Leadership Series

# Didone Sample — Leadership Series An eight-scene vertical reel. Four scenes are recreated from a 20.9s screen recording of a 314x560 preview; four more (`split`, `figure`, `list`, `statement`) are authored from the same primitives, each one bending a different rule of the original. One device carries every scene: a centred stack of high-contrast serif caps, interleaved with italic lines on a highlighter chip. - **A headline line** is white caps at 107px, and it **hard-cuts in** — measured at 60fps, a line goes from nothing to full width inside one frame. No fade, no rise, no type-on. - **A highlighted line** is italic on the accent chip, and it **wipes open left to right** over 0.38s. The plate grows with the reveal because the wipe is a `clipPath` on the per-line pill, which hugs its own words — so it works at any copy length. - **A footnote** is the same chip in white at 37px, and it **unrolls from the left**: the pill slides in by 130% of its own width behind a left-edge clip, so the tail of the string leads and the line arrives last character first. Every line is its own layer, and for the chips that is not tidiness: they overlap their neighbour, and a single text layer cannot overlap its own lines. ## The chip pill, and how far chips may overlap A `backgroundFit: "line"` pill is `lineHeight + 0.32em` tall (0.16em of padding either side) and the italic's ink box is `ascender + descender` = 1.30em, so the type's position inside the pill is fixed by the line height alone: ``` lineHeight 1.092em -> pill 1.412em ink top 0.056 baseline 1.046 descender 1.356 ``` **The pitch to the next chip is therefore a hard floor, not a taste call.** Under 1.046em the pill below starts ABOVE the baseline and slices the bottom off every letter — which is what a first pass at `lineHeight 0.948em` did, where the pill was shorter than the ink it held. Two pitches are used, both in the CHIP's own em because that is the unit the pill and its baseline are measured in: | where | pitch | overlap | costs | | --- | --- | --- | --- | | `series`, the source's own two-line chip | 1.05em | 26% of the pill | lands on the baseline, as the recording does | | `list`, three in a row | 1.30em | 8% | 1.5px off the tip of a descender | The recording only ever stacks two chips, and at 26% it clips the tail of its own `p` in "leadership". Three of them at that pitch is unreadable, so the list opens up to a pitch that keeps every letter whole and still reads as one interlocked block. The pill itself is 34.5px at the source's scale in both — measured off the frames. ## Scenes Four scenes are the recording. The other four each add exactly ONE structural move the recording never makes, so the reel has range instead of eight variations on a centred stack: | scene | | the move | stack | | --- | --- | --- | --- | | `insight` | measured | — | 3 caps + chip, centred low | | `series` | measured | — | cap, 2 chips, cap, plus a dated footnote, centred high | | `split` | new | **the frame splits** — a chip alone at the top, the caps alone at the foot, nothing between. Silent by design (`meta.silent`) | chip at 11%, 2 caps resting on 86% | | `figure` | new | **two cap sizes** — a figure at 2.4x over a normal line, the only place the one-size discipline breaks | XL cap + cap + chip | | `expertise` | measured | — | chips INSIDE the stack rather than under it | | `list` | new | **no white type at all** — the white plate leads and the accent carries the frame, three chips interlocking under it | footnote + 3 chips | | `statement` | new | **not centred** — ragged left on a 6% margin, the block resting on the lower edge | 3 caps + chip, left-aligned | | `close` | measured | — | 3 caps + chip + the standing URL | Every new scene is built from the same three primitives (a hard-cut cap line, a chip that wipes open, a footnote that unrolls) on the measured cadence of ~0.65s an item — what changes is alignment, anchor, scale or which primitive is absent. ## The motion, as measured | beat | at (scene-local) | | --- | --- | | `insight` lines | 0.00 · 1.13 · 2.10, chip 3.13 | | `series` | 0.13, chips 0.87 / 1.53, `SERIES.` 2.17, footnote 2.53 | | `expertise` | 0.43, chip 1.07, 1.57 · 2.13, chip 2.63 | | `close` | 0.00 · 0.73 · 1.43, chip 1.90, footnote 2.70 | The wipe's growth decelerates hard — width reaches 43% of final in the first fifth of the run — which is `power3.out`, not `power2.out` and not linear. The long chip reads ~0.42s and the short ones 0.37s off the frames; that is the ease's tail crossing a detection threshold at different widths, not two durations, so all of them are authored at 0.38s. Nothing is keyed to a scene's end. The recording holds each finished stack until it cuts; none of that is ported, because the narration will not be the length of the placeholder. ## Type Two faces, both as `url` fonts rather than catalogue keys, because both italics and weights the builtin loader does not request are load-bearing here. One href serves them: `css2?family=Instrument+Serif:ital@1&family=Playfair+Display:wght@700;800`. | role | face | size (9:16) | | --- | --- | --- | | headline caps | Playfair Display 800, `-0.04em`, set solid | 97px | | the chip | Instrument Serif italic | 84px | | the footnote | Playfair Display 700 | 36px | **The source's own face exists nowhere.** Fingerprinted against 77 downloaded faces — cap-normalised ink widths over eleven measured strings — every candidate lands within 4% residual and *none* reconciles the two measurements: pinned to the frames' line widths, each one implies a cap height of 17-20px where the frames show 22.5. It is a condensed didone, most likely a proprietary one, and the discrepancy is a consistent ~1.2x — the signature of type that was condensed rather than scaled. So the widths are honoured and the cap height is not, because line width is what a viewer compares at a glance: - **Caps:** Playfair Display 800 at 97px with `-0.04em` tracking lands the eleven headline strings at a 4.3% RMS error (worst +8% on `IS NOT`), at a cap 11% short of the frames'. - **The chip:** Playfair's italic is 46% wider per unit of pill height than the recording's, which is why the first pass drew a yellow bar across the whole frame. Instrument Serif's italic is the closest of 27 italics tested — it hits the six measured pill widths at 5% RMS with a cap only 12% short, against Playfair italic's 30%. Its 400 is the family's only weight, and on dark ink at chip size that reads correctly. - Both faces therefore land ~11% short in cap at matched widths, so the whole composition is the recording at about 89% scale with its line widths intact and every internal proportion — the solid leading, the chip overlap, the chip's own aspect — preserved. Every stored value came out of `template:units`. ## Palette Eight variants on four roles — `accent` (the chip), `light` (the caps and the footnote plate), `ink` (the type on both plates), `ground` (behind the photo). Every accent stays light enough to carry near-black `ink` at chip size, which is the one contrast pair the design cannot lose. ## Backdrops Four generic office stills stand in for the recording's own footage. They are placeholders: each is a `media` slot marked `broll`, so a narrated scene cuts between shots rather than holding one still.