Case Study 2: The Vertical Cut-Down — Turning a Horizontal Talking-Head into a Scroll-Stopping Teaser

Where Case Study 1 took apart a famous technique, this one builds a deliverable from scratch: we take a finished horizontal talking-head — exactly like your Project 1 — and turn it into a vertical teaser that survives a feed. Follow along, and better, do it to your own footage as you read. This is a repurpose-and-shoot walkthrough: reframing, a small native-vertical pickup shoot, captions, and a two-second hook, walked through in real time on the simplest setup in the book.

The brief

We have a 60-second horizontal (16:9) talking-head already shot and cut — a person to camera, clean window light, a lav mic, one clear message. (If you did Project 1, this is your video.) The ask, which lands on every working videographer's desk now: "Can we get a vertical version for Reels, TikTok, and Shorts?"

The constraint is real and instructive. The source was composed for 16:9 — subject on a side third, horizontal breathing room, a slow pace with a gentle spoken intro. None of that is wrong; it is simply wrong for the feed. Our job is not to shrink the video but to re-conceive it for a vertical, muted, scrolling audience while keeping its story intact. One person, a phone or the original camera, free editing software, and a couple of hours. No reshoot of the whole thing — a repurpose, plus one small pickup.

We choose the talking-head because it is the hardest common case to reframe (a single subject, often composed to one side, with information in the audio) and the most requested. If you can crack the talking-head cut-down, the rest of repurposing is easier.

Assessing the source: what we actually have

Before touching a crop tool, we watch the horizontal cut honestly and inventory it, because you repurpose for the edit — you decide the vertical plan before you start dragging keyframes.

FIGURE CS2.1 — Inventory of the horizontal source (60s talking-head)
  0:00–0:08  Gentle intro: "Hi, I'm ___, and I want to talk about ___."   ← throat-clearing; no feed hook
  0:08–0:20  The core claim / most surprising line                         ← THIS is our real hook
  0:20–0:45  The explanation, three points, calm pace                      ← trim hard for vertical
  0:45–0:58  The payoff / call to action                                   ← keep, tighten
  0:58–1:00  Sign-off                                                      ← cut for a teaser
  FRAMING    Subject on the RIGHT third, looking camera-left; window key camera-left.
  AUDIO      Clean lav; good room tone; fully intelligible.
  NO CAPTIONS. NO on-screen text. Composed only for 16:9.

Three findings drive everything. First, our best two seconds are buried at 0:08–0:20, not at the start — the surprising core line. That is our hook, and it needs to move to the front (Case Study 1's front-loading; §23.2). Second, the subject is composed hard on the right third — a naive center-crop to 9:16 would cut half their face off. Third, there are no captions and no text, so as-is it fails the sound-off test (§23.3) completely. The horizontal cut is good; it is just built for a different viewer.

Gear and settings

Repurposing is mostly an edit task, but we will also shoot one small native-vertical pickup (the hook), so we need a minimal capture kit too.

⚙️ Settings Box: the vertical cut-down (edit + one pickup shot).

Setting	Value	Why
Project / timeline ratio	9:16 (1080 × 1920)	Build in the target frame so you compose the crop against real safe zones.
Source handling	Keep original res; scale the crop up to 100%, avoid past it	Don't upscale past native or the crop softens.
Pickup shot orientation	Vertical, phone held upright	Shoot the hook native-vertical so it's sharp full-frame, not cropped.
Pickup resolution / fps	1080p or 4K, 30 fps	Matches the timeline; 4K gives crop room (§23.5).
Pickup light	Same window key, subject facing it	Match the source's look so the pickup cuts in seamlessly (Chapter 13).
Pickup audio	Lav or phone close-mic + room tone	Sound is half the picture even muted; match the source (Chapter 15).
Captions	Burned-in, large, high-contrast, safe-zone	The muted majority reads them (§23.3, ♿).
Loudness	~ −14 LUFS	Platform normalizes anyway; mix clean (Chapter 33).

The whole job needs the original footage, a free editor (DaVinci Resolve is our reference; Appendix E translates the crop and caption tools to any app), and a phone for the pickup. That's it.

The kit: the original talking-head files, a computer with a free editor, and a phone by a window for one pickup shot. A complete cut-down kit — no reshoot, no budget.

The plan: a reframe strategy for every beat

We decide, beat by beat, how each shot becomes vertical — the same way you'd plan coverage before a shoot (Chapter 20). This is the repurpose equivalent of a shot list.

FIGURE CS2.2 — The vertical reframe plan
  BEAT              SOURCE                    VERTICAL METHOD
  ──────────────    ───────────────────────   ────────────────────────────────────────
  Hook (0:00–0:02)  (does not exist yet)      SHOOT a native-vertical hook pickup
  Core line         0:08–0:20, subj. right    CROP-AND-REFRAME: slide crop right to
                                              keep the face centered in 9:16
  Explanation       0:20–0:45, trimmed        CROP-AND-REFRAME; add captions; cut the
                                              three points to two, tighten pauses
  A demo/detail?    (none in source)          Optional: shoot a vertical B-roll insert
  Payoff / CTA      0:45–0:58                  CROP-AND-REFRAME; caption; end on a loop

The key decisions are already made on paper (fix it in pre, not in post — even for a repurpose): we will front-load a shot we haven't filmed yet, crop-and-reframe the talking-head to keep the face centered, and cut the explanation from three points to two so the teaser stays short. Now we execute.

The pickup shoot: filming a native-vertical hook

The source has no feed-worthy hook, so rather than crop a weak one out of the intro, we shoot a strong one native-vertical. Two minutes of setup by the same window.

FIGURE CS2.3 — Vertical hook setup (top-down view)
        [ WINDOW ] ☀  soft key (same as the source, so it matches)
             |
             ▼
           ( S ) subject faces the window, square-ish to camera
             |
           [CAM]  phone upright, ~chest height, framing 9:16;
                  face centered, eyes on the upper-third line

  Same window, same side, same look as the original 16:9 shoot — so the vertical
  pickup intercuts invisibly with the reframed source. Match the light, match the world.

We frame it for the vertical safe zone from the start — this is the whole advantage of shooting native. Here is the overlay we compose against:

FIGURE CS2.4 — The hook, composed inside the 9:16 safe zone
   ┌───────────────────────┐  ← keep title out of the top strip
   │                       │
   │       ( FACE )        │   Face centered, eyes on upper third,
   │      eyes upper 1/3   │   fully inside the action-safe center
   │                       │   column.
   │  ┌─────────────────┐  │
   │  │ TEXT HOOK HERE  │  │   On-screen promise sits center-safe,
   │  └─────────────────┘  │   above the platform's caption strip.
   │·······················│
   │ @username  ▬▬▬▬▬▬▬▬▬▬ │  ← leave this bottom strip for the platform
   └───────────────────────┘

Now we shoot the hook itself. We record the subject delivering the surprising core line cold — no intro, straight into the promise — three or four times, so the edit has choices (§23.2's On Set box, done for real).

FIGURE CS2.5 — The hook pickup, take 3 (the keeper)   [constructed teaching example]
  THE FRAME    9:16 vertical, medium-tight. The subject is centered, eyes on the upper third, looking
               just off-lens. Bold on-screen text sits center-safe: "The one lighting mistake that
               makes phone video look cheap."
  THE MOVE     A hair of handheld life — not locked-off-formal. It reads "now, real," which the feed likes.
  THE LIGHT    The same window key camera-left as the source, so this pickup will intercut invisibly with
               the reframed 16:9 footage. Matching the light is what lets a pickup disappear.
  THE SOUND    First words ARE the hook: "Stop lighting your face from above." No "hi," no setup. For the
               muted viewer, the on-screen text carries the same promise.
  THE EFFECT   In under two seconds the viewer has a clear, slightly contrarian promise and an open loop
               (what's the mistake? what should I do instead?). It works muted and unmuted.
  THE LESSON   When the source has no hook, shoot one — native-vertical, matched to the source's light —
               rather than cropping a weak opening out of an intro that was never built to grab.

Adjustment: take 1 had the subject start with a reflexive "So, um—"; we reset and had them start on the promise. Take 2 drifted the eyes to the lens and it felt like an ad; take 3 kept the off-lens eyeline from the source and matched perfectly. Film, watch, adjust, re-film — the same loop as any shoot (Chapter 1).

⚠️ Common Mistake: a mismatched pickup that won't cut in. The temptation is to grab the vertical hook quickly, anywhere, in different light. Then it clashes with the reframed source — different color, different direction of light, different room — and the cut announces itself as two separate shoots stapled together. The fix is discipline: match the light, the eyeline, the wardrobe, and the background of the original before you roll the pickup. A pickup's whole job is to be invisible; matching is what makes it disappear.

The repurpose, phase by phase

Now we build the vertical timeline. We work in the target 9:16 frame so every crop is judged against the real safe zones.

Phase 1 — front-load the hook

We drop the native-vertical hook pickup (CS2.5) at 0:00. The video now opens on the promise, exactly as Case Study 1 prescribes. Everything from the source's original 0:00–0:08 intro — the "hi, I'm ___" — is deleted without mercy. It was an exit ramp; it's gone.

Phase 2 — crop-and-reframe the talking-head

Now the source footage. The subject sits hard on the right third of the 16:9 frame, so a centered crop would slice them. We slide the 9:16 crop window to the right to center the face.

FIGURE CS2.6 — Crop-and-reframe: sliding the window to the subject
  SOURCE 16:9 (subject on right third)          9:16 CROP (window slid right)
  ┌───────────────────────────────┐             ┌───────────┐
  │                        ┌────┐  │             │  ┌────┐   │
  │                        │ S  │  │   ──────►   │  │ S  │   │  face now centered,
  │        (empty          │    │  │  slide the  │  │    │   │  eyes on upper third,
  │         left side)     └────┘  │  crop → →    │  └────┘   │  inside the safe zone
  └───────────────────────────────┘             └───────────┘
   The empty left side (nose room for 16:9) is simply cropped away; the crop window
   is positioned so the subject lands center-safe in vertical. One static crop per shot,
   set once — no keyframing needed because the subject doesn't move.

Because a seated talking-head barely moves, we don't need to keyframe the crop — we set one static crop position per shot that centers the face, and it holds. (If the subject had moved across frame, we'd keyframe the crop to follow them, or cut between two crop positions, as in §23.5.)

Adjustment: our first crop centered the face perfectly but put the eyes dead-center, which felt low and cramped. We nudged the crop down slightly so the eyes ride the upper third (Chapter 6's headroom rule, alive and well in vertical). Same footage, better vertical frame.

Phase 3 — tighten the story for the feed

The source explained three points at a calm pace. A teaser can't afford three; we cut to the two strongest, and we tighten every pause between sentences (§23.3 — cut the dead air). What was 25 seconds of explanation becomes 12 tight seconds. We're not dumbing it down; we're respecting a scrolling viewer's time — and if they want the full version, that's what the link in the bio is for.

✂️ In the Edit: the teaser sells the full piece. A vertical teaser is not the whole story crammed small — it is a trailer for the full horizontal video. Its job is to hook, deliver one satisfying beat, and leave the viewer wanting the rest (which lives on your channel or site). So we keep the single most valuable point intact and let the teaser point at the full piece. This is the repurpose mindset: one shoot, many deliverables, each aimed at where it lives. (You'll formalize multi-format delivery in Chapter 36.)

Phase 4 — captions and the sound-off pass

Now the accessibility spine of the whole job. We add burned-in captions across the entire teaser: large, high-contrast, with a stroke so they survive the background, placed in the center-safe band (never the bottom strip where the platform's UI sits).

We start from the editor's automatic captions (a fast first pass) and then proofread every word — the auto-caption turned the subject's name into nonsense and mistimed two lines. Corrected, the captions now carry the full message to a muted viewer.

♿ Accessibility & Inclusion: the muted-comprehension test. Before we call captions done, we run the test that matters: watch the whole teaser muted, as a stranger would, and confirm the story lands from picture and text alone. Our first pass failed — a key instruction was spoken but never captioned, so muted it made no sense. We captioned it and re-tested. This single test — can a muted stranger follow it? — is the difference between a video that reaches its audience and one that scrolls past most of them. Captions here are not decoration; they are how the majority of people will experience the entire video.

Phase 5 — the loop and the outro

We cut the source's sign-off ("thanks for watching") — an exit ramp and a dead end. Instead, we end the teaser on a beat that flows back toward the opening promise, so it loops cleanly (§23.3): the last frame's framing and the subject's look match the first, so on repeat there's no visible seam. A clean loop quietly buys watch time — the platform counts the replay.

The edit pass: the vertical timeline

Here is the finished 22-second teaser as a timeline. Notice the structure: hook first, one strong beat, a loop close, captions and a music bed running throughout.

FIGURE CS2.7 — The 22-second vertical teaser (final assembly)
  V2  [ CAPTIONS: big, center-safe, synced, proofread ..................................... ]
  V1  [ HOOK pickup ][ reframed core line ][ point 1 ][ point 2 ][ payoff → loops to hook ]
  A1  [ pickup sync ][ source sync audio ................................................. ]
  A2  [ music bed (licensed) ............................................................. ]
       0s        2s              7s        12s        17s                              22s

  • The native-vertical HOOK (V1, 0–2s) front-loads the promise; the source's intro is gone.
  • Reframed 16:9 footage (V1, 2s→) is cropped so the face stays center-safe throughout.
  • CAPTIONS (V2) run the entire length — the muted majority's version of the audio.
  • A licensed music bed (A2) runs unbroken so the picture cuts cleanly (Chapter 33; license per Chapter 38).
  • The payoff is cut to flow back into the hook — a clean loop.

The key cut decisions, named:

• The hard cut from the vertical pickup into the reframed source is invisible because we matched the light and eyeline on the pickup (Phase 1). Matching on the shoot is what let the edit disappear.
• Every crop position was set to keep the face center-safe (Phase 2) — the single most important repeated decision in the whole cut-down.
• The captions on V2 are the deliverable's accessibility and its reach; without them the teaser fails muted, which is most of the audience.
• The loop replaces the sign-off, converting a dead end into replays.

Delivering the cut-down to every platform

One vertical teaser is rarely the end of the job — the client wanted it "for Reels, TikTok, and Shorts," which is three destinations. Here the repurposing craft meets the delivery craft (Chapter 36), and the good news is that our 9:16 teaser is already built for all three: they share the vertical standard. But a few small, platform-specific decisions separate a lazy "post the same file everywhere" from a considered delivery.

FIGURE CS2.8 — One vertical teaser, three destinations (delivery notes)
  DESTINATION   ASPECT   WHAT WE KEEP / ADJUST
  ───────────   ──────   ─────────────────────────────────────────────────────────
  TikTok        9:16     As built. Consider native/trending audio for reach; keep
                         captions clear of the bottom UI (heavy there).
  Reels         9:16     As built. Instagram's UI also crowds the bottom-right; our
                         center-safe captions already survive it.
  Shorts        9:16     As built. Watch the title/end — Shorts UI differs slightly;
                         our safe-zone margins cover it.
  (Feed post)   4:5      OPTIONAL alt: a 4:5 version for the main feed grid, so the
                         piece also lives outside the short-form tab.

The workflow lesson is keep a master. We export one clean 9:16 master at full quality (1080 × 1920, high bitrate, captions burned in, ~ −14 LUFS), and that single file serves all three short-form destinations. If a platform later wants something different — a 4:5 for the feed, a longer cut, a version without burned captions so its own auto-captions can run — we go back to the timeline, not to a re-compressed upload. Never repurpose from a file that's already been through a platform's transcoder; always go back to the master or the project. (This is delivery discipline you will formalize in Chapter 36, §36.4, "masters vs. deliverables.")

One more delivery habit worth building now: a caption file alongside the burned-in captions. Our teaser has open captions burned in for the muted feed, which is right — but exporting a separate caption sidecar (an .srt, say) as well lets a platform layer its own selectable captions on top for viewers who need a different language or a screen reader's cooperation. It costs a minute and widens who can use the video. Doing both — burned-in for the feed, a caption file for flexibility — is the belt-and-suspenders move a considerate professional makes, and it's the same instinct Chapter 36 (§36.5) formalizes at export. Accessibility is never one box you tick; it's a set of small, cheap habits that compound into a video everyone can actually watch.

🔗 Connection. The full export-and-delivery craft — codecs, bitrates, keeping a master, and the current per-platform spec tables — is Chapter 36 (§36.3–36.4) and Appendix H. What you did here is the creative half of delivery (reframe, hook, caption); Chapter 36 is the technical half (the export dialog and the specs). Together they're how one shoot becomes many deliverables.

What we'd do differently (and why that's the point)

Watching the finished teaser honestly, the next lessons surface — which is exactly how the loop of getting better works.

• We could have shot a vertical B-roll insert. The teaser is all talking-head; a single native-vertical detail shot (the lighting mistake demonstrated on a real face) would have added a visual beat and hidden a cut. We had the window set up; we could have grabbed it in two minutes. Shoot the insert while the light's up — coverage is cheapest on the day (Chapter 20).
• The music bed is generic. It works, but trending/native audio often earns more reach on a feed. Next time: check what audio is native to the target platform before defaulting to a stock bed (and keep the license clean — Chapter 38, Appendix J).
• The hook could be even tighter. The on-screen text is nine words; the strongest feed hooks are often five. We could cut it to "The lighting mistake that looks cheap." Fewer words, faster read, muted.

None of these is a failure; they are the next iteration, surfaced by doing. A creator who repurposes ten videos and critiques each will outrun one who reads ten articles about it.

The bigger lesson

We spent more effort deciding than doing: assessing the source, planning the reframe, matching the pickup. The actual crop-and-caption work took under an hour. That ratio is the point — a good cut-down is mostly judgment (what's the hook, what to cut, how to reframe) executed with simple tools. And notice what made it possible: the original was shot well (clean audio, good light, a clear message), so it had something worth repurposing. You cannot cut a good vertical out of a bad horizontal. The teaser was easy because the source was thoughtful — the whole book's relationship between production and post, proven one more time, now across formats. You did it on a talking-head. You will do the same on all three of your projects, turning each finished piece into the vertical reach it needs.

Discussion questions

1. We shot a new native-vertical hook rather than cropping one out of the source's intro. Defend that choice on both quality and time grounds. When would cropping an existing moment be the better call?
2. The teaser cut three explanation points down to two and tightened every pause. How do you decide what to cut when repurposing — what has to survive, and what can go?
3. The pickup shot "disappears" into the reframed source only because it was matched. List every thing we matched, and what would have given the pickup away if we hadn't.
4. We ended on a loop instead of a sign-off. What did that trade get us, and what kind of video would a sign-off still be right for?
5. The captions were the difference between reaching the audience and scrolling past most of it. Why does the muted-comprehension test belong at the end of every short-form edit?
6. "You cannot cut a good vertical out of a bad horizontal." What does this imply about how you should shoot Project 2 and 3, knowing they'll be repurposed?

Your turn

Produce your own vertical cut-down following this exact process — this is the chapter's Production Checkpoint:

1. Inventory your source (your Project 1 talking-head or a chunk of Project 2). Find your best two seconds and note where the subject sits in the frame.
2. Plan the reframe beat by beat: what gets cropped-and-reframed, what gets stacked, what gets cut.
3. Shoot a native-vertical hook if the source lacks one — matched to the source's light and eyeline, promise in the first two seconds, on-screen text.
4. Build the vertical timeline in a 9:16 project: hook first, face center-safe throughout, story tightened, ending on a loop.
5. Caption everything — burned-in, large, high-contrast, safe-zone, proofread — and run the muted-comprehension test.
6. Export to the vertical spec (1080 × 1920, ~ −14 LUFS; confirm current specs in Appendix H) and watch it back on your actual phone, muted. Write one line in your Frame Log: does it stop your own thumb, and what's the weakest two seconds?

It will be imperfect. Post it anyway (or keep it) — it is your first proof that you can carry a story across the wall between horizontal and vertical, which is a skill the market now pays for on every job.

Key takeaways

• A vertical cut-down is a re-conception, not a shrink. Rebuild the horizontal piece for a vertical, muted, scrolling viewer while keeping its story intact.
• Front-load the hook — shoot one if the source lacks it. A matched native-vertical pickup beats cropping a weak opening out of an intro that was never built to grab.
• Crop-and-reframe to keep the face center-safe; set one crop per static shot, keyframe it only if the subject moves. Ride the eyes on the upper third even in vertical.
• Tighten for the feed. A teaser is a trailer for the full piece — cut to the strongest beats, kill the dead air, and let the full version live behind a link.
• Caption everything and run the muted test. Burned-in, safe-zone, proofread captions are how most of the audience experiences the video — and the muted-comprehension test is non-negotiable.
• End on a loop, not a sign-off, to convert a dead end into replays.
• You cannot cut a good vertical out of a bad horizontal — the cut-down is easy only because the source was shot well. Production and post are one craft, across every format.