nphu0811/pmg-skills

pmg-3d-cinematic

Prepare and compose 3D assets, model orientation, technical cutaways, exploded views and annotations.

소스 보기

사용 방법

  1. 설치 명령을 복사하고 자신의 에이전트 환경에 설치하기 전에 소스 패키지를 확인하세요.
  2. 필요한 도구, 인증 정보, 지원 환경은 원본 문서에서 확인하세요.
  3. 입력과 원하는 산출물을 설명하고 출처와 결과를 대조하세요.
npx skills add https://github.com/nphu0811/pmg-skills --skill pmg-3d-cinematic

시작 전 확인

  • 설치에는 모델 구독, API 한도, 파일 접근 권한이 포함되지 않습니다.
  • 별점 수와 설치 수는 활동 지표이며 독립적인 품질·보안 검증이 아닙니다.
원본 Skill 문서

원본 저장소의 제목, 예시, 코드, 표, 링크, 이미지를 유지해 표시합니다.

PMG 3D Cinematic Engine (pmg-3d-cinematic)

Responsibility and dependencies

3D asset preparation, model placement/orientation, cutaways, exploded views, annotations, model-focused composition and model compatibility. Use pmg-camera-parallax for all camera transforms; choreography owns high-level sequencing.

Dependencies: pmg-engine; pmg-camera-parallax; pmg-morph; core/asset helpers. All generated candidates must pass PMGOrchestrator before delivery; direct save/COM snippets below are lower-level composition examples, not independent production entrypoints.

Transforms PowerPoint into a cinematic 3D documentary experience. Built on Native PowerPoint 3D Models (`.glb`), Morph Camera Rotations, and Macro Dolly Zooms, adaptable to any domain without complex prompts.


1. Native 3D Model Architecture (.glb / <am3d:model3d>)

PowerPoint features a native 3D rendering pipeline supporting industry-standard .glb and .gltf 3D assets:

xml
<am3d:model3d r:embed="rId8">
  <am3d:spPr>
    <a:xfrm><a:off x="0" y="0"/><a:ext cx="7985021" cy="7623432"/></a:xfrm>
  </am3d:spPr>
  <am3d:camera>
    <am3d:pos x="0" y="0" z="56149260"/>
    <am3d:lookAt x="0" y="0" z="0"/>
    <am3d:perspective fov="2700000"/>
  </am3d:camera>
  <am3d:trans>
    <am3d:scale><am3d:sx n="1000000" d="1000000"/><am3d:sy n="1000000" d="1000000"/><am3d:sz n="1000000" d="1000000"/></am3d:scale>
    <am3d:rot rx="1500000" ry="-3000000" rz="0"/>
  </am3d:trans>
</am3d:model3d>

3D Morph Camera Interpolation

When the same 3D model appears on consecutive slides with different camera angles or rotations: $$\mathbf{R}1(\theta{x1}, \theta{y1}, \theta{z1}) \xrightarrow{\text{Morph 60 FPS}} \mathbf{R}2(\theta{x2}, \theta{y2}, \theta{z2})$$ PowerPoint's hardware rasterizer smoothly rotates and scales the 3D model in real time without video renders.

Python COM Insertion

python
# Programmatically insert native 3D GLB model via PowerPoint COM
model = slide.Shapes.Add3DModel("assets/earth_model.glb", Left=100, Top=50, Width=500, Height=500)
model.Model3DFormat.RotationX = 15
model.Model3DFormat.RotationY = -45
model.Model3DFormat.RotationZ = 10

2. Shared camera direction: Macro 3D Dolly Zoom ($100\% \to 400\%$)

Demonstrated in top-tier science decks (e.g. Earth & Space):

+-------------------------------------------------------------------------------+
| STAGE 1: WIDE ESTABLISHING ORBIT (Camera Zoom 100%)                           |
|   - 3D Model / Globe centered at normal scale (w: 8.0", h: 7.6").             |
|   - Deep spatial perspective, titles floating in space.                       |
|                                                                               |
| STAGE 2: THE APPROACH & ROTATION (Camera Zoom 180%)                           |
|   - Model rotates by 45° around Y-axis, scaling to (w: 14.8", h: 14.0").      |
|   - Camera draws closer, revealing regional topography.                       |
|                                                                               |
| STAGE 3: TOUCHDOWN MACRO PUSH-IN (Camera Zoom 400%)                           |
|   - Massive camera plunge! Model scales to (w: 41.0", h: 19.2").              |
|   - The horizon fills the entire viewport; audience feels surface arrival!   |
|                                                                               |
| STAGE 4: HORIZON TRACKING PAN (Camera Zoom 220%)                              |
|   - Model slides laterally along the rim (x: -6.3" -> 32.8").                 |
|   - Stages open space for deep-dive technical callout cards.                  |
+-------------------------------------------------------------------------------+

Universal Domain Adaptation

This 4-beat camera move applies to any subject:

  • Astronomy & Earth Science: Deep space $\to$ Planetary orbit $\to$ Atmospheric touchdown.
  • Consumer Hardware / Tech: Device exterior $\to$ Enclosure opening $\to$ Microchip die macro zoom.
  • Automotive & Engineering: Aerodynamic chassis $\to$ Hood cutaway $\to$ Turbocharger cylinder.
  • Healthcare & Biotech: Human anatomy $\to$ Organ tissue $\to$ Cellular DNA double helix.
  • Ecology & Geography: World map $\to$ National coastline $\to$ Mangrove forest canopy.

3. The 3D Orbiting Carousel (Team & Feature Showcase)

Arranges cards, team members, or feature modules along a virtual 3D elliptical orbit:

                  [ Card 3: Deep Background (Scale 60%, Opacity 40%) ]
                 /                                                     \
   [ Card 2: Left Flank ]                                     [ Card 4: Right Flank ]
      (Scale 85%, Opacity 75%)                                   (Scale 85%, Opacity 75%)
                 \                                                     /
                  [ Card 1: FOREGROUND HERO (Scale 120%, Opacity 100%) ]
  • Morph Dynamics: Advancing the slide rotates all cards around the orbit.
  • The hero card steps aside, and the next card rotates into the foreground, smoothly enlarging to $120\%$ scale.

4. HUD Callouts & Leader Lines

Anchor data cards directly to the 3D model component using vector Target Dots:

[ Target Dot ]  (Dual circle: outer glow 0.25", inner dot 0.08")
      |
      |  (Hairline 1.2pt leader line, angled 45° or 90°)
      v
+-------------------------------------------------------------+
| !!HUD_Card (Frosted Glass Card, 3.8" x 1.3")                |
|   TITLE IN BOLD: ATMOSPHERIC BOUNDARY LAYER                 |
|   Desc: High-velocity jet streams drive regional climates.  |
+-------------------------------------------------------------+
  • Visual-First Rule: Keep text under 25 words per slide. Let the 3D model carry 85% of the cognitive load.

Native GLB/COM and OpenXML snippets describe host-level capabilities, not a fully verified model importer in this release. Test the actual target PowerPoint version, register media/relationships, and use the canonical validation gate. Do not claim support solely from the example XML. Carousel coordination belongs to choreography; this skill supplies model-focused composition.

같은 저장소의 Skills

더 많은 Skills

모든 Skills