# Hard cut from b-roll to a two-shot, then fade two lower thirds in simultaneously under each speaker

Manual DaVinci Resolve / Fusion construction guide generated from the recipe. It is not a tested .drp.
The accompanying Remotion project (recipe.json) is the executable definition; the numbers below are copied from it.
Neither construction establishes source fidelity: this is an original vector approximation of an inspected editing moment.

## Source and clock
Reference: OpenAI's "How California Small Businesses Are Using ChatGPT | OpenAI Academy" (cFyFF3Ad2aM), published 2026-07-22.
Source interval [1840, 1900) on the 24000/1001 fps master: 60 frames, 2.502500 s. Local frame 0 = source frame 1840.
Create a 1920x1080 progressive timeline at 24000/1001 fps and a Fusion composition exactly 60 frames long. Keep integer frames; do not round the frame rate.
Background field: #d9d6d0. Export silent H.264 yuv420p to match the intended playback container.

## Observation
Ten frames of b-roll (a seated speaker with a microphone at a dark table, crowd behind) then a hard cut at local frame 10 (master 1850) to a bright two-shot: two people standing in front of a cream wall and a curved black stair rail, banner at right. Starting at ~12 both lower thirds fade in at once, one under each person (left at x≈200, right at x≈1066), reaching full opacity by ~20 (master 1860). Each is a white semibold name over a lighter role line, no plate. The shot then holds.

## Mechanism
Cut then paired identification: a hard cut lands the two-shot, and two independent lower thirds share one fade so both subjects are named in a single beat.

## Phases (local frames)
- 0: B-roll
- 10: Hard cut to two-shot
- 12: Both lower thirds start fading in
- 20: Lower thirds fully on
- 45: Hold

## Layer inventory (absolute 1920x1080 top-left pixels, back to front)
- plate-a [group] at (0, 0) size 1920x1080 visible [0, 10)
  - a-wall [rect] at (0, 0) size 1920x1080 fill #b9b3aa
  - a-window [rect] at (700, 0) size 500x420 fill #e8ecea
  - a-window-2 [rect] at (1300, 0) size 300x420 fill #e8ecea
  - a-table [rect] at (0, 700) size 1920x380 fill #1f1f22
  - a-crowd [ellipse] at (400, 300) size 1300x500 fill #6f6a66
  - a-figure [ellipse] at (300, 250) size 600x900 fill #3a3838
  - a-figure-2 [ellipse] at (1150, 350) size 500x700 fill #4a4f6a
  - a-bottle [ellipse] at (780, 620) size 120x260 fill #9fc8be
- plate-b [group] at (0, 0) size 1920x1080 visible [10, 60)
  - b-wall [rect] at (0, 0) size 1920x1080 fill #d9d6d0
  - b-column [rect] at (1000, 0) size 180x1080 fill #e6e3de
  - b-banner [ellipse] at (1500, 0) size 600x900 fill #efd7e3
  - b-rail [path] at (1240, 560) stroke #1a1a1a 14px
  - b-figure-l [ellipse] at (80, 80) size 620x1200 fill #5a6288
  - b-figure-r [ellipse] at (780, 60) size 640x1240 fill #e0725f
- lt-left [group] at (200, 866) size 700x80 visible [12, 60)
    keyframes opacity: f12=0, f20=1 (out)
  - lt-left-name [text] at (200, 866) text "Cyrus Delgado" 30px weight 600 #ffffff
  - lt-left-role [text] at (200, 906) text "Owner · Delgado Architecture" 18px weight 400 #f2f2f2
- lt-right [group] at (1066, 866) size 700x80 visible [12, 60)
    keyframes opacity: f12=0, f20=1 (out)
  - lt-right-name [text] at (1066, 866) text "Rebeca Okafor" 30px weight 600 #ffffff
  - lt-right-role [text] at (1066, 906) text "CEO · Verde Design Studio" 18px weight 400 #f2f2f2

## Construction steps
1. Timeline 1920x1080 at 23.976 fps; Fusion composition 60 frames. Local 0 = master frame 1840.
2. Edit: clip A (b-roll) frames 0-9, clip B (two-shot) from frame 10. Straight cut, no handles.
3. Two Text+ pairs on clip B: names 30 px semibold white at (200,866) and (1066,866); roles 18 px near-white 40 px lower. No bars.
4. Both pairs share one Blend keyframe set: 0 at frame 12, 1 at frame 20, ease-out. Position the left block under the left subject and the right block under the right subject.
5. Hold to end of composition; nothing animates out.
6. Easy mistake: cascading the second lower third after the first; the source fades both in together two frames after the cut.

## Content variants
Default copy:
- nameL: Cyrus Delgado
- roleL: Owner · Delgado Architecture
- nameR: Rebeca Okafor
- roleR: CEO · Verde Design Studio
Adaptation changes:
- nameL: Marcus Lee
- roleL: Founder · Northshore Coffee
- nameR: Dana Whitfield
- roleR: Director · Whitfield Tutoring

## Limits
Easing names map to Fusion spline shapes: linear = linear, in/out/inOut = ease curves on the same frames, hold = step, spring/overshoot = slight overshoot past the target before settling.
Text reveal frames give the first and last unit; distribute characters/words evenly between them. Cursor click rings last 14 frames.
Vector illustration is an approximation of layout and motion, not the product UI or recovered production assets.