ambience bed
   ├─ MOTION continues?  → cut on the action, mid-movement, SAME screen direction
   ├─ FORM continues?    → graphic match (shared shape / line / mass)
   ├─ IDEA continues?    → match cut (two things revealed as one thought)
   └─ RHYTHM is held?    → montage on a music bed / cross-cut on a shaped tempo
  Nothing continues? → the cut will show. Then choose: hide it, or EXPOSE it on purpose.

F.3 The editing order of operations

The whole path from a pile of cards to a finished, frozen cut, in order. Steps 1–4 are edit prep (Chapter 27 — get in safely and fast); steps 5–9 are the edit (Chapters 26, 28–30 — make it a story). Run prep as a checklist every time; the failures it prevents are failures of a tired person skipping a step, not of skill.

THE MASTER PIPELINE — from card to lock (run it in order)

  ── EDIT PREP (Chapter 27) ────────────────────────────────────────────────
  1  INGEST      offload cards → VERIFIED copy → TWO drives, then import
  2  ORGANIZE    build numbered folders; mirror as bins; label by scene / day / cam
  3  SYNC        marry double-system audio to picture (clap ► waveform ► timecode)
  4  PROXY       if playback stutters, generate proxies; edit on them, deliver on originals
  ── THE EDIT (Chapters 26, 28–30) ─────────────────────────────────────────
  5  SELECTS     watch EVERYTHING; flag the keepers into a Selects bin
  6  STRING-OUT  selects laid end to end in rough story order  (= "the assembly")
  7  ROUGH CUT   order into a story SHAPE; cut the obvious fat; watch it whole   (= "first cut")
  8  FINE CUT    trim to the frame — enter late, leave early; PROTECT the emotion
  9  LOCK        freeze the picture → hand off to color / sound / titles / captions
  ──────────────────────────────────────────────────────────────────────────►
     ...then Part VII (Ch.31–36) makes the locked film look and sound finished.

"String-out" and "the assembly" are the same rung under two names. The book calls the raw first pass the assembly for a scripted scene (Chapter 26) and a string-out when it is built from interview selects (Chapter 27). Both mean the same thing: keeper material laid end to end, rough, so you can watch the whole and judge its shape. Get to that whole before you perfect any single scene — you cannot judge a scene's pace until you have felt the scenes around it.

F.4 Media management & proxies — the safety checklist

The unglamorous discipline that decides whether an edit flows or fights you (Chapter 27). Full backup theory — the 3-2-1 rule, naming that survives, archiving — is Chapter 37.

The offload (do this before you edit a single frame):

[ ] Copy each card → a VERIFIED (checksum) copy, not a naked drag  (catches silent corruption)
[ ] Land it on TWO separate drives — a working drive AND a backup drive
[ ] Offload the audio recorder's card with the SAME paranoia as the camera
[ ] Only NOW format a card for reuse — never before two good copies exist
[ ] Import points the project AT the offloaded files — don't move or rename them after

The one rule that separates people who lose footage from people who don't: never work off the card, and never wipe a card until its contents live in at least two other verified places. A card is transport, not storage — the only place a shoot can die. If it can nearly happen to a major studio (it has), it can happen to your documentary.

Folders on disk, mirrored as bins in the project (numbered so they sort in workflow order, not alphabetically):

  ON DISK                          IN THE PROJECT (bins)
  PROJECT-NAME/                     📁 PROJECT
  ├── 01_FOOTAGE/  (Cam A/B, B-roll)   ├── 01 FOOTAGE
  ├── 02_AUDIO/    (Recorder, Room tone)├── 02 AUDIO
  ├── 03_MUSIC-SFX/                     ├── 03 Music & SFX
  ├── 04_GRAPHICS-STILLS/               ├── 04 Graphics & stills
  ├── 05_PROJECT-FILES/                 ├── 05 SELECTS     ◄ points at clips filed above
  ├── 06_PROXIES/                       ├── 06 TIMELINES   ◄ String-out / Rough / Fine
  ├── 07_EXPORTS/                       │
  └── 08_DOCS/  (brief, script, releases)
   Disk = WHERE the media lives (once). Bins = HOW you find it (from as many views as useful).

Proxies — the round-trip rule: generate proxies if playback stutters; edit on the light copies; switch back to originals to export.

[ ] Playback stutters / lags / drops frames?  → generate proxies (or "optimized media")
[ ] Proxy ≈ half-res, low-bitrate, an easy all-intra codec — light enough to scrub instantly
[ ] EDIT on the proxies; keep them in 06_PROXIES/ (disposable, regenerable)
[ ] Before the FINAL render: confirm the project is on ORIGINALS  (first item on the export checklist)
[ ] NEVER delete the originals — if you delete anything, delete the proxies

Phone-first surprise: modern phones shoot highly compressed Long-GOP HEVC/H.265, which is genuinely harder to decode than footage from a "real" camera — so phone footage often stutters more on a laptop. If your phone edit feels sluggish, proxies are the free fix (Chapter 27).

F.5 The paper / radio edit — the interview quick-method

The method for turning twelve rambling minutes into a tight three-minute story (Chapter 30). The raw interview is not a rough draft of your film — it is the quarry. You build the film out of it, one soundbite at a time, in an order the subject never spoke. The radical, counterintuitive core: cut the words first, with your eyes closed to the picture.

THE INTERVIEW-EDIT PIPELINE — from transcript to lock (audio & story FIRST; picture enters at step 5)

  1  RADIO EDIT      transcript → highlight KEEPERS, cut the mush → order the cards on PAPER
        │             (eyes closed to picture — the story must work as WORDS first)
  2  SELECTS         pull each keeper soundbite, in/out ON THE BREATH, labeled by CONTENT
  3  RADIO STRING-OUT selects end to end, in STORY order → CLOSE YOUR EYES: does it hold as radio?
  4  SPINE           commit the order (the ARC, not chronology); cut the questions out; tighten honestly
        │             (done as audio — deliberately BROKEN as picture: it jumps at every seam)
  5  B-ROLL LAYER    cover EVERY seam with MATCHED B-roll on the track above; wild sound beneath
        │             (voice leads, unbroken; picture proves the claim — a B-roll-driven narrative)
  6  PACE            reveal the person → roam the B-roll → RETURN to the face for the emotional lines
  7  FINE CUT        enter late / leave early on words AND pictures; protect the living silences
        │             → run BOTH quality passes:  eyes-closed (radio)  +  sound-off (mute)
  8  LOCK            freeze the picture → finishing begins
  ────────────────────────────────────────────────────────────────────────────────────────►

When a locked interview piece feels boring and you can't say why, walk back down the ladder. Nine times out of ten a dull piece has fine B-roll and a broken spine — the fix is back to the words and their order (step 1), not more pretty pictures.

F.6 Murch's Rule of Six

The compass for why any cut works or fails, from editor Walter Murch's In the Blink of an Eye (Chapter 29). When a cut cannot satisfy everything at once — and it almost never can — you protect the top of the list and sacrifice from the bottom.

  EMOTION alone (51%)  >  the other five COMBINED (49%).      Emotion + Story = 74%.
  Sacrifice from the BOTTOM, never the top. A crossed line the head forgives a beat late;
  a wrong emotional note the body feels instantly and never quite forgives.

How to use it in the chair. When you're agonizing over a cut, the first two questions answer most of it: What should they feel here? What do they need to know or feel next? Get those right and you can be a little wrong about rhythm — and quite wrong about the bottom three — and still have a cut that works. Get emotion wrong and no amount of clean continuity will save it. The emotional cut is a decision about the audience, not the footage: cut in on the feeling, hold past comfort, and leave before it resolves.

F.7 Diagnose a bad edit — symptom → cause → fix

The timeline diagnostic. Most "something's off and I can't say what" problems are on this table.

F.8 The reference card — decision rules to keep

The rules worth tattooing on your trim hand:

• The cut is a thought, not a transition — ask "what should they understand, and when?", not "how do I join these clips?"
• Default to the hard cut (~99% of edits); a transition must be motivated — finish "I'm dissolving here because ______."
• You shoot for the edit — dissolves need handles, cuts-on-action need overlap, L-cuts need room tone and B-roll, captured on set or not at all.
• Trim to the meat; enter late, leave early — first assemblies are almost always ~20% too slow.
• Tighten without gutting — the scene gets shorter everywhere except the moment that matters. Cut dead air; protect living silence.
• Ask "what carries the viewer across?" at every seam — sound, motion, form, idea, or rhythm hides a cut; nothing bridging it exposes one.
• Emotion outranks continuity — when priorities collide, protect the feeling and give up the crossed line.
• Cut the words first in any interview; kill your darlings (save them for the reel, Ch.39); lock the picture, then finish.

The one sentence: get the media in safe and organized, string your selects into a whole you can watch, find the story before you polish, then rough cut → fine cut → lock — defaulting to the invisible hard cut, carrying something continuous across every seam, and protecting the feeling above all — because the cut is where meaning is made.