Bay lượn giữa terminal 3D

Mọi thứ trên trang này đều có bản máy đọc. Đưa agent link JSON, hoặc thêm MCP server gimgs rồi bảo nó tìm motion template.

1. Tải JSON — trong đó có công thức cảnh quay (timing, easing, bảng tham số, lỗi hay gặp) và toàn bộ source TSX.
2. Clone repo video-shotcraft; component import fixture chung từ demos/_fixtures và helper từ assets/lib.
3. Đăng ký component thành <Composition> trong project Remotion rồi render bằng `npx remotion render`, hoặc dựng lại công thức bằng engine của bạn.

curl -s https://gimgs.net/motion/terminal-3d.json

git clone https://github.com/Vincentwei1021/video-shotcraft
# component: demos/camera/terminal-3d/Terminal3D.tsx
npx remotion render <CompositionId> out.mp4

claude mcp add --transport http gimgs https://gimgs.net/mcp
# then: search_motion_templates / get_motion_template("terminal-3d")

Use the video-shotcraft shot recipe "Bay lượn giữa terminal 3D" (Remotion composition, Apache-2.0) published on gimgs.net.
Recipe JSON (prompt = timing/easing/parameter table/pitfalls, files = TSX source): https://gimgs.net/motion/terminal-3d.json
Raw TSX: https://gimgs.net/motion/terminal-3d/template.html
Repository: https://github.com/Vincentwei1021/video-shotcraft  (component: demos/camera/terminal-3d/Terminal3D.tsx; imports demos/_fixtures and assets/lib)
Spec: 1920x1080, 30 fps, 6.5s

1. Read the JSON and follow its "prompt" section (when to use, intention, parameter table, known pitfalls).
2. Adapt it for: <describe what you want>.
3. Either clone the repo, register the component as a <Composition> in a Remotion project and run npx remotion render <CompositionId> out.mp4 — or re-implement the recipe in your own engine (HTML/WAAPI works: keep the same timing and easing).

JSON · Markdown · llms.txt

// ===== demos/camera/terminal-3d/Terminal3D.tsx =====
// terminal-3d — Terminal 3D 命令执行叙事流（motion-lab 定稿转原生 Remotion）
// 多个通用桌面风格终端窗散布 3D 空间不同位置与角度，相机在窗间飞行（途中略微
// 拉远）；每到一窗，窗内打字机敲出命令并逐行吐出结果，形成命令执行叙事流。
// 设计坐标 480×270（DesignStage 等比放大），参数表数值以此坐标系标定。
import React from 'react';
import { DesignStage, E, lerp, seg, useT } from '../../_fixtures/Motion';

export const TERMINAL_3D_DURATION = 180; // 6000ms @30fps

// 光栅化补偿系数：3D 子树里 Chrome 按布局尺寸取样文字，经 DesignStage 放大 4x
// 后小号等宽字会糊。把整个 3D 场景数值 ×K 建模、再用外层 scale(1/K) 抵消——
// 透视投影对整体缩放自相似（p·x/(p-z) 等比），画面几何与 effect.js 逐像素一致，
// 仅文字光栅化分辨率提升到与原样片（4x device scale 截帧）相同。
const K = 4;

// 关键帧累加器：base 起步，每段 [at0,at1] 内朝 to 推进（效果源文件级共享 helper）
type Vec = Record<string, number>;
const acc = (t: number, base: Vec, kfs: { at: number[]; to: Vec }[], keys: string[], ease: (x: number) => number) => {
  const out: Vec = {};
  for (const k of keys) out[k] = base[k];
  let prev = base;
  for (const kf of kfs) {
    const u = seg(t, kf.at[0], kf.at[1], ease);
    for (const k of keys) out[k] += u * (kf.to[k] - prev[k]);
    prev = kf.to;
  }
  return out;
};

// 三个终端窗：3D 位姿 + 标题 + 命令 + 输出行（数值为 480×270 设计坐标）
const DATA = [
  {
    pose: { x: -300, y: -34, z: -110, ry: 24 },
    title: '~/workspace — zsh',
    cmd: '$ git status -sb',
    out: ['## main...origin/main', ' M src/timeline.ts', ' M src/camera.ts', '?? fx/b01.js'],
  },
  {
    pose: { x: 96, y: 62, z: 90, ry: -16 },
    title: 'dev server',
    cmd: '$ npm run dev',
    out: ['vite v5.2.0  ready in 312 ms', '➜  local:   http://localhost:3000', '➜  network: 192.0.2.10:3000', 'watching 148 modules'],
  },
  {
    pose: { x: 402, y: -74, z: -60, ry: -32 },
    title: 'logs',
    cmd: '$ tail -f server.log',
    out: ['12:04:11 GET /api/render 200 41ms', '12:04:12 POST /api/queue 201 88ms', '12:04:14 worker#3 frame 240/270', '12:04:15 done → out/final.mp4'],
  },
];

// Step 时序：每窗 [飞行起, 飞行止]；TYPE 为各窗打字起点
const STEP = [[0, 0.02], [0.30, 0.44], [0.64, 0.78]];
const TYPE = [0.05, 0.47, 0.81];
const PK = ['x', 'y', 'z', 'ry'];

export const Terminal3D: React.FC = () => {
  const t = useT();

  // 相机姿态 = 关键帧累加（两段飞行，inOutCubic）
  const v = acc(t, DATA[0].pose, [
    { at: STEP[1], to: DATA[1].pose },
    { at: STEP[2], to: DATA[2].pose },
  ], PK, E.inOutCubic);
  // 飞行途中拉远：两段过渡各一个正弦鼓包
  let pull = 0;
  for (let i = 1; i < 3; i++) {
    const u = seg(t, STEP[i][0], STEP[i][1]);
    pull += Math.sin(u * Math.PI) * 210;
  }

  return (
    <DesignStage bg="#07080e">
      {/* 反缩放包裹层：把 ×K 建模的场景缩回设计坐标（见 K 注释） */}
      <div
        style={{
          position: 'absolute',
          left: 0,
          top: 0,
          width: 480 * K,
          height: 270 * K,
          transform: `scale(${1 / K})`,
          transformOrigin: 'top left',
        }}
      >
        {/* scene：透视容器 + 深空渐变背景 */}
        <div
          style={{
            position: 'absolute',
            inset: 0,
            background: 'radial-gradient(120% 90% at 50% 0%,#141826,#07080e 70%)',
            perspective: `${900 * K}px`,
            overflow: 'hidden',
            WebkitFontSmoothing: 'antialiased', // 原样片截帧为灰度平滑，避免笔画偏粗
          }}
        >
          {/* world：相机逆变换载体 —— 先拉远/推近，再反转角，再反平移 */}
          <div
            style={{
              position: 'absolute',
              inset: 0,
              transformStyle: 'preserve-3d',
              transform: `translateZ(${(300 - pull) * K}px) rotateY(${-v.ry}deg) translate3d(${-v.x * K}px,${-v.y * K}px,${-v.z * K}px)`,
            }}
          >
            {DATA.map((d, i) => {
              const p = d.pose;
              // 离焦：与当前相机 x 距离越远越暗越糊
              const focus = 1 - Math.min(1, Math.abs(v.x - p.x) / 420);
              // 打字机：命令逐字点亮，光标闪烁，输出逐行滑入
              const ty = seg(t, TYPE[i], TYPE[i] + 0.09);
              const n = Math.floor(ty * d.cmd.length + 0.0001);
              const caretOp = ty >= 1
                ? (Math.floor(t * 26) % 2 ? 0.15 : 0.9)
                : (Math.floor(t * 40) % 2 ? 0.35 : 1);
              return (
                <div
                  key={i}
                  style={{
                    position: 'absolute',
                    left: '50%',
                    top: '50%',
                    width: 300 * K,
                    height: 176 * K,
                    margin: `${-88 * K}px 0 0 ${-150 * K}px`,
                    borderRadius: 9 * K,
                    background: '#0e1017',
                    overflow: 'hidden',
                    boxShadow: `0 ${24 * K}px ${60 * K}px rgba(0,0,0,.7),inset 0 0 0 ${K}px #2a3040`,
                    transform: `translate3d(${p.x * K}px,${p.y * K}px,${p.z * K}px) rotateY(${p.ry}deg)`,
                    opacity: 0.34 + focus * 0.66,
                    filter: `blur(${(1 - focus) * 2.2 * K}px) brightness(${0.7 + focus * 0.3})`,
                  }}
                >
                  {/* 标题栏：红黄绿灯 + 居中标题 */}
                  <div
                    style={{
                      position: 'absolute', left: 0, top: 0, width: '100%', height: 22 * K,
                      background: 'linear-gradient(180deg,#242a38,#1b202b)', borderBottom: `${K}px solid #2c3242`,
                    }}
                  >
                    {['#ff6058', '#ffbd2e', '#28ca42'].map((c, k) => (
                      <div
                        key={c}
                        style={{
                          position: 'absolute', left: (9 + k * 13) * K, top: 8 * K, width: 7 * K, height: 7 * K,
                          borderRadius: '50%', background: c,
                        }}
                      />
                    ))}
                    <div
                      style={{
                        position: 'absolute', left: 0, top: 0, width: '100%', height: 22 * K, textAlign: 'center',
                        font: `600 ${8 * K}px/${22 * K}px Courier,monospace`, color: '#77809b',
                      }}
                    >
                      {d.title}
                    </div>
                  </div>
                  {/* 命令行：逐字 opacity + 尾随光标 */}
                  <div
                    style={{
                      position: 'absolute', left: 12 * K, top: 32 * K,
                      font: `600 ${9.5 * K}px/1 Courier,monospace`, color: '#9dffcf', whiteSpace: 'pre',
                    }}
                  >
                    {Array.from(d.cmd, (ch, c) => (
                      // 空格换成 nbsp，对齐 effect.js 的 textContent 处理
                      <span key={c} style={{ opacity: c < n ? 1 : 0 }}>{ch === ' ' ? ' ' : ch}</span>
                    ))}
                    <span
                      style={{
                        color: '#9dffcf',
                        opacity: caretOp,
                        display: 'inline-block',
                        transform: `translateX(${(d.cmd.length - n) * -0.1 * K}px)`,
                      }}
                    >
                      ▌
                    </span>
                  </div>
                  {/* 输出行：命令敲完后逐行 stagger 淡入 + 左滑归位 */}
                  {d.out.map((o, k) => {
                    const ou = seg(t, TYPE[i] + 0.10 + k * 0.022, TYPE[i] + 0.145 + k * 0.022, E.outCubic);
                    return (
                      <div
                        key={k}
                        style={{
                          position: 'absolute', left: 12 * K, top: (52 + k * 16) * K,
                          font: `500 ${9 * K}px/1 Courier,monospace`,
                          color: k === 0 ? '#c9d3ea' : '#7f8aa6',
                          whiteSpace: 'pre',
                          opacity: ou,
                          transform: `translateX(${lerp(ou, -7, 0) * K}px)`,
                        }}
                      >
                        {o}
                      </div>
                    );
                  })}
                </div>
              );
            })}
          </div>
        </div>
      </div>
    </DesignStage>
  );
};

1. Read the recipe
   Prompt = when to use, intention, parameter table with typical values and how each one feels, known pitfalls. Treat values as calibrated starting points.
2. Get the code
   tool: git
   git clone https://github.com/Vincentwei1021/video-shotcraft && cd video-shotcraft && npm install
3. Register + render
   tool: remotion
   npx remotion render <CompositionId> out.mp4 # component: demos/camera/terminal-3d/Terminal3D.tsx
4. Or re-implement
   Port the timing/easing from the recipe to HTML + WAAPI (gimgs motion contract) or any engine; keep the same segment windows and easing.