569「세로 모션 쇼릴」 — 점 하나가 15초 내내 장면을 잇고, 내 이름으로 끝나는 세로 영상

Make a 15-second vertical motion design showreel for my name and render it to MP4. Do the whole job now without asking questions.

[Inputs (use the default when empty)]
Name: "【이름】" — default 프롬프트판
Four short words, shown one by one (slash-separated): "【단어 네 개】" — default 칸 / 채우면 / 끝 / 프로급
Closing line: "【한 줄】" — default 칸만 채우면 프로급
Palette: "【색 조합】"
  기본 = orange #FF4B2B, ultramarine #2B3FFF, yellow #FFC93C, black #0A0A0A, off-white #F1EDE4
  형광 = acid lime #D7FF3E, hot pink #FF3E9A, cyan #3EE6FF, black #0A0A0A, off-white #F1EDE4
  차분 = terracotta #D96B4B, deep teal #1F5B5B, sand #E8C88A, ink #121417, paper #F4F0E8

[What makes it good]
- One hero: a single accent-colored dot. It splits, orbits, stamps, wipes, and finally lands as the period after the name. No scene starts from an empty frame; something from the previous moment always causes the next one.
- Every moment takes over the whole phone screen with a full-bleed color or one graphic idea. Never a small card floating in the middle.
- Type has weight: each word lands like a heavy object, stacks under the previous one, and the layout rebuilds itself to make room instead of cutting. Words never overlap or touch while moving.
- One pure idea at a time: a dot splitting into an orbit; a shape morphing circle → rounded square → triangle → cluster; a field of large dots with a color wave that folds into a sphere, a ring and a wire cube; a single ribbon drawing a loop; the words riding a large spinning ring.
- Rhythm with one real rest: around 9.5 to 11 s, one calm full-screen moment with a single slow ribbon. The music keeps playing there.
- Nothing but the motion: no corner labels, no frame counter, no year, no captions or watermarks.

[Storyboard — follow these times and grounds exactly]
Palette roles, in the order listed above: ACCENT, SECOND, THIRD, DARK, LIGHT.
0.00–1.75 · DARK ground. One ACCENT dot splits into children that orbit on a tall ellipse.
1.75–2.00 · ACCENT grows from their center and covers the whole screen.
2.00–3.75 · ACCENT ground. Word 1 lands huge with weight; word 2 swings in beneath it; thin corner brackets snap around them; both rise to make room as word 3 drops in below; then everything contracts into the ACCENT dot.
3.75–5.50 · DARK ground. The dot grows into a circle, morphs to a rounded square (SECOND), a triangle (THIRD) and a grainy cluster, inside thin corner guides.
5.50–7.25 · LIGHT ground. The cluster spreads into a tall field of large dots; an ACCENT wave climbs through it.
7.25–9.50 · SECOND ground. The same dots fold into a surface, a sphere, a ring and a wire cube; the camera pushes through the cube.
9.50–11.25 · THIRD ground, the calm moment. One ACCENT ribbon slowly draws a tall loop and withdraws. No text.
11.25–13.25 · DARK ground. The loop becomes a large elliptical track; words 1 to 4 ride it upright and readable, rotate, then contract into the ACCENT dot.
13.25–15.00 · LIGHT ground. The dot moves to the end of the name, which resolves very large; the closing line settles beneath it and holds.
Every one of these grounds must fill the entire screen. Do not keep one background color for the whole film.

[Layout for a phone]
1080×1920, 60 fps, exactly 15 seconds. Keep everything that matters inside 90 px from the sides and 180 px from top and bottom; nothing important is ever cut by the screen edge, including rotating text. Big words are at least 220 px tall and single short words at least 320 px; ring words at least 110 px and always upright enough to read, never mirrored; the final name fills about 86% of the width and the closing line is at least 64 px. Dot fields use dots big enough to read on a phone. Use a heavy Korean display font (install @fontsource/black-han-sans, SIL Open Font License; fall back to Noto Sans KR Black) and check every glyph renders. Print each input exactly once and add no other words.

[Sound]
Compose an original beat in code at about 120 BPM: tight kick, soft clap, simple bass line, clean synth stab; cut the picture to the same beat data. During the calm moment remove only the drums and keep bass and a soft chord about 6 dB lower, so it stays one continuous song with no gap. Measure loudness in 0.5-second windows; no window may fall more than 8 dB below the average. Integrated about -14 LUFS, true peak below -2 dB.

[Build]
Node.js, @napi-rs/canvas and FFmpeg; no browser needed. All drawing comes from a pure seek(t). Keep all inputs and colors in config.json. Write render.mjs so `node render.mjs` makes a new MP4 (H.264 + AAC, yuv420p).

[Check before finishing]
Write plan.md first: the three great moments, the calm moment, and for every transition the object that carries across. After rendering, export frames every 0.25 s in 5×4 strips stamped with time, open them, and fix any step that looks like an unrelated new picture, any overlap, any text cut by the edge, and any dead hold. Then render again. Deliver motion.mp4, render.mjs, config.json and plan.md.

[Final output format]
Output size 1080x1920 pixels, 9:16 portrait story. Keep all text inside the central 80% of the height: compress every zone percentage below into the 10-90% range.
This selected output format takes priority over any fixed canvas size or aspect ratio elsewhere in this template. Adapt the composition, text zones and margins to the selected canvas without stretching or cropping essential content. Device safe areas apply only when a device is selected. Never print these format instructions on the image.